Ethanol-wet Bonding and Chlorhexidine Improve Resin-Dentin Bond Durability: Quantitative Analysis Using Raman Spectroscopy Supitcha Talungchita / Julie L.P. Jessopb / Deborah S. Cobbc / Fang Qiand / Saulo Geraldelie / David H. Pashleyf / Steven R. Armstrongg
Purpose: To directly test the effectiveness of ethanol-wet bonding (EW) in improving monomer infiltration into demineralized dentin through quantitative measurement of bis-GMA and TEG-DMA molar concentrations within hybrid layers, and to comprehensively evaluate the effect of EW and chlorhexidine on durability of resin-dentin bonds compared to conventional water-wet bonding (WW). Materials and Methods: A three-step etch-and-rinse adhesive (70% bis-GMA/28.75%TEG-DMA) was applied to coronal dentin using a clinically relevant ethanol-wet bonding protocol (EW) or the conventional water-wet bonding (WW) technique. Bis-GMA and TEG-DMA molar concentrations at various positions across the resin/dentin interfaces formed by EW and WW were measured using micro-Raman spectroscopy. The experiment was repeated at the same positions after 7-month storage in phosphate buffer solution containing 0.1% sodium azide. The μTBS and hybrid layer morphology (TEM) of bonding groups with and without chlorhexidine application were compared immediately and after 1-year storage in terms of nanoleakage, collagen fibril diameter, collagen interfibrillar width, and hybrid layer thickness. Results: Specimens bonded with EW showed significantly higher monomer molar concentrations and μTBS throughout the hybrid layer immediately and after storage, providing direct evidence of superior infiltration of hydrophobic monomers in EW compared to WW. Microscopically, EW maintained interfibrillar width and hybrid layer thickness for resin infiltration and retention. The application of chlorhexidine further preserved collagen integrity and limited the degree of nanoleakage in EW after 1-year storage. Conclusion: EW enhances infiltration of hydrophobic monomers into demineralized dentin. The results suggest that a more durable resin-dentin bond may be achieved with combined usage of a clinically relevant EW and chlorhexidine. Keywords: bonding, collagen(s), matrix metalloproteinases (MMPs), ultrastructure, dentin, adhesives. J Adhes Dent 2014; 16: 441–450. doi: 10.3290/j.jad.a32695
Submitted for publication: 23.10.13; accepted for publication: 17.4.14
a
Full-time Faculty Member, Department of Conservative Dentistry, Prince of Songkla University, Songkhla, Thailand. Hypotheses, experimental design, performed the experiments in partial fulfillment of requirements for PhD, wrote the manuscript, contributed substantially to discussion.
f
Emeritus Regents’ Professor of Oral Biology, Department of Oral Biology, College of Dental Medicine, Georgia Regents University, Augusta, GA, USA. Idea, consulted on ethanol-wet bonding and chlorhexidine, proofread the manuscript, contributed substantially to discussion.
b
Associate Professor, Department of Chemical and Biochemical Engineering, College of Engineering, University of Iowa, Iowa City, IA, USA. Idea, consulted on the experiment using Raman spectroscopy, proofread the manuscript, contributed substantially to discussion.
g
Professor and Chair, Department of Operative Dentistry, College of Dentistry, University of Iowa, Iowa City, IA, USA. Idea, hypotheses, experimental design, proofread the manuscript, contributed substantially to discussion.
c
Associate Professor, Department of Operative Dentistry, College of Dentistry, University of Iowa, Iowa City, IA, USA. Proofread the manuscript, contributed substantially to discussion.
Correspondence: Professor Steven R. Armstrong, Department of Operative Dentistry, College of Dentistry, University of Iowa, S229A DSB, Iowa City, IA, USA 52242. Tel: +1-319-335-7211, Fax: +1-319-335-7267. e-mail: [email protected]
d
Associate Research Scientist, Department of Preventive and Community Dentistry, College of Dentistry, University of Iowa, Iowa City, IA, USA. Performed statistical analyses, interpreted statistical findings, proofread the manuscript.
e
Clinical Associate Professor, Department of Restorative Dental Sciences, College of Dentistry, University of Florida, Gainesville, FL, USA. Idea, consulted on the experiment using TEM, proofread the manuscript, contributed substantially to discussion.
