Highlights
- •
We characterize optical and mechanical properties of enamel surfaces during erosion and remineralization.
- •
3D imaging of enamel surfaces undergoing erosive wear requires instrumentation with lateral resolution significantly less than 5 μm.
- •
3D surface texture parameters are able to successfully characterize textural changes in enamel during erosive demineralization and salivary remineralization.
- •
The role of surface texture in remineralization is less clear but suggests that in vivo remineralized lesions remain rougher despite the surface microhardness recovering to baseline levels.
Abstract
Objectives
This paper investigates the application of confocal laser scanning microscopy to determine the effect of acid-mediated erosive enamel wear on the micro-texture of polished human enamel in vitro .
Methods
Twenty polished enamel samples were prepared and subjected to a citric acid erosion and pooled human saliva remineralization model. Enamel surface microhardness was measured using a Knoop hardness tester, which confirmed that an early enamel erosion lesion was formed which was then subsequently completely remineralized. A confocal laser scanning microscope was used to capture high-resolution images of the enamel surfaces undergoing demineralization and remineralization. Area-scale analysis was used to identify the optimal feature size following which the surface texture was determined using the 3D (areal) texture parameter Sa.
Results
The Sa successfully characterized the enamel erosion and remineralization for the polished enamel samples ( P < 0.001).
Significance
Areal surface texture characterization of the surface events occurring during enamel demineralization and remineralization requires optical imaging instrumentation with lateral resolution <2.5 μm, applied in combination with appropriate filtering in order to remove unwanted waviness and roughness. These techniques will facilitate the development of novel methods for measuring early enamel erosion lesions in natural enamel surfaces in vivo .
1
Introduction
Currently, no reliable technique exists for the quantification of early enamel erosion in vivo . In vitro , microhardness is considered the gold standard for characterising the early enamel erosion lesion , however as this technique cannot be used in vivo there has recently been increased interest in characterising textural changes that occur during early erosive demineralization and remineralization . In order to reliably quantify dynamic changes occurring in the enamel surface during acid mediated erosion, appropriate surface texture instrumentation and software would need to be carefully chosen in order to image the micro-scale features of exposed enamel rods or prisms .
Recent investigations into early erosive wear have suggested that enamel surface texture characterization could be a suitable target for therapeutic oral care products . Moreover, in order to apply these analytical techniques in vivo , data on the optimal scale at which to measure these surface events are required in order to guide the choice of instrument selection, especially in terms of minimum lateral resolution required. To date, scale dependant relative area analysis used in anthropological micro-wear measurement have yet to be applied to determine how to optimally employ 3D ‘areal’ surface texture analysis in order to characterize the surface events occurring in human enamel during acid initiated erosive wear and salivary mediated remineralization .
The aim of this study was therefore to determine the optimal scale at which enamel surface textural changes from citric acid demineralization and salivary remineralization can be observed in vitro , using confocal laser scanning microscopy. The objective was to utilize this clinically relevant enamel erosion model to determine the optimal surface texture measurement workflow to best characterize the enamel surface events occurring during erosion in vitro . The null hypothesis was that surface texture analysis with a confocal laser scanning microscope will not be able to characterize the development of early erosive lesions and their remineralization by human saliva in polished human enamel in vitro .
2
Materials and methods
Twenty enamel specimens (5 mm × 3 mm × 2 mm) were prepared from the mid-coronal portion of the buccal and lingual surfaces of extracted caries-free human third molar teeth, using a diamond wafering blade (XL 12205, Benetec Ltd., London, UK). Research ethics approval for use of Human Tissue in this study had been granted (REC number 09/H0808/109). From pilot data a sample size calculation revealed that at 5% level of significance, to test the null hypothesis of correlation between two measures as −0.5 against an alternative of −0.76, requires a total sample of 20 samples to achieve the power of 80% to test the significance of correlation, assuming the bi-variate normal model. The power calculation was carried out using the statistical freeware Gpower (version 3.1.5) .
The samples were embedded in Protemp4 ® (3M ESPE, Germany) using a dedicated mould former (SyndicadIngenieurbüro, München, Germany) and subjected to a standardized previously published grinding-polishing protocol which resulted in an area of enamel, which was around 5 mm × 3 mm in area with a flatness tolerance of 0.4 μm and homogenous baseline roughness values . All samples were then subjected to an in vitro erosion-remineralization model in order to simulate an early enamel erosion lesion in vitro , as described by Young and Tenuata . A 0.3% citric acid solution was prepared by adding citric acid powder (Sigma–Aldrich, Poole, Dorset, UK) to distilled water, following which the pH was adjusted to 3.2 using a sodium hydroxide buffer and a calibrated pH meter and electrode (WD-35801-00 pH electrode Eutech Instruments, Nijkerk, Netherlands). The solution had a titratable acidity of 19.5 ml, measured as the volume of 0.1 M solution of sodium hydroxide required to raise 20 ml of citric acid solution to pH 7.0 by adding increasing volumes of sodium hydroxide solution followed by agitation and equilibrium for 2 min until the pH reached 7.0.
