Dentin collagen denaturation status assessed by collagen hybridizing peptide and its effect on bio-stabilization of proanthocyanidins

Abstract

Objectives:

To assess dentin collagen denaturation from phosphoric acid and enzyme treatments using collagen hybridizing peptide (CHP) and to investigate the effect of collagen denaturation on bio-stabilization promoted by proanthocyanidins (PA).

Methods:

Human molars were sectioned into 7-μm-thick dentin films, demineralized, and assigned to six groups: control with/without PA modification, H₃PO₄-treated collagen with/without PA modification, enzyme-treated collagen with/without PA modification. PA modification involved immersing collagen films in 0.65% PA for 30 seconds. H₃PO₄ and enzyme treatments were used to experimentally induce collagen denaturation, which was quantitated by fluorescence intensity (FI) from the fluorescently-conjugated-CHP (F-CHP) staining (n=4). FTIR was used to characterize collagen structures. All groups were subject to collagenase digestion to test the bio-stabilization effect of PA on denatured collagen using weight loss analysis and hydroxyproline assay (n=6). Data were analyzed using two-factor ANOVA and Games-Howell post hoc tests (α=0.05).

Results:

FTIR showed collagen secondary structural changes after denaturation treatments and confirmed the incorporation and cross-linking of PA in control and treated collagen. F-CHP staining indicated high-degree, medium-degree, and low-degree collagen denaturation from H₃PO₄-treatment (FI=83.22), enzyme-treatment (FI=36.54), and control (FI=6.01) respectively. PA modification significantly reduced the weight loss and hydroxyproline release of all groups from digestion (p<0.0001), with the results correlated with FI values at r=0.96–0.98.

Significance:

A molecular method CHP is introduced as a sensitive technique to quantitate dentin collagen denaturation for the first time. PA modification is shown to effectively stabilize denatured collagen against collagenase digestion, with the stabilization effect negatively associated with the collagen denaturation degree.

1. Introduction

Modern composite restorations, while being more esthetically pleasing than amalgam, have the disadvantage of a significantly shorter lifespan [1]. One major cause of restoration failure is due to the breakdown of the dentin-adhesive hybrid layer, which is formed during the etching and adhesive resin penetration steps of the procedure [2]. Specifically, acid etching removes mineral from the top surface of dentin substrate, leaving behind a protein network mostly composed of type I collagen. Due to incomplete infiltration of the adhesive resin into the demineralized dentin layer, the exposed collagen fibrils in the hybrid layer are subject to enzymatic degradation. Endogenous matrix metalloproteinases (MMPs) within dentin are thought to play a major role in the degradation of exposed collagen within the hybrid layer [3]. They are activated by the low pH brought on by caries or the acid etching step of composite restorations [4]. Prevention of collagen degradation in the hybrid layer is of great clinical importance and could be accomplished by inhibiting the endogenous MMPs or by stabilizing the collagen via cross-linking in the hybrid layer [5, 6]. Proanthocyanidins (PA) from grape seed extract are a natural cross-linker known for their dual functionality of collagen cross-linking and MMP inhibition [7]. Application of PA to demineralized dentin has been shown to increase its stiffness, ultimate tensile strength, elastic modulus and bond strength [8–11]. PA has also demonstrated the ability to prevent demineralized dentin from enzymatic degradation [12, 13]. The quick cross-linking time and low toxicity of PA further promote its potential use in clinical restorative practice [7, 14].

In dental practice, dentists perform restorations mostly on diseased teeth, such as carious or sclerotic teeth. The disease pathologies modify the structure and composition of the teeth, resulting in the need for restoration. The process of dentin caries is a complex event involving demineralization and matrix degradation [15]. Carious dentin consists of two distinct layers: the outer caries-infected dentin layer (CID) and the inner caries-affected dentin layer (CAD). The CID is bacterially infected, highly demineralized, physiologically un-remineralizable and contains irreversibly denatured collagen with a virtual disappearance of cross-linkages. On the other hand, the CAD is partially demineralized, physiologically remineralizable, and consists of matrix alterations such as decreased cross links and altered secondary structure of collagen [16, 17]. Traditional caries management recommends removal of CID to prevent further cariogenic activity and provide a mineralized dentin substrate for restoration. Recent research on caries especially deep caries management recommends less invasive strategies such as partial or selective removal of carious tissues to reduce the risk of pulp exposure [18]. Either strategies could result in bonding surfaces consisting of non-ideal dentin substrate with structurally altered collagens that can negatively affect adhesive bonding [19, 20]. Bonding to dentin can also be affected by the sclerosis of the substrate, which refers to wasting diseases of dental hard tissue in the cemento-enamel junction region [21]. The dentin composition on the surface of non-carious sclerotic lesions also contains denatured collagens which together with other factors can negatively affect adhesive bonding [22–25]. It is imperative to develop a simple, in situ, and quantitative method for the characterization and evaluation of collagen denaturation.

