Site specific properties of carious dentin matrices biomodified with collagen cross-linkers

Abstract

Purpose

To assess in non-cavitated carious teeth the mechanical properties of dentin matrix by measuring its reduced modulus of elasticity and the effect of dentin biomodification strategies on three dentin matrix zones: caries-affected, apparently normal dentin below caries-affected zone and sound dentin far from carious site.

Methods

Nano-indentations were performed on dentin matrices of carious molars before and after surface modification using known cross-linking agents (glutaraldehyde, proanthocyanidins from grape seed extract and carbodiimide).

Results

Statistically significant differences were observed between dentin zones of demineralized dentin prior to surface biomodification (P< 0.05). Following surface modification, there were no statistically significant differences between dentin zones (P< 0.05). An average increase of 30-fold, 2-fold and 2.2-fold of the reduced modulus of elasticity was observed following treatments of the three dentin zones with proanthocyanidin, carbodiimide and glutaraldehyde, respectively.

Introduction

Carious dentin has been traditionally characterized into two layers: caries-infected and caries-affected dentin. Caries-affected dentin is the inner layer of the carious tissue¹ that contains collagen fibrils believed to retain their triple helical conformation² and inter-molecular cross-links. This layer of carious dentin presents different degrees of demineralization³ and is potentially remineralizable.⁴ Caries-infected dentin (outer layer), is largely demineralized and the collagen fibrils are mostly denatured.³ Caries-infected dentin is usually removed due to limited tissue reparative/regeneration alternatives.

Variable findings regarding the biochemical composition has shown decreased collagen contents of hydroxyproline and glycine, indicative of collagen biodegradation^3,5 as well as no significant differences in amino acid composition among carious dentin layers and sound dentin.^2,6 Differences in the intermolecular cross-links of collagen fibers from carious dentin layers and sound dentin have also shown variable results.^2,3,7 Discrepancies observed among these studies were likely related to the challenges in working with such substrate due to high variability, caries activity (arrested or active), type of teeth (permanent vs. primary), site specificity of the lesions, and biochemical assays. But a consensus is apparent regarding the lower biomechanical properties of the carious tissue when compared to sound dentin. The hardness and elastic modulus of carious dentin have been directly linked to the mineral content of dentin.⁸ Correlation of the stainability of carious dentin in different zones of two types of carious processes (moderately active and arrested), showed that light pink stained intertubular dentin had lower values of hardness and reduced elastic modulus when compared to the transparent and apparently normal dentin (both measured underneath the lesion).⁹ Caries-affected dentin was also found to be less stiff and contained more water than sound dentin; a large range of shrinkage values indicates variable degrees of demineralization.¹⁰

Collagen cross-linking is the post-translational modification to collagen that provides tensile strength and determines the biodegradation rates of collagen. Dentin modification using chemical agents that affect collagen and non-collagenous proteins has been proposed as a bioinspired therapy to modify the tissue for reparative and preventive purposes.¹¹ Specific synthetic and natural cross-linking agents have been shown to enhance the biomechanical properties of the healthy tissue and greatly reduce biodegradation rates of dentin matrix,^11–15 improve resin dentin bonds^6,16 and have the ability for mineral nucleation.^17,18

Restorative/reparative therapies are mostly done in altered forms of dentin substrate, instead of healthy teeth. In vitro studies^19–21 have indicated that the bonding of resin composite restoration is affected by different dentin substrates. Unlike most tissue engineering strategies where one expects the engineered scaffold to be replaced by host tissue, in restorative dentistry, the dentin matrices need to be stabilized so that they will last for decades. The endogenous proteases of dentin need to be inactive or inhibited to prevent proteolysis of the dentin matrix. These matrices cannot remineralize if the endogenous proteases remain active because they destroy the very matrix that is necessary for mineralization.

This investigation measured the modulus of elasticity of longitudinal sections of human carious teeth using nano-indentation to characterize the stiffness values of two zones underneath the caries-infected and one zone of sound dentin before and after application of cross-linking agents. The test null hypothesis was that cross-linking agents had no effect on the reduced modulus of the three dentin zones.

