Synergistic Effects of Bacteria, Enzymes, and Cyclic Mechanical Stresses on the Bond Strength of Composite Restorations

Abstract

Predicting how tooth and dental material bonds perform in the mouth requires a deep understanding of degrading factors. Yet, this understanding is incomplete, leading to significant uncertainties in designing and evaluating new dental adhesives. The durability of dental bonding interfaces in the oral microenvironment is compromised by bacterial acids, salivary enzymes, and masticatory fatigue. These factors degrade the bond between dental resins and tooth surfaces, making the strength of these bonds difficult to predict. Traditionally studied separately, a combined kinetic analysis of these interactions could enhance our understanding and improvement of dental adhesive durability. To address this issue, we developed and validated an original model to evaluate the bond strength of dental restorations using realistic environments that consider the different mechanical, chemical, and biological degradative challenges working simultaneously: bacteria, salivary esterases, and cyclic loading. We herein describe a comprehensive investigation on dissociating the factors that degrade the bond strength of dental restorations. Our results showed that cariogenic bacteria are the number one factor contributing to the degradation of the bonded interface, followed by cyclic loading and salivary esterases. When tested in combinatorial mode, negative and positive synergies toward the degradation of the interface were observed. Masticatory loads (i.e., cycling loading) enhanced the lactic acid bacterial production and the area occupied by the biofilm at the bonding interface, resulting in more damage at the interface and a reduction of 73% in bond strength compared to no-degraded samples. Salivary enzymes also produced bond degradation caused by changes in the chemical composition of the resin/adhesive. However, the degradation rates are slowed compared to the bacteria and cyclic loading. These results demonstrate that our synergetic model could guide the design of new dental adhesives for biological applications without laborious trial-and-error experimentation.

1. INTRODUCTION

After nearly seventy years of discovery, resin-based composite restorations continue to demonstrate limited long-term clinical effectiveness [1–3]. More than half fail within a decade of placement [4–6]. In addition, replacing these failed restorations constitutes about 60% of the daily workload in dental practices that saturate the healthcare system [1, 7, 8]. The estimated cost of managing oral biofilm-associated diseases (e.g., recurrent caries) amounted to $81 billion/year [9]. The limited longevity of composite restorations is not only about biomaterials but also about patient factors, including bruxism [8]. In fact, caries at the restorative margins and biomaterial fractures are the primary reasons for the early failure of composite restorations [10]. The bond formed at the interface between the tooth (dentin/enamel) and the restorative materials weakens prematurely [5].

Bacteria, saliva, and cyclic mechanical masticatory stresses induce the fast degradation of all bonded biomaterials [4, 11–13]. Resins are primarily degraded by the acids and enzymes produced by acidogenic bacteria (e.g., Streptococcus mutans) [14–17] and by hydrolysis sourced from the aqueous content of saliva [18]. Elution of unreacted monomers up-regulates bacterial activity, aggravating the production of acids/enzymes [19, 20]. The esterase enzymatic activities from saliva [14, 17, 21] and the cyclic masticatory forces also contribute to the bond weakening [22–24]. Dentin is demineralized by the acids produced by bacteria [16, 25] and enzymatically degraded by the activation of the dentinal matrix metalloproteinases (MMP) [26, 27]. This breakdown of biomaterials results in widening the restoration margins, allowing additional bacteria/saliva to infiltrate the bonded interface (restoration margin) [28]. This vicious cycle exacerbates the degradation rates, resulting in the fast reduction of the bond strength and the early failure of composite restorations [24, 29].

Assessing the bond strength of resin composites to dental hard tissues is a known challenge with multiple laboratory approaches developed over the years [30]. Specifically, replicating the interplay of the different mechano-bio-chemical sources (bacteria, cyclic loading, saliva) insulting and degrading the bond has been difficult, primarily due to the bacteria- and enzyme-rich characteristics of the oral cavity [31, 32]. Current approaches that evaluate the degradation of the bond study each challenge independently or in sequence [30, 33–37]. For example, bonded interfaces are initially exposed to biofilm or salivary challenges, then subjected to repetitive loading (fatigue), and finally subjected to failure to measure the bond strength [34, 36, 38]. Unfortunately, this sequence does not accurately reflect the realistic conditions inside the mouth. The assumptions made in these laboratory methods do not closely replicate the concurrent degradative environment, which could lead to an underestimation of the clinical performance of the restoration or bond [23, 35, 39, 40].

Clinically, bonded biomaterials are continuously and simultaneously exposed to multiple degradative factors, all of which could contribute to the breakdown of biomaterials and the low durability of the restoration [12]. Clinical studies may provide an accurate assessment of the durability of dental restorations. However, these studies are laborious, need a large sample size, require long time, patient compliance, and are difficult to control since restorations are also influenced by the operator, defect size, and defect location [29, 35]. Thus, laboratory approaches that predict the bond of composite restorations considering the different mechanical, chemical, and biological degradative challenges working simultaneously are needed.

In recent years, there has been a significant surge in the development of bioactive resin composites/adhesives offering antimicrobial properties, remineralization capabilities, self-healing/repair, and resistance against enzymatic activities [41–47]. Some of these biomaterials claim to improve the restoration’s durability by deterring attacks from the different degradative sources. While we have gained insights into various sources of bond degradation to inform the design of bioactive therapies, understanding the intricate interplay, potential synergies among different sources, and the hidden mechanisms driving bond degradation still eludes us. Without a proper laboratory approach that closely replicates the challenges attacking the bond, the prediction of bioactive biomaterials to improve the clinical service of the bond remains questionable. Therefore, it is crucial to have accurate laboratory models to assess how bioactive therapies enhance bonding with restorative materials effectively.