Most research work that evaluated the effects of PA on dentin collagen have utilized sound dentin samples in the experiments [26–28]. Only a small number of studies investigated the effects of PA on the mechanical properties and adhesive bonding properties of non-ideal dentin substrates such as caries-affected and eroded dentin. One study found that PA treatment of etched dentin prior to bonding significantly enhanced the bond strengths of both sound and CAD dentin [29]. Another study showed that PA treatment of CAD significantly increased their elastic modulus [30]. A recent study investigated the effect of PA on the bonding properties of soft drink and citric acid eroded dentin and reported that the addition of PA to phosphoric acid etchant improved both short-term and long-term adhesive performance at the resin and eroded-dentin interface when compared with conventional phosphoric acid etchant [31]. However, the collagen status in those non-ideal dentin substrates was unknown. The effect of PA on well-characterized denatured dentin collagen has not been investigated before. The non-ideal dentin substrates encountered in clinical settings are highly variable due to different caries activity (arrested or active), type of teeth (permanent vs. primary) and site specificity of the lesions, which present challenges for well-controlled quantitative evaluation. Two experimental protocols were developed (phosphoric acid and enzyme treatments) to induce denaturation in sound dentin collagen for quantitative investigation.

The objectives of the current study were to characterize the collagen denaturation status from these two treatments using a molecular technique called Collagen Hybridizing Peptide (CHP) and to investigate the effect of collagen denaturation on bio-stabilization promoted by PA. CHP is a synthetic single strand collagen mimetic peptide (CMP) (sequence: (GPO)x, x=6–10, O: hydroxyproline) that can specifically bind to unfolded collagen chains of denatured collagen through hydrogen bonding [32]. The binding interaction originates from the unique triple helical structure of the collagen molecules and the inherently strong triple helical folding propensity of the CHP. The conjugation of a fluorescein to the CHP allows for fluorescence detection with a microscope. The null hypotheses tested were 1) that dentin collagens treated by the two experimental protocols (phosphoric acid and enzyme treatments) would show the same CHP fluorescence intensity as control, and 2) that PA would not stabilize experimentally denatured dentin collagen from collagenase digestion.

2. Materials and Methods

All chemicals used in the current study were purchased from Sigma-Aldrich (St. Louis, MO, USA) unless otherwise noted. The grape seed extract was donated by Mega Natural (Madera, CA, USA). 0.96% phosphate buffered saline (PBS, PH=7.4) was prepared using Dulbecco’s Phosphate Buffered Saline powder packet (P3813) and 0.002% sodium azide was added to prevent bacterial growth. TESCA buffer was prepared by dissolving 5.75 g of TES (N-tris (hydroxymethyl) methyl-2-aminoethanesulfonic acid, Lot 103181, Fisher Scientific, Pittsburgh, PA, USA), and 26.5 mg of CaCl₂ (Lot 876772, Fisher Scientific) in 500 mL of distilled water, and the PH was adjusted to 7.4 using NaOH. Bacterial Collagenase (from Clostridium histolyticum - type I, ≥125 CDU/mg solid) was dissolved in TESCA solution for collagen digestion. The collagen hybridizing peptide (CHP) was purchased from 3Helix (Salt Lake City, UT, USA). The overall experimental design and procedures are illustrated in Figure 1.

Figure 1.

The flow chat of experimental design and procedures

2.1. Dentin film preparation and demineralization

Six non-carious human third molars were collected with no associated patient identifiers, collection protocol determined as not human subject research (NHSR 12–50) as per the University Adult Heath Sciences Institutional Review Board. Teeth were stored in 0.96% PBS containing 0.002% sodium azide at 4°C. The occlusal enamel and roots were removed from the teeth and the remaining of the teeth were cut into dentin blocks approximately 6 × 5 × 2 mm³ using a water-cooled diamond saw (Buehler, Lake Bluff, IL, USA). A light microscope (Nikon Eclipse LV150NL, Visual Dynamix, Chesterfield, MO, USA) was used to ensure that no enamel remained. Any remaining enamel was polished away using 600-grit silicon carbide paper. Each dentin block was further sectioned into 7-μm thick dentin films using a microtome (Leica Biosystems, Buffalo Grove, IL, USA). The dentin films were demineralized by 10% phosphoric acid for 30 minutes and then rinsed with deionized water for 30 minutes, changing water every 10 minutes. After rinsing, the collagen films were stored in 0.96% PBS at 4°C.