Materials and Methods

Specimen preparation

The use of human molars extracted for dental reasons was approved by the UIC institutional review board. A total of 15 human molars with non-cavitated occlusal carious lesions were selected and stored frozen for no longer than 3 months. Each tooth was sectioned longitudinally, through the occlusal central groove in the mesial-distal orientation. Two additional cuts were then made 2 mm in either direction, from the initial cut, to yield two disc-like samples per tooth.

The exposed enamel and dentin of the obtained sections were polished using 320, 600, 1,200 and 2,400 grit silicon carbide abrasive paper (Carbimet 2^a) in a polisher (EcoMET 3000^a) under running water, to achieve a smooth finish without imperfections. Using standard metallographic polishing techniques, the exposed dentin was further polished to a mirror-like finish with polycrystalline diamond suspensions of grades 9, 6, 3 and 1 μm (MetaDi^a) and 0.05 μm alumina suspension polish (MasterPrep^a) on soft polishing pads (MicroCloth^a). Between each polishing stage, specimens were thoroughly cleaned in an ultrasonicator (Ultrasonic Cleaner Kendal CD 4800^b) for 300 seconds to remove any debris.

A combined visual assessment and use of caries detector solution^c applied for 20 seconds was used to identify carious zones. The use of caries detector solutions is limited clinically due to high false positive readings.²² However, it is an auxiliary tool in laboratory studies to identify the location of different layers of caries.^9,21,23 Based on the criteria described above, three different zones were identified: Zone 1 – caries-affected dentin, (faded pink stained); Zone 2 −1 mm below Zone 1 – apparently normal; and Zone 3 – apparently caries-free site far from the carious lesion –“sound” dentin (Fig. 1).

Fig. 1.

Representative image of a sliced carious tooth showing the testing sites for each dentin zone.

Surface biomodification

Indentations were performed immediately following (1) superficial demineralization (using 32% phosphoric acid gel (Uni-Etch^d) for 60 seconds and rinsed with distilled water for 180 seconds (n=15); and (2) 10-minute treatment with cross-linking agents described below (n=5): GD – 5% glutaraldehyde solution^e (GD), GSE – 6.5% grapeseed extract (GSE) solution (94% proanthocyanidins (MegaGold Natural^f) and EDC/NHS – 0.3M 1-Ethyl-3-[3-dimethylaminopropyl] carbodiimide hydrochloride/0.12M N-hydroxysuccinimide solution. All solutions were adjusted to pH 7.2. Specimens were thoroughly washed with water for 180 seconds following surface treatment. The application of 32% phosphoric acid resulted in ~20 μm deep demineralization of dentin.²⁴

Nanoindentation

Nanoindentation testing was performed using a TI 700 Ubi NanoIndenter.^g The indenter and probe system were carefully calibrated (the tip area function of the probe tip) by making repeated indents on a standard fused quartz surface.^g Specimens were attached to an 18 mm sample mounting metal disc^h using cyanoacrylate glue (Loctite Super Glueⁱ). A well-like boundary was formed around the mounted samples using dental boxing wax, for storing Hank’s balanced salt solution (Lonza^j) in order to keep the samples hydrated throughout the testing procedure.²⁵ The indents were made using an elongated Berkovich diamond fluid tip; a maximum load of 200 μN was used for all evaluation periods using a trapezoidal load function 5–10–5 seconds for loading, holding and unloading times, respectively. During each test, a series of five indents, spaced apart at approximately 20 μm, were made at three different points in each of the dentin zones for a total of approximately 15 indents in each zone. The starting points for the indent lines were selected using scanning probe microscopy to ensure that all five indents were well within the zone being tested.

The test area was scanned and imaged post-indent to confirm that the test was performed on intertubular dentin. Indentation depths varied between 3,600–3,700 nm at baseline and between 600–3200 nm post-treatment. The indentation test data was collected and analyzed and the reduced elastic modulus calculated based on the Oliver & Pharr²⁶ method, using the companion software of the nanoindenter system. The reduced elastic modulus (E_r) was determined from the following equation:

$$E_r=\frac{1}{2}S\sqrt{\frac{\pi }{A}}$$

where S is the contact stiffness that is measured from the load displacement data as the slope of the upper portion of the unloading curve of the resulting load-displacement plot and A is the contact area of the indentation at maximum indentation force.