To illustrate, consider a scenario where a resin composite adhered to dentin is simultaneously attacked by cariogenic biofilms and enzymatic-rich environments while concurrently withstanding the repetitive (mastication) forces for millions of cycles. Through this innovative modeling, it was possible to assess the individual and combined synergistic contribution of bacterial/salivary activities (acidity and enzymatic actions) and mechanical deterioration (cyclic loading) on the strength of the bond of dental restorations.

2. EXPERIMENTAL SECTION

2.1. Preparation of the Dentin-Adhesive-Composite Interface

To study the degradation of the dentin-adhesive-composite interfaces, “twin interface” samples were prepared as described previously [48, 49]. Briefly, healthy extracted molars were obtained from the Kornberg School of Dentistry and the University of Maryland School of Dentistry clinics. After extraction, teeth were refrigerated (4°C) in Hank’s Balanced Salt Solution (HBSS) and stored for maximum 1 month [50]. Teeth were sectioned (Leco VC-50) to obtain dentin blocks (2×2×2 mm³). These dentin sections were consistently cut from the mid-coronal region to guarantee similar tubule density and mineral/collagen ratio. Dentin blocks were transferred to a metal mold with the occlusal surface facing upwards, so the composites were bonded on the mesial and distal sides (Figure 1a). This orientation led to dentinal tubules positioning parallel to the interface (e.g., θ = 0°) [51], similar to the orientation when a class I restoration is prepared.

Figure 1. Integrated modeling of the multiple degradation sources affecting adhesive-composite-dentin bonded interfaces.

(a) Schematic diagram showing the preparation of twin-interface samples. (b) A three-factor experimental design was used to evaluate the contribution of each degradative source, including bacteria, saliva (i.e., pseudocholinesterase (PCE) and cholesterol esterase (CE) enzymatic activities), and cyclic loading and their combinations/interactions on the reduction of the bond strength. (c) Bacteria supplementation during experiments. After bacterial adhesion, samples were incubated in phosphate-buffered saline (PBS) at 37°C. BHI media and sucrose supplementation were conducted 5 times/day by temporarily submerging the colonized samples for 1 minute.

The total-etch technique was employed to bond the dentin block with an adhesive/composite system, simulating the steps and procedures conducted in the clinic [52]. We chose to use the total-etch technique because this method has been shown to provide higher fatigue strength compared to self-etch systems [53]. This choice ensures that the bonded interfaces are mechanically robust, allowing us to better assess the degradation mechanisms under the combined challenges of bacterial, enzymatic, and cyclic loading. First, the dentin’s mesial and distal proximal surfaces were etched with 32% phosphoric acid (3M^™ Scotchbond^™ Universal Etchant Etching Gel) for 15 seconds, followed by a 10-second gentle water rinse. Second, a single layer of adhesive (3M Scotchbond Universal Adhesive) was applied to the etched surface with a micro-brush, then gently air-dried for 5 seconds and light-cured for 20 seconds (radiant exposure of 15.4 J/cm²) following the manufacturer’s instructions. After the adhesive application, a resin composite (3M^™ Filtek^™ Supreme Ultra Universal Restorative) was added in 1 mm increments. Each layer was light-cured for 20 sec (radiant exposure of 165 J/cm²) with a calibrated LED unit (Cure TC-3, Spring Health Products). Mylar film was placed over the final top/bottom biomaterial surfaces to guarantee homogenous roughness along the sample surface.

After curing, samples were immersed in HBSS for 24 hours before testing for the release of unreacted monomers [54]. The average roughness of the composites was verified as <0.2 μm to prevent the influence of roughness on bacterial adhesion/growth [55] (see Supplemental Information (SI-1). Samples were sterilized by a 1 min immersion in 70% ethanol and air-dried inside a biological safety cabinet during UV light exposure (SI-2).

2.2. Experimental Design

The bonded interfaces of the twin interface samples were challenged to the three primary degradative sources found in the oral cavity, including bacteria, salivary esterases, and cyclic loading (see Figure 1b). To evaluate the effect of each degradative source and their combination on the reduction of the bond strength, a three-factor experimental design was employed [56]. Eight experimental groups were evaluated, including bacteria-only, saliva-only, cyclic loading-only, bacteria+saliva, bacteria+cyclic loading, saliva+cyclic loading, bacteria+saliva+cylic loading (all contributions), and a control (after preparation without degradation). After challenging the interfaces to the different degradation sources, the residual bond strength and the bacterial activity at the bonded interface (degree of invasion/surface coverage, L-lactate concentration, metabolic activity, and number of viable cells) were measured as response variables. Scanning electron microscopy (SEM) was used to observe the bonded interface condition, and optical photothermal infrared (O-PTIR) spectroscopy was used to evaluate changes in the chemistry of the biomaterials at the interface.