2.2. Treatments of dentin collagen

The dentin collagen films were randomly assigned into three groups (control, H₃PO_4-treatment, and enzyme-treatment) with 140 films each. The collagen films in the control group were immersed in 0.96% PBS. The films in the H₃PO₄-treatment group were treated with 35% phosphoric acid in aqueous solution for 3 hours at 37 °C. The films in the enzyme-treatment group were subject to 0.1 M lactic acid (pH 4.0 in PBS) for 24 hours followed by 30 μg/ml collagenase treatment for 6 hours at 37 °C. The samples were placed on an orbital shaker inside the incubator during the incubation at 37 °C. The films were rinsed with deionized water for 30 minutes and stored in 0.96% PBS at 4°C.

2.3. Biomodification of dentin collagen with PA

Using the ultra-thin dentin film model (≤ 10 μm), our group can quantitatively evaluate the stabilization effects of various natural cross-linkers on dentin collagen with relatively low cross-linker concentrations (0.4%–2%) and clinically relevant treatment time (30 s) [33]. For the current study, all three groups were divided into two subgroups of 70 collagen films each, resulting in a total of six test groups (Fig. 1). One subgroup was modified with 0.65% PA aqueous solution for 30 seconds and the other subgroup was not modified with PA. Specifically, the collagen films were spread flat on a clean glass substrate with the aid of a small paintbrush or a dental explorer. Then one drop (~15 μL) of PA solution was pipetted onto each film to cover the entire surface and manually agitated throughout the 30 seconds modification time. Afterwards, the films were rinsed with deionized water for 10 minutes three times and placed on labeled coverslips (Fisher Scientific, Pittsburgh, PA, USA). The films were subsequently dried in a vacuum desiccator for 48 hours

2.4. Fourier transformed infrared (FTIR) spectroscopy

FTIR was used to elucidate biochemical and secondary-structural information of the dentin collagen sample. Two sampling modes of FTIR were used in the current study – the Attenuated Total Reflection (ATR) mode and the transmission mode. A small volume of dry PA powder was placed on the top of the ATR diamond-ZnSe crystal to get the spectrum of PA. FTIR spectra of dentin collagen films were acquired in transmission mode using a Perkin Elmer FTIR Spectrum Spotlight system (Spectrum one, Spotlight 300, Perkin Elmer, Waltham, MA, USA). Specifically, dried collagen films were placed on BaF₂ polished discs (Reflex Analytical Corporation, Ridgewood, NJ, USA) and FTIR scans were carried out in the range of 1800 – 950 cm⁻¹ at a resolution of 4 cm⁻¹. Each spectrum was the average of 16 co-adding spectra. The original FTIR spectra were pre-processed by baseline adjustment and Amide I band normalization using the Perkin Elmer Spectrum software. The average spectrum was calculated from three films in each test group (n=3). Comparisons were made among the average spectra of control, H₃PO₄-treated and enzyme-treated samples with or without PA biomodification. The average spectra of the control and experimentally denatured dentin collagen without PA modification was subtracted from the corresponding average spectra of PA-modified collagen to observe the incorporation and cross-linking of PA in dentin collagen qualitatively.

2.5. Collagen hybridizing peptide (CHP) assay

CHP assay was used to examine the denaturation status of dentin collagen. Specifically, 60 μg collagen hybridizing peptide, 5-FAM conjugate (F-CHP) purified lyophilized powder was dissolved in 400 μL pure water to make 50 μM stock CHP solution and stored at 4 °C until use. At the time of assay, the 50 μM stock solution was further diluted to 5 μM staining solution. The staining solution was heated in a digital dry bath (Thermo Scientific, USA) at 80 °C for 10 mins to disassociate the recombined CHP molecules in the solution, followed by an immediate ice-water bath for 10 s to quench the solution to room temperature. Then 20 μL CHP staining solution was quickly pipetted onto each dentin collagen film to cover the entire surface. Afterward, the films were sealed in a container and incubated with the staining solution at 4 °C for overnight.

The stained films were then rinsed with water for 5 minutes three times at room temperature. After air drying, the films were mounted on a glass slide. A Keyence all-in-one Fluorescence microscope BZ-X800 (Osaka, Japan) equipped with a FITC filter cube was used to take images of the films. Images were taken with a 20 × PlanFluor objective at a fluorescence exposure time of 0.1 s. The ImageJ software (1.52a, National Institutes of Health, USA) was used to measure the fluorescence intensity (FI) of the images. Four films from each test group were used for the analysis (n=4).

2.6. Weight loss analysis

Weight loss analysis was used to measure the weight change of collagen films before and after collagenase digestion. 60 films from each test group were randomly divided further into six sample groups of 10 films each (n=6). The net weight of 10 films in each sample group was weighed using an analytic balance (d=0.01 mg, Mettler Toledo AG285, Zurich, Switzerland). The weight of the control group (C) and the two treated groups (P & E) were compared to see if experimental treatments resulted in any weight change for the collagen films.