Statistical analysis was performed using SPSS^k statistical software. A total of 10–15 measurements were used on each zone of each tooth. Indentations were averaged per tooth to calculate percent increase in E_r following dentin biomodification. Two and one-way ANOVAs and post-hoc Scheffé and Games Howell tests were used for homogenous and non-homogenous data respectively, at a 95% confidence interval.

Results

The results are summarized in the Table. Statistically significant differences were observed between dentin zones of demineralized dentin prior to treatment. No statistically significant differences were observed between dentin Zones 1 and 2, while dentin Zone 3 showed significantly higher E_r than Zones 1 and 2 (P< 0.05).

Table.

Reduced modulus of elasticity of three dentin zones on carious teeth of demineralized dentin before and after dentin biomodification strategies.

	Reduced modulus of elasticity (Er)
	Er (GPa) Mean (SD)	Statistics analysis	P values
Demineralized dentin surface
Dentin Zone 1	0.01 (0.01)	Zone 1 × Zone 2	0.416
Dentin Zone 2	0.01 (0.00)	Zone 1 × Zone 3	0.002
Dentin Zone 3	0.02 (0.01)	Zone 2 × Zone 3	0.000
Dentin biomodification strategies following demineralization 6.% Proanthocyanidin rich grape seed extract – GSE
Dentin Zone 1	0.42 (0.25)	Interaction (Zones × Treatments)	0.851
Dentin Zone 2	0.42 (0.26)	Zones	0.987
Dentin Zone 3	0.40 (0.25)	Treatments	0.000
5% Glutaraldehyde – GD		GD × EDC	0.829
Dentin Zone 1	0.03 (0.01)	GD × GSE	0.000
Dentin Zone 2	0.02 (0.01)	GSE × EDC	0.000
Dentin Zone 3	0.04 (0.01)
Carbodiimide −0.3M EDC/0.12M NHS
Dentin Zone 1	0.02 (0.01)
Dentin Zone 2	0.02 (0.01)
Dentin Zone 3	0.03 (0.01)

N = 15. Dentin Zone 1 = zone just beneath soft caries-infected dentin that stained faintly pink with caries-detector solution (caries-affected); Dentin Zone 2 = Zone 1 mm below zone 1 that did not pick up any caries-detector stain (it is regarded as deep normal dentin); Dentin Zone 3 = superficial normal dentin far away from carious lesion. PA = phosphoric acid gel. *Statistical significant differences among groups are indicated by P< 0.05.

Demineralized dentin surface modification by cross-linkers resulted in increased E_r values. The highest increase of E_r was observed for GSE treatment, while no differences were observed between GD and EDC/NHS. Following treatment, there were no statistically significant differences between dentin zones, indicating a reinforcement mechanism for dentin under the affected carious zone. The percentage increase in the E_r (calculated as E_r post treatment − E_r pre-treatment × 100) was significantly higher for GSE when compared to the other two agents (P< 0.001) (Fig. 2).

Fig. 2.

Percentage increase of the reduced modulus of elasticity (E_r) of demineralized dentin following dentin biomodification.

Discussion

It has been well documented that natural carious dentin that has been subjected to repeated cycles of demineralization and remineralization is softer than normal healthy dentin. Presumably, the lower hardness and modulus of elasticity reflects lower mineral content in intertubular dentin.^8,9,21,23 The present data however showed that the E_r of dentin matrix also varied between zones, indicating that not only changes to the mineral content but biochemical and structural changes to dentin matrix may compromise the mechanical properties of the tissue. It also highlights that the damage to the dentin matrix due to caries extended to areas considered apparently sound (below the boundaries of the caries-affected dentin).