2.2.1. Bacteria-Induced Degradation Model

To study the contribution of bacterial attacks on the degradation of the dentin-adhesive-composite interfaces, we optimized a S mutans single-species biofilm model previously employed [57]. This pathogen was selected due to its strong involvement in triggering dental caries development, is highly acidogenic, has aciduric properties, and possesses cell surface proteins associated with solid adherence to biomaterial surfaces [58, 59]. In addition, specific esterases produced by the bacterium have degradative activity toward methacrylate-based resin monomers [14]. To grow biofilms over the bonded interfaces, the S. mutans strain UA159 (isolated from a caries-active child) was plated in a brain-heart infusion (BHI) agar and incubated at 37°C for 24 hours. A single colony was harvested and incubated overnight in fresh BHI stirred at 130 rpm. The solution was diluted in phosphate-buffered saline (PBS) to obtain OD₆₀₀=0.1 (~8×10⁷ CFU) [60] (see SI-3).

Twin interface samples, with the pulpal side facing upward, were submerged in 3 mL of the calibrated liquid bacterial culture. To allow initial bacterial adhesion, beams were incubated at 37°C for 2 hours. Next, the calibrated medium was removed from the wells. The beams were then gently rinsed three times with fresh PBS by repeatedly aspirating and adding with fresh PBS. This process was done to remove any unattached or dead cells loosely adhered to the exterior surfaces of the samples. After rinsing, fresh PBS medium was added to the wells, and the incubation period began. Samples were incubated for 6 days at 37°C in PBS with replacement every 24 h. Bacteria supplementation was conducted by placing the samples in a fresh solution with BHI mixed with 10% of sucrose for 1 min, 5 times a day, representing eating periods with a high cariogenic diet (Figure 1c) [61, 62]. The pH levels of the media were measured every 15 min and kept at pH=7.0 (see SI-4). Overall, this protocol mimicked the oral cavity’s pH fluctuating dynamics, guaranteeing bacterial viability and preventing overfeeding or sustained lower pH levels around the restoration. After the incubation period, all external surfaces of the beam were gently cleaned using a cell scraper to evaluate only the cells that invaded/penetrated the bonded interface.

2.2.2. Salivary Esterases Induced Degradation Model

The enzymatic degradation caused by saliva to adhesives/composites was assessed by submerging samples in a simulated human salivary esterase (SHSE) solution following previous protocols [21, 28]. SHSE simulates the pseudocholinesterase (PCE) and cholesterol esterase (CE) activity from saliva and other body fluids [63]. PCE and CE enzymatic activity hydrolyzes the composite and adhesive’s ester bonds, weakening the interface and inducing microleakage [17]. The SHSE was prepared by mixing cholesterol esterase (Sigma C3766) and pseudocholinesterase (Sigma B4186) in sterile PBS to obtain a solution with similar enzymatic units that human saliva (16 units/mL and 0.01 units/mL CE and PCE activity, respectively) [63, 64]. The media was sterilized through a 0.22 μm filter. SHSE activity levels were verified using nitrophenyl butyrate (pNPB) and butyryl thiocholine (BTC) substrates (see SI-5) [64]. Samples were incubated for 6 days at 37°C in this salivary esterase solution. The SHSE was replaced daily to maintain concentration throughout the incubation. After incubation, samples were washed 3× with PBS for further evaluation.

2.2.3. Bacteria and Salivary Esterases Combinatory Degradation Model

A model to assess the combined degradation caused by bacteria and salivary esterases to the reduction of the bond strength of restorations was developed. Briefly, twin interface samples were incubated at 37°C for 2 hours in 3 mL of a liquid calibrated media (OD₆₀₀=0.1) of S. mutans to allow bacterial adhesion. Next, the beams were gently rinsed three times with fresh PBS to remove unattached/dead cells adhered to the external surfaces of the samples (Section 2.2.1). After rinsing, fresh SHSE was added to the wells (3 mL) (Section 2.2.2), and the incubation period started for 6 days. The salivary esterase-rich media was replaced daily. To feed bacteria, supplementation was performed 5 times/day by submerging the samples in a fresh solution containing 10% sucrose BHI broth for 1 min. After the incubation period, all external surfaces of the beam were gently cleaned using a cell scraper to evaluate only the cells that invaded/penetrated the bonded interface.

2.2.4. Cyclic Loading Induced Degradation Model

To assess the contribution of mastication forces (i.e., cyclic loading) on the degradation of the dentin-adhesive-composite bonded interfaces, twin interface beams were repeatedly loaded using a 4‐point bending configuration (span: 20 mm) with a frequency of 2 Hz and a stress ratio of R=0.1 [33]. This bending configuration ensures uniform stress distribution across bonded interfaces and the stress ratio simulate tension-tension cycles [49, 65]. The stress amplitude (σ_a) was set to 3 MPa, which is only 45% of the endurance limit (6.5 MPa) reported for this sample and media configuration. This magnitude allowed samples to endure 1 million cycles over 6 days without premature failure due to the synergistic effects of cyclic loading and bacterial degradation. These parameters were chosen based on prior studies and further detailed in the SI-6. Samples were submerged in the different solutions for the various testing groups, including PBS for the bacteria- and cyclic loading-only models and SHSE for the saliva and the bacteria+saliva models.

2.2.5. Evaluations of the Bonded Interface: Residual Bond Strength, Bacterial Activity, and Composite Degradation

The condition of the bonded interface was evaluated using different approaches, including residual bond strength measurement, bacterial activity inside the bonded interface, and changes in the chemical composition (degradation) of the biomaterials at the interface. Immediately after incubation, the residual bond strength of the interfaces was measured by subjecting the twin interface samples to failure using 4-point bending configuration (span=20 mm) and quasistatic loading (TA Instruments ElectroForce 5500, 200 N, sensitivity 0.001 N). The cross-head speed was set at 0.06 mm/min [66]. The failure load was recorded, and the flexural strength was calculated [66]. The bond strength was also assessed in an “As Prepared” group, where samples were tested after 24 hours of preparation and immersion in PBS without exposure to degradative challenges.