After weighing, films were immersed in 300 μL of 0.1% bacterial collagenase solution in TESCA buffer for 1 hour at 37 °C. Films were then rinsed with distilled water for 30 mins and dried in a vacuum desiccator for 48 hours. Dried films were weighed and the percentage weight loss from collagenase digestion was determined by the dry weight change before (W₀) and after (W₁) collagenase digestion using the following equation: W_L%=(W₀-W₁)/W₀×100%.

2.7. Hydroxyproline assay

Hydroxyproline assay was used to measure the amount of hydroxyproline released from dentin collagen samples during collagenase digestion. Briefly, the digestion solution from weight loss analysis was hydrolyzed with 6M HCl at 110 °C for 24 h, followed by vacuum-drying. The dried hydrolysates (free hydroxyproline and other amino acids) were dissolved with distilled water and pre-treated by neutralization, oxidation and then subject to 5% Ehrlich’s reagent to develop the color. The samples were transferred to 96-well microplates (Corning, NY, USA) and the absorbance was measured at 555 nm with a microplate spectrophotometer (Biotek Instruments, Winooski, VT, USA). The trans-4-hydroxy-L-proline (analytical standard, Sigma-Aldrich, St. Louis, MO, USA) was used as the standard to provide a reference curve for calculating the hydroxyproline release from each sample (μg/mg) during the digestion (n=6). The detailed assay procedure can be found in the reference [33].

2.8. Statistical Analysis

Data were expressed as means ± standard deviation for CHP FI values, weight loss percentages and hydroxyproline releases for all six groups. Kolmogorov-Smirnov and Shapiro-Wilk tests were employed for data normality testing and Levene’s test was performed to check the homogeneity of variance of data. Then all the data were subject to two-factor analysis of variance (ANOVA) Games-Howell test (α=0.05). The two factors were the dentin collagen treatment /denaturation status (factor 1) and PA modification (factor 2). The correlation between the CHP FI values and digestion results (weight loss and hydroxyproline release) was analyzed using correlation coefficient r (α=0.05). Statistical analysis was performed in IBM SPSS v27 (IBM SPSS Inc., Chicago, IL, USA) and Microsoft Excel (Microsoft 365, Microsoft Corporation, Redmond, WA, USA).

3. Results

The FTIR spectral analysis results are summarized in Fig. 2. All spectra are average spectra of 3 samples. Fig. 2a shows the spectra of control (C) and PA powder. The characteristic bands for dentin collagen were identified in all samples as following: amide I at ~1660 cm⁻¹, amide II at ~1545 cm⁻¹, CH₂ bending at ~1450 cm⁻¹, and amide III at ~1235 cm⁻¹. The amide I band is dominantly attributed to the stretching vibrations of peptide C=O groups. And the amide II absorbance arises from the N–H bending vibrations coupled to C–N stretching vibrations. The Amide III is assigned to the C–N stretching and N–H bending vibrations from amide linkages, as well as wagging vibrations of CH₂ groups in the glycine backbone and proline side chains [34]. The spectrum of PA powder showed three distinct bands at 1604, 1519 and 1441 cm⁻¹, which are attributed to aromatic ring stretching (C=C-C).

Figure 2.

Average FTIR spectra for (a) C (control) and PA, (b) C, E and Δ = E − C, (c) C, P and Δ = P− C, (d) C, C+PA, Δ = (C+PA) − C and PA, (e) E, E+PA, Δ = (E+PA) − E and PA, and (f) P, P+PA, Δ = (P+PA) − P and PA. The inserts in (b) and (c) are the deconvoluted spectra for amide I and amide II. C: untreated control collagen, E: enzyme-treated collagen, P: H₃PO₄-treated collagen, PA: proanthocyanidins, Δ: difference spectrum

Fig. 2b shows the spectra of control (C), enzyme-treated collagen (E), and collagen [35, 36]. Fig. 2e shows the spectra (Δ = E - C), which are very similar to each other, with subtle changes at the right shoulders of amides I and II, being better observed in the difference spectrum and in the deconvoluted spectra (insert). Those subtle spectral changes indicate moderate collagen secondary structural alterations from enzyme treatment. Fig. 2c shows the spectra of control (C), H₃PO₄-treated collagen (P), and their difference spectrum (Δ = P - C), which show several distinct differences between C and P. Specifically, compared to the control, the H₃PO₄-treated spectrum exhibits a new band in the region of 1774 – 1704 cm⁻¹, a slight red shift of the right shoulder of amide I toward lower wavenumbers, an increase in the band ratio of amide I to amide II, a shift of the 1403 cm⁻¹ band to 1408 cm⁻¹ with decreased intensity, a new tiny band at 1378 cm⁻¹ and changes in the region of 1162 – 1080 cm⁻¹. The overall changes can be observed in the difference spectrum and the changes in the amide I and amide II region can be better visualized in the deconvoluted spectra (insert). The spectral changes resulted from the H₃PO₄ treatment are more substantial than those from the enzyme treatment, indicating more collagen secondary structural alterations from the H₃PO₄ treatment.