The lower elastic modulus of various dentin zones in carious dentin has been previously shown to be affected by the hydration of the tissue.²³ Type I collagen is the most abundant protein in dentin and plays a major role in the mechanical properties and fracture mechanics²⁷ of the tissue. Four different types of post-translational enzymatic collagen cross-linking (two reducible and two non-reducible) between non-helical and helical domains within and between adjacent collagen molecules, primarily define the tensile strength and biodegradation rate of the matrix. Hydroxylation of Lys and Hyl is indicative of presence of collagen cross-linking.²⁸ A consensus seems to indicate that irreversible destruction to collagen in caries-infected dentin is reflected by a remarkable decreased number of reducible cross-linking (dehydro-dihydroxylysinonorleucine – deH-DHLNL) and dehydro-hydroxylysinonorleucine – deH-HLNL)² and increased amounts of precursors (DHNL and HLNL). Varying results, however, have been reported for caries-affected dentin, where no changes to hydroxyproline,³ or Lys and Hyl residues⁶ were observed in primary and permanent molars’ carious-affected dentin, respectively. Conversely, a study showed a decreased amount of reducible cross-links (i.e. DHLNL and HLNL) and increased oxidation and Maillard reaction-related products⁷ in caries-affected dentin. The present biomechanical data supports structural changes to the dentin matrix that significantly reduced the elastic modulus of the caries-affected dentin (Zone 1) and apparently normal dentin underneath the arrested lesion (Zone 2). Besides structural changes to dentin type I collagen, altered distribution of proteoglycans may contribute to the damaging effect of dentin caries and deserve further investigation. Reduced antigenicity of proteoglycans has been reported in sclerotic dentin,²⁹ potentially decreasing the mechanical properties of the tissue.³⁰

The current work showed that dentin matrix can be reinforced by chemical collagen cross-linkers. Interestingly, interactions detected by increased E_r were more pronounced in damaged dentin matrix (Zones 1 and 2). Besides enzymatic pathways, covalent and non-covalent collagen cross-linking can be induced by non-enzymatic reactions. Synthetic and naturally occurring collagen cross-linking agents induce non-enzymatic cross-linking and increase biomechanical properties of sound dentin^12,13 and greatly reduce biodegradation rates.^6,11 The ability of these agents to chemically interact with non-collagenous components (i.e. proteoglycans, cathepsins, MMPs) has greatly increased their impact in biomedical applications. Naturally occurring proanthocyanidins present very complex chemical interactions with collagen, which is evident by the 30-fold increase in dentin matrix E_r when compared to synthetic agents. The presence of high molecular weight oligomeric proanthocyanidins in GSE is highly significant for the very potent interaction with dentin matrix components, even when compared to other proanthocyanidins sources.¹⁵ Many benefits already reported for GSE proanthocyanidins make this agent a powerful instrument for tissue reinforcement and stability. Ongoing research is in progress to reduce the exposure time (~1 minute) and/or develop different delivery systems to establish clinically relevant protocols. It is important to note that different sources of proanthocyanidins will interact differently with the tissue since the concentration and type of proanthocyanidins vary greatly between extraction techniques and material sources.

GD was used as a positive control, due to its well known cross-linking mechanisms to dentin collagen,³¹ but its clinical application is limited by high toxicity. Despite the fact that changes to E_r were much lower for EDC/NHS; its rapid, residue free and stable reaction with collagen may be useful for specific dentin reparative strategies, such as dentin stabilization for remineralization^17,18 and reduced collagen biodegradability.¹¹ While GD and EDC cross-linking mechanism take place at distinct sites, both synthetic agents are limited by the availability of free amino and carboxyl groups for cross-linking which can explain the lower E_r values when compared to GSE. Therefore, EDC cross-linking is limited by the proximity of reactive residues.³²

In clinical situations clinicians must deal with teeth that present caries or abraded-sclerotic dentin and therefore the need to utilize a non-ideal substrate. The challenges to restore caries-affected dentin^6,21,33 have been well-documented. In addition, new reparative/regenerative strategies for remineralization of dentin caries has drawn great interest to dentin matrix composition, collagen nature state and stability to function as scaffold for tissue remineralization.^17,18 Nano-mechanical characterization of the tissue interaction with surface modification agents has been herein demonstrated to be possible in damaged and healthy dentin. Since the current model used to determine the nanomechanical properties does not fully account for time-dependent properties of the demineralized dentin, including its viscoelasticity and creep behavior³⁴ or for pile effects,³⁵ it is important to note that the values of the nanomechanical properties evaluated in this study were estimates for comparison purposes and should not be considered absolute.³⁶ The interaction of the biomodification agents and dentin matrix has been shown elsewhere to be long-lasting,^11,15 and therefore may be a powerful and efficient mechanism to reinforce damaged dentin for clinical applications.

Clinical significance.

Dentin biomodification is an effective strategy to biomechanically reinforce carious teeth by inducing changes to collagen biochemistry. Reinforcement of carious tissue may increase success of restorations to caries-affected dentin.