The activity of bacteria inside the bonded interface was evaluated, including measuring the amount of bacteria invasion (surface coverage), metabolic activity, viable cells (CFU), and L-lactate concentration. Measurements were conducted on the fracture surface, which corresponds to the interior of the bonded interface. To visualize and quantify the bacterial invasion along the interface, confocal microscopy (CLSM) (Olympus Corporation, FV1200) in combination with live/dead staining (LIVE/DEAD BacLight, Thermofisher Scientific L7007) was performed following our previous protocols [57]. The L-lactate concentration at the bonded interface was measured using the L-Lactate Assay Kit (Sigma MAK329) [67], which is the form of lactate typically produced during cellular metabolism of S. mutans [68]. To measure the L-lactate levels at the bonded interface, fractured beams were placed in 200 μL of PBS. To detach the bacterial cells from the fracture surface, samples were sonicated for 1 min at 40 kHz and then vortexed for 30 s. Aliquots of 20 μL were used to prepare the colorimetric reactions following the manufacturer’s instructions. A standard curve was developed for calculating the L-lactate concentration (see SI-7). The metabolic activity of the bacterial cells at the bonded interface was measured using the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay following our previous protocols [57]. Cell viability was assessed by counting the CFU [69].

To characterize the chemical degradation of the biomaterials at the bonded interface, Optical photothermal infrared (O-PTIR) spectroscopy was employed. Selected samples of each group were analyzed using a mIRage Microscope (Photothermal Spectroscopy Corporation). This system utilizes a quantum cascade laser pump source, a 532 nm visible probe laser, and an avalanche photodiode detector to monitor localized photothermal effects. High-speed spectra were collected at 500 cm⁻¹/s over the 1850– 800 cm⁻¹ range, with 2 cm⁻¹ and three averaged scans. The mIRage power settings were 44% for the IR with a pulse rate of 100 kHz. The probe was set to 27%. To visualize compositional differences among samples, hyperspectral image data from the bonded interface was collected in a 25 μm × 50 μm area using a 5 μm resolution. 66 spectra were collected per sample. Images were then analyzed using the mIRage PTIR Studio 4.0 software. Chemical degradation of the biomaterials at the interface was evaluated by measuring changes in the intensity of peaks related to the stretching of the carbonyl group (C=O) of the esters bonds (1720 cm⁻¹) [70] and the overlapping contributions of the asymmetric stretching of the Si-O-Si bonds of the silica fillers and silane coupling agent (1,100 cm⁻¹) [71]. Degradation of dentin was evaluated by measuring changes in the intensity of the υ₁,υ₃PO₄ peak at 1048 cm⁻¹, which is associated with the phosphate stretching and bending of hydroxyapatite, and the peak at 1660 cm⁻¹ related to the movement of the amide I present in collagen [72–74]. Color scale bars on the hyperspectral image reflect relative O-PTIR intensity, with dark blue as the lowest and red as the highest concentration. Selected samples were observed using scanning electron microscopy (EI Quanta 450 EFG) in secondary electron imaging mode with different magnifications at an accelerating voltage of 10 kV.

2.3. Statistical Analysis

All data is presented as box plots. The box corresponds to the standard error of the mean, and the whiskers to the standard deviation. Generalized linear models (GLM) were used to study the impact of the degradative factors (i.e., bacteria, saliva, cyclic loading) and their interactions on the dependent (response) variables [75]. The model included one three-way interaction (i.e., bacteria+ saliva+cyclic loading) and three two-way interactions (i.e., bacteria+saliva, bacteria+cyclic loading, and saliva+cyclic loading) as follows:

$$Y={\beta }_0+\left({\beta }_{1}A\right)+\left({\beta }_{2}B\right)+\left({\beta }_{3}C\right)+\left({\beta }_{4}AB\right)+\left({\beta }_{5}AC\right)+\left({\beta }_{6}BC\right)+\left({\beta }_{7}ABC\right)+\epsilon$$ Eqn 1

where, Y corresponds to the response variables including A: Bacteria, B: Saliva, C: Cyclic loading. β_x corresponds to the linear regression coefficient associated to each factor and ε is the standard error. The calculated regression coefficients, depending on their sign (positive/negative) and effect on the degradation of the bonded interface was classified as detrimental or beneficial to the bond of the interface. Multi-factor analysis of variance was used to determine the significance of effects and their interactions (significance, p <0.05) (SI-8). All statistical analyses were completed using STATGRAPHICS Centurion XVII.