Fig. 2d shows the spectra of control (C), control with PA modification (C+PA), their difference spectrum (Δ = (C+PA) − C), and PA spectrum. The difference spectrum shows a general characteristic of the PA powder spectrum, indicating the incorporation of PA into the demineralized dentin collagen. The difference spectrum also shows an additional band at ~1682 cm⁻¹ which does not come from PA. This band likely indicates a new covalent bonding attributable to the imine (C=N) stretching of the Schiff base formed between PA component and collagen [35, 36]. Fig. 2e shows the spectra of enzyme-treated collagen (E), enzyme-treated collagen with PA modification (E+PA), their difference spectrum (Δ = (E+PA) − E), and PA spectrum. Similarly, a characteristic of the PA spectrum is seen in the difference spectrum, indicating the incorporation of PA into the enzyme-treated dentin collagen. The difference spectrum also shows several spectral features that differ from the PA spectrum at 1682, 1362, 1215, and 1556 cm⁻¹, indicating new bond formation via cross-linking interactions between PA and enzyme-treated collagen. Fig. 2f shows the spectra of H₃PO₄-treated collagen (P), H₃PO₄treated collagen with PA modification (P+PA), their difference spectrum (Δ = (P+PA) − P), and PA spectrum. Once again, the difference spectrum is similar to the PA spectrum, indicating PA incorporated into the H₃PO₄-treated collagen. Moreover, several spectral features that differ from the PA spectrum at 1723, 1682, 1645, 1392, 1240 and 1220 cm⁻¹ are seen in the difference spectrum, indicating new bond formation via cross-linking interactions between PA and H₃PO_4-treated collagen. Notably, the new band in the region of 1774 – 1704 cm⁻¹ that appeared after H₃PO₄ treatment disappears after PA modification, as indicated in the red rectangular boxes in Fig. 2c and 2f.

The CHP analysis results are shown in Fig. 3. The demineralized dentin films showed low fluorescence intensity of 6.01 ± 5.34, the enzyme-treated dentin collagen showed medium fluorescence intensity of 36.54 ± 5.46, and the H₃PO₄-treated dentin collagen showed high fluorescence intensity of 83.22 ± 2.1, all of which were reduced to zero with PA treatment. From the results, phosphoric acid treatment induces significantly higher degree of collagen denaturation than enzyme treatment (p=0.0002), which induces significantly higher degree of collagen denaturation than demineralization treatment (p=0.0023).

Figure 3.

(a) Representative F-CHP staining images for each group. (b) Quantitative fluorescence intensity (FI) analysis of CHP-stained dentin collagen films (n=4). C: control group, E: enzyme-treated group, P: H₃PO₄-treated group, +PA: with PA modification. The letters a, b, c, and d indicate statistically significant differences.

The weight for the control group was 1.903 ± 0.086 mg (6 sample groups of 10 films each). The weight for the enzyme-treated group was 1.855 ± 0.280 mg. The weight for the H₃PO₄-treated group was 1.343 ± 0.110 mg. The enzyme treatment did not change the weight of the dentin collagen, but the phosphoric acid treatment significantly reduced the weight of the dentin collagen (p < 0.0001).

The results of weight loss from collagenase digestion for all test groups are shown in Fig. 4a. The weight loss for the control group was 100% without PA treatment (C) and 4.51 ± 0.88 % with PA treatment (C+PA). The weight loss for enzyme-treated group was 100% without PA treatment (E) and 15.34 ± 3.45 % with PA treatment (E+PA). The weight loss for H₃PO₄-treated group was 100% without PA treatment (P) and 23.00 ± 3.31 % with PA treatment (P+PA). The PA modification significantly reduced the weight loss of both control and denatured (E & P) dentin collagen (p < 0.0001), with the stabilization effect in the order of C > E > P (p < 0.03).

Figure 4.

(a) Percentage of dentin collagen weight loss after collagenase digestion. (b) Hydroxyproline release (μg/mg) during dentin collagen digestion (n=6). C: control group, E: enzyme-treated group, P: H₃PO₄-treated group, +PA: with PA modification. The letters a, b, c, d, and e indicate statistically significant differences.