Each sample was immersed in 50 ml of the citric acid solution at room temperature for the following time points: 30 s, 1 min, 2 min and 5 min, after which the samples were rinsed in distilled water (pH 6.8) and allowed to dry before measurement.
Following erosion, the samples were rinsed in distilled water and then immersed in pooled human saliva to allow remineralization of the eroded enamel lesions. Paraffin-stimulated whole mouth saliva samples were collected from 20 healthy volunteers, following previously published protocols . The collected saliva was ice-chilled and pooled immediately after collection at −80 °C for long-term storage. Prior to use, the frozen natural saliva was defrosted in ice time at room temperature 22 ± 1 °C. The pH of the saliva was 7.1 and the calcium content was 1.4 mmol/l as measured using inductively coupled plasma mass spectroscopy (ICP-MS) . Each 5 samples were immersed in 20 ml of the saliva at room temperature for the following time intervals 1 h, 6 h and 12 h. After each rinsing period the samples were removed from the saliva, rinsed in distilled water and allowed to dry before measurement.
The enamel surface microhardness and surface texture was measured at baseline (prior to the erosion/remineralization model) and again after 30 s, 1 min, 2 min and 5 min of immersion in citric acid (erosion) and after 1 h, 6 h, 12 h of immersion in pooled human saliva (remineralization). For microhardness, an average Knoop Hardness number (KHN) was calculated from three indentations made using a Duramin-5 Hardness Tester (Struers Inc., Rotherham, UK) with dwell time 5 s, load 0.981 N and each indentation made 100 μm apart, to ensure that there were no interaction between indents next to each other. During sample repositioning, the live video interface was used to examine the surface to ensure that indents were placed no closer than 100 μm from adjacent indents. The accuracy of the tester was 39.33 KHN as measured using a 600 KHN calibrated transfer standard block (Staatliches Materialsprufungsamt Nordrhein-Westfalen, Dortmund, Germany). For surface texture measurement, five 129 μm × 129 μm measurements were made using the ×50 objective, 0.95 NA lens of a confocal laser scanning microscope (LEXT OLS4100, Olympus, Tokyo, Japan) employing a 0.2 μm diameter, 405 nm wavelength laser beam.
In order to select the optimal scale at which to carry out the 3D surface texture analysis for the erosion/remineralization time points, a correlation analysis between the changes in microhardness and the changes in surface texture at varying relative area-scales was carried out (Sfrax 1.0 http://www.surfract.com ). This analysis was conducted in order to determine the optimal area scale (in μm 2 ) at which the surface texture parameters would best highlight the enamel surface features, with reference to the analytical technique microhardness.
This area-scale/microhardness correlation data were then used to in order to optimally highlight textural data regarding the relevant features (i.e. the eroded interprismatic enamel pattern) which corresponded to a scale of approximately 20 μm 2 . This information guided the selection of the appropriate filters which were applied to discard unwanted waviness and noise data from the 3D profiles thus ensuring that only data on the relevant feature of interest (i.e. the eroded enamel prisms) was included in the texture analysis. The refinement of the filters was carried out using an iterative process within MountainsMap ® whereby representative sample images were taken from each erosion/demineralization stage and the image analysis workflow was subsequently performed with the 3D data displayed at each of the time points (baseline, during erosion and during remineralization). In order to confirm the optimal filters, which would highlight pertinent data of the feature of interest in the present study (i.e. the enamel prism structure) most useful low-pass and high-pass filters were determined. As a result of this iterative process, the following filters were applied using MountainsMap ® surface texture software (Premium v7.1, Digital Surf, France), as shown in Fig. 1 . Firstly, a 1 μm cut-off robust Gaussian low-pass filter was applied in order to remove high spatial frequencies of the measurement noise. This cut off was chosen to be 1/5th the feature size in that it would not cause any distortion. Following this, a 30 μm cut-off Gaussian high pass filter (i.e. six times the feature size) was applied to remove the irrelevant long wavelength spatial components, i.e. waviness. This allowed the mean (SD) roughness parameter Sa to be used to characterize the enamel surface texture at each erosion/remineralization timepoint.
2
Materials and methods
Twenty enamel specimens (5 mm × 3 mm × 2 mm) were prepared from the mid-coronal portion of the buccal and lingual surfaces of extracted caries-free human third molar teeth, using a diamond wafering blade (XL 12205, Benetec Ltd., London, UK). Research ethics approval for use of Human Tissue in this study had been granted (REC number 09/H0808/109). From pilot data a sample size calculation revealed that at 5% level of significance, to test the null hypothesis of correlation between two measures as −0.5 against an alternative of −0.76, requires a total sample of 20 samples to achieve the power of 80% to test the significance of correlation, assuming the bi-variate normal model. The power calculation was carried out using the statistical freeware Gpower (version 3.1.5) .