3. RESULTS

Bacterial activity at the bonded interface:

The invasion of bacteria inside the bonded interfaces was assessed by visualizing the live/dead cells and quantifying the number of viable microbes and their metabolism (Figure 2). As expected, the saliva and PBS control groups showed no bacteria at the bonded interface. Overall, cyclic loading of the bonded interfaces facilitated the invasion, the number of bacteria, and the production of lactic acid along the adhesive/dentin margin (Figure 2a). In the cyclic loaded groups, bacteria were preferentially located at the center of the fracture surface in the bacteria-only groups. On the other hand, when salivary esterase was added (bacteria+saliva), it resulted in the dispersal of microbes primarily along the outer edges of the sample. This suggests a distinct impact on the position and distribution of bacteria along the restoration margin. In the static groups, there was no significant evidence of bacteria adhering to the interior of the bonded interface. Moreover, the surface area covered by microbes after invasion is presented in Figure 2b. The highest coverage (28% of the fracture surface) was observed for bonded interfaces subjected to the combination of all degradative challenges (bacteria, saliva, cyclic loading). The lowest bacteria coverage was observed for the static interfaces challenged by saliva and bacterial attacks. To determine interactions between the degradative factors, the statistical GLM model confirmed that the combination of bacteria+cyclic loading significantly contributes to the increased quantity of microbes invading and covering the restorative margin (p<0.001) (Figure 2c). In addition, the interaction between all the degradative sources also synergically helps to increase the bacterial surface coverage over the bonded interface. In other words, cyclic loading significantly exacerbates the quantity of bacteria accessing and colonizing bonded interfaces regardless of the presence of salivary enzymes.

Figure 2. Bacteria activity at the bonded interfaces.

(a) Fracture surface CLSM images of selected samples subjected to different challenges, including phosphate-buffered saline (PBS), saliva, bacteria, cyclic loading, and their interactions. Samples were stained with Syto-9 (green) and propidium iodide (red) to indicate live and dead bacteria, respectively. (b) Quantification of the surface area occupied/covered by bacteria. N= 3 samples for each group. (c) Coefficient estimates from GLM statistical analysis for bacteria surface coverage response. (d) L-lactate quantity along the bonded interface for all tested groups. Data is expressed as millimolar (mM) of L-lactate. N = 6 samples for each group. (e) Coefficient estimates from GLM for L-lactate concentration responses. (f) Metabolic activity of S. mutans cells at the bonded interface for all evaluated conditions. N = 4 samples for each group. (g) Coefficient estimates from GLM for metabolic activity responses at the bonded interface. (h) Number of viable bacteria (CFU) present at the bonded interface. N= 4 samples for each group. (i) Coefficient estimates from GLM for the number of viable cells at the bonded interface. All GLM coefficients were classified as detrimental (red bars) or beneficial (green bars) to the degradation of the bonded interface. Significance of the factors and interactions were assessed by multifactor ANOVA (<0.05).

Furthermore, the concentration of L-lactate along the bonded interface was also increased by applying cyclic loading compared to static conditions (Figure 2d). Looking at the potential interactions between all the degradative sources, cyclic loading significantly aggravates the lactic acid concentration along the bonded interfaces regardless of the presence of salivary enzymes (Figure 2e). Lactic acid is known to amplify biofilm formation, virulence, dental tissue demineralization, and breakdown of resins, therefore exacerbating the destruction of the bonded interface [59, 76, 77]. Bonded interfaces held static (no application of cyclic mechanical loading) and did not report any lactic acid at the interface. Moreover, a similar response was observed for the metabolic activity of the microbes covering the bonded interface (Figure 2f). Bacteria, cyclic loading, and their combination significantly contribute to increased bacteria metabolism (Figure 2g). Looking into the viable microbes at the bonded interface, cyclic loading and saliva contributed to the significant increase in bacteria compared to static (Figure 2h). These results indicate that saliva helped increase the number of microbes but under a lower metabolism, invading the restoration margin.

Bond strength after challenged by multiple degradative sources:

The flexural bond strength of dentin-adhesive-composites bonds before and after the degradation challenges is presented in Figure 3a. As expected, the highest bond strength (27 MPa) was observed for interfaces not subjected to any degradation challenge (after sample preparation). The lowest bond strength was observed for interfaces subjected to simultaneous cyclic loading and bacterial attacks, with a 73% reduction compared to the no-degraded group (residual strength 7 MPa). The combinatorial attacks from saliva, bacteria, and cyclic loading rendered a bond strength reduction of 56% compared to the no-degraded samples. On average, the saliva and PBS (control media) groups, regardless of the mechanical loading condition, showed a reduction of the bond strength of 30% after 6 days. Under static conditions, the lowest strength was measured for bonds incubated in the presence of bacteria+saliva (14 MPa), corresponding to a 46% reduction compared to the no-degraded interfaces. Looking at the interactions between all the degradative factors, the GLM models showed that bacteria, cyclic loading, and their combined interaction negatively affect the bond strength of the composites (p< 0.001) (Figure 3b). In other words, bacteria and cyclic loading work together to amplify the bond strength reduction significantly. A positive effect on the bond strength (no reduction) was found when all degradative challenges work simultaneously, suggesting that saliva retard the rate of degradation.

Figure 3. Bond strength responses and bonded interface conditions after challenged by multiple degradative sources.

(a) Comparison of the bond strength for interfaces subjected to different challenges, including phosphate-buffered saline (PBS), saliva, bacteria, cyclic loading, and their interactions. The duration of the challenges was 6-days (or ~1 M cycles for cyclically loaded cases). N= 6 samples for each group. (b) Coefficient estimates from GLM for residual bond strength responses. Coefficients were classified as detrimental (red bars) or beneficial (green bars) to the degradation of the bonded interface. Significance of the factors and interactions were assessed by multifactor ANOVA (<0.05). (c) Representative micrographs of the bonded interfaces after different degradative challenges, including bacterial and bacteria/saliva under cyclic loading conditions and phosphate-buffered saline (PBS) and bacteria under static conditions. “C” indicates the composite material, “A” represents the adhesive interface, and “D” corresponds to dentin. The yellow and orange arrows on the micrographs correspond to the adhesive and composite material damage, respectively.