The results of the hydroxyproline release during digestion for all test groups are shown in Fig. 4b. The hydroxyproline release for the control group was 81.44 ± 5.08 (C) and 6.28 ± 1.69 (C+PA) μg hydroxyproline per milligram of dentin collagen without and with PA modification respectively. The hydroxyproline release for enzyme-treated group was 74.02 ± 8.94 (E) and 20.49 ± 2.40 (E+PA) μg hydroxyproline per milligram of dentin collagen without and with PA modification respectively. The hydroxyproline release for H₃PO₄-treated group was 105.38 ± 9.02 (P) and 28.70 ± 3.65 (P+PA) μg hydroxyproline per milligram of dentin collagen without and with PA modification respectively. PA modification significantly reduced the hydroxyproline release of both control and denatured (E & P) dentin collagen (p<0.0001). The hydroxyproline release among the three PA modified groups increases in the order of C+PA < E+PA < P+PA (p ≤ 0.001). Two-way ANOVA analysis for weight loss and hydroxyproline release showed a significant difference (p<0.001) for both dentin collagen treatment/status (factor 1) and PA crosslinker modification (factor 2). The analysis also revealed a statistically significant interaction between the dentin collagen treatment/status and PA crosslinker modification (p<0.001).

Fig. 5a shows the molecular structures of F-CHP and PA; Fig. 5b illustrates collagen denaturation, F-CHP staining, and PA cross-linking processes as well as the representative CHP fluorescence images of demineralized (control), enzyme and H₃PO₄ denatured, and PA cross-linked collagen. When the denatured collagen is modified/cross-linked with PA, F-CHP is no longer able to hybridize with it.

Figure 5.

(a) molecular structure of F-CHP (top) and PA (bottom). (b) schematic illustration of CHP staining of demineralized control collagen, denatured collagen, and PA cross-linked collagen.

The correlation coefficient was r=0.98 (p=0.14) and r=0.96 (p=0.17) respectively between the denaturation degrees (CHP FI values) of the three treatment groups (C, E, and P), and the weight loss and hydroxyproline release respectively after PA modification (C+PA, E+PA, and P+PA), indicating a negative correlation between the PA stabilization effect and the collagen denaturation degree. The correlations were non-significant (p>0.05) due to the small number of treatment groups (n=3) in the current study.

4. Discussion

Dentin collagen matrix comprises primarily of collagen type I (90% of organic matrix) [37], which is a heteropolymer of two α1(I) chains and one α2(I) chain in the form of a triple-helical structure that is ~300 nm in length and ~1.5 nm in diameter [38]. Each collagen chain consists of repeating (Gly–X–Y) units, where peptides in X and Y positions are frequently proline (Pro) and hydroxyproline (Hyp), respectively [39]. In a single collagen molecule, three polypeptide chains, arranged in a left-handed polyproline-II (PPII) helical conformation, are assembled to form a right-handed superhelix structure. This complex triple helical structure is stabilized by interchain hydrogen bonds, including direct hydrogen bonds between the backbone N-H group of glycine and the back-bone C=O group of a residue in the X position of the neighboring chain, and indirect hydrogen bonds, mediated by water bridges between the backbone N-H group of proline in the X position and the C=O group of glycine [39, 40]. Covalent intermolecular cross-links in collagen provides the dentin organic matrix with stability and tensile strength. [41]. Dentin collagen is the most cross-linked collagen in the body. It is not as easily degraded as other bodily collagens such as the dermal collagen. But unlike insoluble collagen in other parts of the human body, dentin collagen does not metabolically turn over. Therefore, once it is degraded, it cannot be metabolically replaced [42].

This triple-helical conformation makes collagen resistant to most proteinases. Collagenase is a type of enzymes that can cleave the triple-helical structure at body temperature of 37 °C. Collagenase such as MMP-1 can bind and locally unwind the triple-helical structure of collagen before hydrolyzing the peptide bonds. As a result, the collagen denatures at body temperature and can be further degraded by other nonspecific tissue proteinase [43]. The collagen denaturation process involves a rearrangement of the triple helix into a random coil configuration (conformational changes) caused by the breaking of different cross-links present at both intermolecular and intramolecular levels [44] [45]. Dentin collagen denaturation happens naturally in diseased teeth such as carious and sclerotic teeth [16] [23].

The heterogeneity of clinically collected diseased teeth poses challenges for quantitative study of collagen biostability. In the current study, the stabilization effects of PA on experimentally denatured dentin collagen were investigated. Specifically, two experimental treatment protocols were developed – the enzyme treatment (0.1 M lactic acid treatment for 24 hours followed by 30 μg/ml collagenase treatment for 6 hours at 37 °C) and the phosphoric acid treatment (35% phosphoric acid treatment for 3 hours at 37 °C). The current enzyme-treatment protocol was modified from the protocol used previously by Dung et al. [46] and Hannig [47], in which 15 μg/ml collagenase treatment for 24 hours at 37 °C was used. In comparison, the previous protocol resulted in 57% degradation (weight loss of collagen) while the current protocol did not cause any loss of collagen.