The samples were embedded in Protemp4 ® (3M ESPE, Germany) using a dedicated mould former (SyndicadIngenieurbüro, München, Germany) and subjected to a standardized previously published grinding-polishing protocol which resulted in an area of enamel, which was around 5 mm × 3 mm in area with a flatness tolerance of 0.4 μm and homogenous baseline roughness values . All samples were then subjected to an in vitro erosion-remineralization model in order to simulate an early enamel erosion lesion in vitro , as described by Young and Tenuata . A 0.3% citric acid solution was prepared by adding citric acid powder (Sigma–Aldrich, Poole, Dorset, UK) to distilled water, following which the pH was adjusted to 3.2 using a sodium hydroxide buffer and a calibrated pH meter and electrode (WD-35801-00 pH electrode Eutech Instruments, Nijkerk, Netherlands). The solution had a titratable acidity of 19.5 ml, measured as the volume of 0.1 M solution of sodium hydroxide required to raise 20 ml of citric acid solution to pH 7.0 by adding increasing volumes of sodium hydroxide solution followed by agitation and equilibrium for 2 min until the pH reached 7.0.
Each sample was immersed in 50 ml of the citric acid solution at room temperature for the following time points: 30 s, 1 min, 2 min and 5 min, after which the samples were rinsed in distilled water (pH 6.8) and allowed to dry before measurement.
Following erosion, the samples were rinsed in distilled water and then immersed in pooled human saliva to allow remineralization of the eroded enamel lesions. Paraffin-stimulated whole mouth saliva samples were collected from 20 healthy volunteers, following previously published protocols . The collected saliva was ice-chilled and pooled immediately after collection at −80 °C for long-term storage. Prior to use, the frozen natural saliva was defrosted in ice time at room temperature 22 ± 1 °C. The pH of the saliva was 7.1 and the calcium content was 1.4 mmol/l as measured using inductively coupled plasma mass spectroscopy (ICP-MS) . Each 5 samples were immersed in 20 ml of the saliva at room temperature for the following time intervals 1 h, 6 h and 12 h. After each rinsing period the samples were removed from the saliva, rinsed in distilled water and allowed to dry before measurement.
The enamel surface microhardness and surface texture was measured at baseline (prior to the erosion/remineralization model) and again after 30 s, 1 min, 2 min and 5 min of immersion in citric acid (erosion) and after 1 h, 6 h, 12 h of immersion in pooled human saliva (remineralization). For microhardness, an average Knoop Hardness number (KHN) was calculated from three indentations made using a Duramin-5 Hardness Tester (Struers Inc., Rotherham, UK) with dwell time 5 s, load 0.981 N and each indentation made 100 μm apart, to ensure that there were no interaction between indents next to each other. During sample repositioning, the live video interface was used to examine the surface to ensure that indents were placed no closer than 100 μm from adjacent indents. The accuracy of the tester was 39.33 KHN as measured using a 600 KHN calibrated transfer standard block (Staatliches Materialsprufungsamt Nordrhein-Westfalen, Dortmund, Germany). For surface texture measurement, five 129 μm × 129 μm measurements were made using the ×50 objective, 0.95 NA lens of a confocal laser scanning microscope (LEXT OLS4100, Olympus, Tokyo, Japan) employing a 0.2 μm diameter, 405 nm wavelength laser beam.
In order to select the optimal scale at which to carry out the 3D surface texture analysis for the erosion/remineralization time points, a correlation analysis between the changes in microhardness and the changes in surface texture at varying relative area-scales was carried out (Sfrax 1.0 http://www.surfract.com ). This analysis was conducted in order to determine the optimal area scale (in μm 2 ) at which the surface texture parameters would best highlight the enamel surface features, with reference to the analytical technique microhardness.
This area-scale/microhardness correlation data were then used to in order to optimally highlight textural data regarding the relevant features (i.e. the eroded interprismatic enamel pattern) which corresponded to a scale of approximately 20 μm 2 . This information guided the selection of the appropriate filters which were applied to discard unwanted waviness and noise data from the 3D profiles thus ensuring that only data on the relevant feature of interest (i.e. the eroded enamel prisms) was included in the texture analysis. The refinement of the filters was carried out using an iterative process within MountainsMap ® whereby representative sample images were taken from each erosion/demineralization stage and the image analysis workflow was subsequently performed with the 3D data displayed at each of the time points (baseline, during erosion and during remineralization). In order to confirm the optimal filters, which would highlight pertinent data of the feature of interest in the present study (i.e. the enamel prism structure) most useful low-pass and high-pass filters were determined. As a result of this iterative process, the following filters were applied using MountainsMap ® surface texture software (Premium v7.1, Digital Surf, France), as shown in Fig. 1 . Firstly, a 1 μm cut-off robust Gaussian low-pass filter was applied in order to remove high spatial frequencies of the measurement noise. This cut off was chosen to be 1/5th the feature size in that it would not cause any distortion. Following this, a 30 μm cut-off Gaussian high pass filter (i.e. six times the feature size) was applied to remove the irrelevant long wavelength spatial components, i.e. waviness. This allowed the mean (SD) roughness parameter Sa to be used to characterize the enamel surface texture at each erosion/remineralization timepoint.