The bonded interfaces were observed using SEM (Figure 3c). Overall, interfaces subjected to bacteria and bacteria+saliva attacks showed poor integrity of bonded biomaterials and significant damages without clearly distinguishing the bonded layers (Figures. 3c.1,2,4). Uniform adhesive/composite layers without defects or gaps, in addition to even distribution of fillers in the composite, were observed in the group defined as static PBS for 6 days (Figure 3c.3). Overall, bonded interfaces attacked by bacteria showed significant destruction of the adhesive layer regardless of the loading condition (yellow arrows) (Figures 3c.1 and c.4). Considerable damage in the composite material was observed for all samples regardless of the degradative media or loading condition (see orange arrows). The greatest damage in the composite material was observed for samples under bacteria/cyclic loading conditions, where detachment of filler particles and polymer matrix were observed (Figure 3c.1). Overall, the extensive interfacial damage is linked to the application of cyclic loading. Bonded interfaces subjected to all the degradative sources showed less adhesive damage, which correlates with the results observed with the bond strength (Figure 3c.2).

Changes in the chemical composition of the biomaterials at the bonded interface:

The chemical degradation of the bonded biomaterials (dentin-adhesive-composite) was evaluated using O-PTIR imaging (Figure 4). Color variations in the peak area/intensity are related to concentration variations, chemical transformation, and structural changes [78]. We selected four specific peaks for analysis since they indicate the degradation of these biomaterials. First, changes in the ester groups (stretching of the carbonyl group (C=O)) of the Bis-GMA (bisphenol A-glycidyl methacrylate) and the urethane dimethacrylate (UDMA) of the resin composite were observed at 1720 cm⁻¹ (Figure 4a) [79]. The highest number of ester bonds (i.e., redder image) was observed in the after-preparation (no degradation) composite (Figure 4a.1). Overall, all degradative challenges reduced the number of ester bonds (decreased peak intensity towards blue/purple) compared to the PBS-reference (Figure 4a.4). A reduction in the number of ester bonds suggests hydrolysis of the methacrylate-based polymers due to salivary and bacterial esterases [80]. The lowest number of ester bonds (darker blue/purple) was observed in the static composites attacked by bacteria (Figure 4a.6), especially in the region closer to the adhesive layer (white arrow). Composites submerged in PBS showed the highest number of ester bonds, indicating the least degraded condition (Figure 4a.4).

Figure 4. Chemical changes of the biomaterials at the bonded interfaces after challenged by multiple degradative sources, including phosphate-buffered saline (PBS), bacteria, saliva, cyclic loading, and their interactions.

Hyperspectral images representing changes in the intensity of peaks associated with the (a) ester (C=O), (b) Si-O-Si, (c) ν₁,ν₃ PO₄, and (d) Amide I bonds. Degraded samples were compared with “As prepared” (no-degraded) samples. Color scale bars on the hyperspectral image correspond to the relative O-PTIR intensity, with dark blue as the lowest and red as the highest peak intensity.

Chemical changes associated with the silica fillers and silane coupling agent of the resin composite and adhesives (Figure 4b) can be evaluated at the 1100 cm⁻¹ band, which is related to the overlapping contributions of the asymmetric stretching of the Si-O-Si bonds [71]. Highest and similar peak intensities were observed for the no-degraded (as prepared) and bonded interfaces subjected to salivary enzymes (Figures 4b.1 and b.5). These results suggest that the presence of salivary esterases is not cleaving the Si-O-Si bonds. Lower peak intensities were observed in regions closer to the adhesive material (green arrows). Significant degradation (decreased peak intensity) was observed in composites attacked by bacteria regardless of the loading condition and presence of saliva (Figures 4b.2, b.3, b.6, and b.7). Incubation of samples in PBS for 6 days resulted in decreased Si-O-Si bonds, which are also susceptible to hydrolytic attacks [81]. Chemical changes associated with dentin (i.e., demineralization) can be observed at 1048 cm⁻¹ in relation to the υ₁,υ₃PO₄ phosphate stretching and bending of hydroxyapatite [74] (Figure 4c). Chemical changes associated with collagen can be observed at 1664 cm⁻¹ [72–74] (Figure 4d). A similar degradation pattern was observed for both tissues. Compared to the no-degraded dentin, all degradative challenges caused a reduction in the peak intensity of the mineral and collagen tissue (more blue images). For both tissues, the higher degradation was associated with bacteria (Figures 4c.2–6 and 4d.2–6). The acid produced by bacteria induces demineralization [82], and disrupts the collagen structure by denaturing or hydrolyzing its peptide bonds [83]. Saliva appears to have different effects on both tissues. No significant changes in the mineral peak intensity were observed in samples placed on saliva for 6 days (Figure 4c.4). However, a decrease in the peak intensity of the band related to collagen was observed (Figure 4d.4). Selected spectra of the bonded interface (dentin-adhesive-composite) and the individual biomaterials are presented in SI-9.