PA has been recognized as an effective collagen cross-linker to improve mechanical properties and biostability of dentin collagen matrix for restorative dentistry [7]. However, it is unknown if PA can protect and stabilize denatured dentin collagen. Although the effects of PA on the bonding properties of adhesive to altered dentin substrates (e.g., by acid etching, carious process, soft drink and citric acid erosion) have been investigated in a small number of studies [29–31, 48, 49], the collagen alteration/denaturation status in these studies was undetermined due to the lack of sensitive characterization techniques. The current study is the first one to quantitatively investigate the effect of PA on the biostability of altered dentin collagen with well-characterized denaturation status.

The structural changes from collagen denaturation and interaction between collagen and PA were evaluated by FTIR. The difference spectra between denatured collagen and control, as well as their deconvoluted spectra indicate more secondary structural changes from phosphoric acid treatment than from enzyme treatment (Fig. 2b and 2c). Particularly, the phosphoric acid denatured collagen shows a broad band between 1774 and 1704 cm⁻¹, which likely comes from newly formed C=O ester bond. The FTIR spectra of PA modified dentin collagen showed general features of the PA spectrum on top of the collagen spectrum, including the appearance of a bulge at approximately 1110 cm⁻¹ and a slight shift and widening of the amide II band at approximately 1550 cm⁻¹. The difference spectra between dentin collagen with and without PA modification for both control and denaturation groups clearly reveal the characteristic of the PA spectrum, evidencing the incorporation of PA into both demineralized and denatured dentin collagen. A new band at ~1682 cm⁻¹ appeared in all PA-modified collagen, which could come from the imine (C=N) stretching of the Schiff base formed between PA and collagen [35, 36]. The difference spectra in Fig. 2e and 2f also show several additional spectral features, indicating additional new bond formation between PA and denatured collagen. Interestingly, the C=O ester bond formed with phosphoric acid treatment as shown in Fig. 2c disappeared after PA modification, as shown in Fig. 2f, indicating the breakage of the C=O ester bond by PA modification.

The structural changes and denaturation of collagen can be assessed by several other techniques. Circular dichroism is one of the most common methods for studying the secondary structure of protein. It reveals a decrease of the characteristic peak of the triple helix at about 225 nm during denaturation [50]. Optical rotation shows a significant change in specific optical rotation on denaturation [51]. Denaturation can also be detected by changes in hydrodynamic parameters such as intrinsic viscosity [52] and sedimentation coefficient. The kinetics of denaturation can be studied by differential scanning calorimetry for both insoluble fibers and the corresponding molecules in solution [44]. Microscopically, second harmonic generation (SHG) is a nonlinear optical technique that can detect the presence of non-centrosymmetric structures (e.g., fibrillar collagen types I). The transition of fibrillar collagen to single stranded collagen fibrils during the denaturation process results in the loss of SHG signal [53]. For dentin collagen, its denaturation has been evaluated using atomic force microscope and transmission electron microscopy, as evidenced by the loss of collagen fiber periodicity (67 nm banding) and change of morphological integrity [47].

In the current study, the dentin collagen denaturation was quantitated using a molecular approach called collagen hybridizing peptides (CHP). CHP single strand has a strong propensity to fold into a triple helix and can bind to unfolded collagen chains by forming a hybridized triple helix through inter-strand hydrogen bonds. The CHP has no affinity to intact collagen and other biomolecules, due to lack of binding sites and the neutral and hydrophilic amino acids composition of the peptide. Fluorescently conjugated CHPs have been previously shown to bind to collagen molecules denatured by heat or by enzymes in skin, cornea and bone [54] [55]. Compared to other techniques mentioned above, the CHP technique does not require a special instrument other than a regular light microscopy. Moreover, it allows for in situ visualization and quantitation of collagen denaturation, including the location and distribution of the denatured collagen matrix.

The CHP results indicate low-level CHP hybridization (FI=6.01) from untreated control (demineralized only), medium-level CHP hybridization (FI=36.54) from enzyme treatment, and high-level CHP hybridization from phosphoric acid treatment (FT=83.22). The first null hypothesis is rejected. The low-degree denaturation detected by CHP after demineralization with 10% phosphoric acid for 30 minutes could be due to acid-collagen interaction or surface mechanical damages generated during the microtome cutting process. Acids used to etch dentin substrates and remove smear layers in adhesive dentistry have been found to induce dentin collagen conformational changes and denaturation in some studies [56–58]. CHP has been previously shown to hybridize to mechanically damaged collagen fibrils in tendons [59]. The quantitative CHP results indicate that phosphoric acid treatment used in this study induced more collagen denaturation than the enzyme treatment, which agrees with the qualitative FTIR results.