4. DISCUSSION

Approaches that realistically predict composite restorations’ degradation and clinical performance are lacking [12, 35]. Current models are primarily limited to evaluating one or two degradative factors separately or in sequence but rarely in combination [17, 23, 31, 36, 84–86]. In addition, it remains unclear which degradative factor holds the greatest significance and whether there is a synergistic/antagonistic effect when combined. In this study, we developed a lab approach that assessed simultaneous degradative factors such as bacteria, salivary esterases, and cyclic loading on the bond of composite restorations. When working in concert, all these factors affect differently against the dentin-adhesive-composite interface. Three major findings are reported in this study: First, each degradative factor, when assessed individually, had a detrimental impact on the bond strength of the restoration (Figure 3). However, when assessed in combination, both negative and positive synergies contributed towards the interface degradation and bond strength (Figures 2–4). For example, bacteria and cyclic loading collaborate to intensify restoration damage and reduce bond strength, whereas salivary esterases were found to decelerate bacterial metabolism, mitigating bond degradation. Second, salivary esterases have both beneficial/detrimental effects on the bonded interface degradation. When combined with cyclic loading and bacteria, saliva assists in bacterial interfacial penetration, increasing the number of viable cells at the interface but with reduced metabolism (Figure 2) (SI-10). This decrease in metabolic activity acted as a “protection”, reducing the damage in the chemical structure of the composites and the matrix-filler interface (e.g., Si-O-Si) (Figure 4), which prevented additional bond strength reduction. Third, ranking the individual degradative sources in reducing bond strength revealed that bacterial attack was the main contributor to interface degradation, followed by cyclic loading and salivary esterases. Our model captured the intricate interplay of degradative sources impacting dental restoration’s durability. This model is a significant advancement, integrating all degradation challenges simultaneously and providing a realistic representation of the oral environment affecting dental restorations. The inclusion of cyclic loading represents a notable leap forward, highlighting the profound influence of mechanical stress in concert with biological factors on the bond degradation process.

Cyclic loading and bacteria synergically degraded bonded interfaces:

Our data showed a powerful synergy between cyclic loading and bacterial attacks that is detrimental to the bond strength and integrity of the bonded interface. Results indicated that bacterial attacks on bonded interfaces without mastication forces (static) caused microbial penetration covering less than 10% of the margin, an acid concentration of 0.1 mM, and moderate interfacial damage, leading to a 46% reduction in bond strength compared to non-degraded samples. However, when cyclic loading was factored in, bacteria coverage increased to 28% of the interface, with an acid concentration increase of 1 mM, resulting in a 73% reduction in bond strength. The combination of cyclic loading and bacterial attacks rendered the bonded interface’s lowest bond strength and highest damage along the tested groups. These observations agree with previous work [13, 24, 87]. The application of cyclic loading represents daily masticatory activities. Repeated stresses applied to dental materials have been relevant to studying fatigue failure [33]. In the context of microbe-biomaterial interactions, cyclic mechanical stresses were found to exacerbate fungal infections by increasing the virulence of Candida albicans [88] and modifying the subgingival microbiome composition [89]. In the context of bonded interfaces, repetitive loading facilitates bacterial penetration by creating a hydraulic pumping effect [24]. The continuous opening and closing of the marginal gap may introduce fresh nutrients and bacterial cells into the restoration margin. Additionally, the hydrolytic degradation of the adhesive and the demineralization of dentin (Figure 4) leads to increased material roughness (SI-11) and the formation of gaps, creating extra spaces for bacterial penetration [4, 90]. More bacteria at the interface will produce more acid and induce additional degradation of biomaterials, amplifying bond destruction. These findings also correlate with clinical observations. Patients with bruxism are known to have short durability of composite restorations [8]. Results from this work emphasize the need to consider repetitive mechanical forces alongside microbial and enzymatic challenges when evaluating the bond of a restoration, which will represent real-world conditions.

Salivary esterases and their effect on the degradation of bonded interfaces:

Results from this study showed that salivary esterases have both beneficial/detrimental effects on the interface degradation and bond strength of composite restorations. It is known that salivary esterases (CE and PCE) are responsible for the hydrolysis of carboxylic esters and have a degradative activity towards the methacrylate-based organic matrix of dental composite resins [21, 91, 92]. In addition, the by-products produced after hydrolysis modulate the growth of oral bacteria [93]. Our results agree with these observations (Figures 3–4). The effect of salivary enzymes on bond degradation was reflected by a decrease in the bond strength (~30%) regardless the application of cyclic loading and changes in the chemical composition of composites from hydrolysis (Figure 4). However, additional effects were triggered when salivary enzymes and bacteria were together. Saliva notably facilitated bacterial penetration at the interface (Figure 2a.2), evidenced by the increase in bacteria (Figure 2b). Bacteria preferentially were positioned along the edges of the restoration rather than at the center or interior. This suggests that the hydrolytic degradation may have created additional spaces within the adhesive, providing an ideal environment for bacterial colonization. However, these additional microbes reported lower metabolism in the presence of salivary enzymes (Figure 2f) (SI-10). The Bis-GMA degradation products such as 2,2-Bis [4(2,3-hydroxypropoxy) phenyl] propane (BisHPPP) and methacrylic acid (MA) have been reported to inhibit the growth of S. mutans [93, 94]. In contrast, the biodegradation product of triethylene glycol dimethacrylate (TEGDMA), triethylene glycol (TEG), promotes the growth of S. mutans [95]. This reduction in the bacterial metabolic activity was shown to be beneficial in maintaining the bond strength of the bonded interface (Figure 3b), reducing the damage and biomaterials integrity (Figure 3c.2), and protecting the chemical structure of the composite and the matrix-filler interface (i.e., Si-O-Si) (Figure 4) compared to bacterial attacks only. Saliva is essential in deterring the degradation of the bond of restorations. Patients with decreased salivary secretion showed faster failure rates of direct restorations [96]. Despite the fact that saliva can induce hydrolytic degradation and reduce to some extent the bond strength, its net effect is positive, which was unveiled with this model.

Chemical hydrolysis as a baseline to measure the degradation of bonded interfaces:

Hydrolysis caused by water is known to reduce the mechanical properties of dental composites [19]. Results from this work support these observations in the context of bonding. Bonded interfaces challenged by a neutral environment (PBS) exhibited a 24% reduction in bond strength compared to the no-degraded samples (Figure 3). This reduction is explained by the changes observed in the chemical composition due to chemical hydrolysis (Figure 4). Water is known to facilitate the chemical degradation of adhesives and polymer matrix [97]. Hydrolysis caused by water involves its diffusion through the polymer matrix and matrix/filler interface, leading to plasticization, softening, swelling, hydrolysis of ester bonds (Figure 4a.4), and cleavage of the siloxane bonds in the silane coupling agent (Figures 4b.4) [98, 99]. Hydrolytic degradation in dentin’s collagen I breaks collagen bonds within the hybrid layer [100]. Recent studies have shown that only chemical hydrolysis is related to reduced flexural strength, diametral tensile strength, and hardness [101, 102]. Degraded adhesives and composites can significantly impact the bond strength due to this reduction of native properties. Since water is ubiquitous in the mouth, the longevity of dental materials and bonded interfaces should be evaluated under wet conditions and in water-saturated interfaces.

Despite the potential impact of our findings, this study has limitations. First, our experiments were performed in a S. mutans single-species model. We selected this specific bacteria strain since it is the major pathogen related to dental caries [58, 103]. Future studies on our model should assess the degradation of bonded interfaces using more complex and heterogeneous inoculum to understand better possible microbial interactions and microbiome changes derived from biomaterials at the interface. Expanding our model to encompass a wider range of microbial species and environmental conditions will increase its clinical relevance. In these studies, a more detailed investigation into the role of nanoleakage is required. We plan to incorporate specific assessments of nanoleakage, such as silver nitrate penetration or confocal laser scanning microscopy, to understand its impact on bond durability better. Second, our study evaluated the effect of salivary esterases such as CE and PCE on the degradation of the interfaces. Future studies should include additional salivary enzymes that can degrade dental materials or interact with bacterial biofilms. For example, salivary albumin can cleave ester linkages and catalyze the degradation of resin-based dental materials [104], and salivary peroxidase participates in the antimicrobial protection of oral surfaces [105]. In addition, our model did not include collagenase degradation sources due to the limited contribution to bond degradation over time [106, 107]. However, these enzymes might affect bacterial metabolism and influence bond degradation, which warrants further evaluation. Third, cariogenic bacteria, such as S. mutans, produce esterases that can also induce degradation of dental resins. Additional studies should focus on discriminating salivary and bacterial enzymes attacking the composites and adhesives, providing a better association between degradative sources and their consequent damage. Furthermore, thermal cycling or water aging are important factors in the degradation of dental restorations and the microbe’s metabolism [108]. Future studies might consider adding this layer of complexity into the model to close the gap between lab models and clinical scenarios. Fourth, the long-term stability is an important factor to consider for a more comprehensive understanding of the degradation of the bond after being subjected to the proposed synergistic challenges. Future studies should include extended experimental durations including 3, 6, and 12 months. Finally, in our study, we used the novel O-PTIR spectroscopic technique to assess the degradation of the bonded interfaces. In the hyperspectral images, we could see the differences between the chemical composition of the composite, adhesive, and dentin. However, specific changes in the hybrid layer were not identified. Future O-PTIR measurements should be performed using a submicron resolution that allows us to identify more specific changes along the degraded interface.

5. CONCLUSIONS

In this study, we developed a new approach to evaluate composite restorations’ degradation and bond strength, which was challenged by simultaneous insults from bacteria, salivary esterases, and cyclic loading. For the first time, results showed that bacteria is the number one factor promoting the bonded interface’s degradation and the bond strength reduction, followed by cyclic loading and salivary esterases. Each degradative factor, when assessed individually, had a detrimental impact on the bond strength of the restoration. However, when assessed in combination, both negative and positive synergies contributed towards the interface degradation and strength. For example, cyclic loading enhanced the bacteria invasion at the bonded interface, increasing the production of lactic acid, demineralization of dentin, and degradation at the composite/filler interface, rending the highest reduction in the bond strength (73%) compared to no-degraded interfaces. Individually, salivary enzymes reduced the bond strength of the interface by producing changes in the chemical composition of the adhesives. However, in combination with cyclic loading and bacteria, saliva aids the invasion of microbes to the bonded interface but affects the bacteria metabolism, mitigating aggressive bond strength reduction. Findings of this study are crucial for identifying the pathways through which degradation of dental materials occurs, offering a new perspective on the breakdown processes that are instrumental in developing more resilient dental composites.

Highlights.

• An approach to evaluate the bond degradation of composite restorations was created

• Simultaneous biological, chemical and mechanical challenges degraded bonded interfaces

• Bacterial attacks ranked first in degrading bonded interfaces followed by mastication

• Cyclic loading (mastication) and bacteria synergically degraded bonded interfaces

• Salivary esterases have both beneficial/detrimental effects on the interface degradation

DATA AVAILABILITY

The datasets supporting this study’s findings are available from the corresponding author upon reasonable request.