The interesting finding that PA cross-linking completely prevented the hybridizing of CHP to denatured dentin collagen (Fig. 3) provides direct evidence of the interactions between PA and denatured collagen. PA stabilizes collagen molecule primarily through hydrogen bonding [60] and hydrophobic interactions [61], which results in increased cross-links between collagen fibrils [6, 62]. The C=O and N-H groups on the collagen backbone, and the side chain hydroxyl (−OH) groups provide potential interacting sites for the formation of hydrogen bonds with the phenolic hydroxyl groups (−OH) of PA [63]. After the denaturation treatments, some interchain hydrogen bonds are broken, resulting in partially unwound α chains with free glycine N-H groups and proline C=O groups on the polypeptide backbone that are readily available to form new intermolecular hydrogen bonds with the mobile single strand CHP molecule and therefore can be detected by CHP. However, when the denatured collagen is modified with PA first, the free glycine N-H groups and proline C=O groups on unwound α chains will interact with the phenolic hydroxyl groups (-OH) of PA through hydrogen bonding to form cross-links, which prevents further hybridization of CHP to the denatured collagen (Fig. 5). The CHP molecular staining has become a promising approach for detecting collagen denaturation/damage [32, 59]. Our finding also indicates that extra caution should be taken to analyze the results when using CHP in the presence of other collagen binding reagents.

Weight loss analysis and hydroxyproline assay were used to quantitate the stabilization effects of PA against collagenase digestion of dentin collagen. The results show that both demineralized (control) and denatured collagen modified with PA was significantly more resistant to collagenase digestion than corresponding collagen without PA modification. The findings reject the second null hypothesis and suggest that PA is effective in stabilizing denatured dentin collagen against collagenase digestion.

The strength and durability of adhesion to dentin depends on both the adhesive system used and the status of dentin substrate. Dentists often deal with carious teeth and therefore need to bond to non-ideal dentin substrates, which contain structurally altered or denatured collagen matrix. It has been shown that the bond strengths of adhesives to sound dentin are significantly higher than those to CAD, which, in turn are significantly higher than those to CID [64]. Adhesive bond strength to sclerotic dentin is also shown to be much weaker than to sound dentin substrate [23]. The weaker bonding to those non-ideal substrates could be partially due to the damaged/denatured collagen matrix on the surface of the substrates. Previously, PA has been shown to stabilize collagen in clinical CAD substrates. Specifically, Macedo et al. separated and powdered CAD collagen to evaluate its digestibility with and without PA modification. The authors found that CAD treated with 6.5% PA for 1 hour had a 17% weight loss compared to 91% weight loss without PA modification when exposed to 1% collagenase for 24 hours. No difference in PA stabilization was found between CAD and sound dentin collagen [29]. In the current study, experimentally denatured dentin collagen modified with 0.65% PA showed ~15% (enzyme treatment) and 23% (phosphoric acid treatment) weight loss when exposed to 0.1% collagenase for 1 hour compared to 100% weight loss without PA modification. Although many experimental parameters are different between the two studies, both showed the efficacy of PA modification in stabilizing altered dentin collagen from collagenous digestion. However, due to the lack of techniques for direct monitoring of collagen denaturation in the previous study, the status of collagen in the CAD substrates was unknown, prohibiting further understanding of bonding in these clinically relevant substrates. The use of CHP molecular staining for in situ characterization of the collagen alteration status in CAD, CID or sclerotic dentin substrates is currently undergoing in our laboratory. We aim to understand how PA and other cross-linkers interact with altered/denatured collagen and to develop a strategy that can effectively stabilize the hybrid layer of those clinically relevant non-ideal dentin substrates during adhesive bonding.

Denaturation and cross-linking are two processes that significantly affect the structures of collagen matrix - denaturation weakens the structural integrity of collagen matrix while cross-linking reinforces it. The current work shows that experimentally denatured dentin collagen matrix can be reinforced and protected from collagenase digestion by 0.65% natural cross-linker PA in 30 seconds, and the stabilization effect depends on the degree of collagen denaturation. Dentin biomodification using PA cross-linker could be an effective strategy to biomechanically reinforce altered/denatured dentin collagen matrix in clinically relevant substrates to enhance their restorative/reparative abilities.

The limitations of the current study include the following: 1, that the denaturation induction mechanisms of the two experimental treatment protocols were not investigated, and 2, it is not clear how well the experimental treatment protocols simulate the collagen denaturation induced by the carious process. Those are the areas of future research.

5. Conclusions

The current study demonstrates for the first time that collagen hybridizing peptide is a sensitive technique to quantitate dentin collagen denaturation. Within the limitations of the study, it is also concluded that natural cross-linker PA can stabilize denatured dentin collagen from collagenase digestion and the stabilization effect depends on the degree of collagen denaturation.