Vol 16, No 5, 2014
Parts of this study were presented at the IADR 86th General Session and Exhibition, Toronto, Canada, 2008, Abstract #0356 (Poster session), IADR 87th General Session and Exhibition, Miami, FL, USA, 2009, Abstract #1517 (Oral presentation session), and IADR 91st General Session and Exhibition, Seattle, WA, USA, 2013, Abstract #1668 (Oral presentation session).
441
Talungchit et al
B
onding effectiveness to dentin relies on resin infiltration and encapsulation of the demineralized collagen fibril network, creating a hybrid layer.23 Deterioration of the hybrid layer is a consequence of hydrolysis of the polymer matrix and degradation of unencapsulated demineralized collagen.2,20 Therefore, enhanced resin infiltration and sustained integrity of both resin matrix and collagen are crucial for a durable resin-dentin bond. Hydrophobic monomers are preferable to hydrophilic monomers to prevent water sorption and hydrolytic degradation of the resin matrix.16 However, hydrophobic monomers fail to infiltrate hydrophilic water-saturated demineralized dentin using the conventional water-wet bonding technique (WW).1 The ethanol-wet bonding technique (EW) replaces water in demineralized dentin with ethanol, thus increasing miscibility and solubility of subsequently applied hydrophobic resin adhesives, yielding strong and durable bonds.28 Ethanol, a transitional solvent, maintains interfibrillar width by preventing interpeptide H-bonding in collagen, thereby preventing the collapse of the collagen matrix during solvent evaporation and resin adhesive application.25 Various EW protocols have been used in recent studies.15,26,29 However, the results were non-uniform, and there is no study which quantitatively measured and compared the amount of dimethacrylate monomers, ie, bis-GMA and TEGDMA, in hybrid layers formed by EW and WW. Once activated, host-derived matrix metalloproteinases (MMPs)22 and cysteine cathepsins24 slowly degrade unencapsulated collagen within hybrid layers, compromising resin-dentin bond durability. Chlorhexidine (CHX) was introduced in bonding to inhibit proteolytic degradation.10,32 The use of CHX with WW produced a more stable collagen morphology and better bond strength compared to no CHX both in vivo13 and in vitro.5,7 Although limited amounts of denuded collagen are expected when using EW, CHX may still be beneficial in preserving collagen integrity. The purpose of this study was to evaluate the shortand long-term (1-year) effects of EW and CHX on resindentin microtensile bond strength (μTBS) and hybrid layer quality as measured by nanoleakage, collagen fibril diameter, interfibrillar width, hybrid layer thickness, and the presence of collagen banding using TEM. Micro-Raman spectroscopy was used to evaluate monomer infiltration in both EW and WW specimens by measuring the bis-GMA and TEG-DMA molar concentrations (CM) across hybrid layers immediately and after 7-month storage in phosphate buffer solution. The null hypotheses were that EW and CHX had no effect on resin-dentin μTBS nor on hybrid layer quality and that EW had no effect on CM within hybrid layers compared to WW.
MATERIALS AND METHODS Microtensile Bond Strength Test Institutional Review Board (IRB)-approved, sound human molars were stored in 0.5% chloramine-T at 4°C and used within 6 months of extraction. Teeth were rinsed free of chloramine-T and stored in PBS before being randomly divided into WW and EW groups (n = 32 per 442
group) and ground with a carbide bur (#55, Brasseler; Savannah, GA, USA) in a computer numerical control (CNC) specimen former (University of Iowa, Iowa City, IA, USA) to expose a flat surface of coronal dentin. A 0.5-mm-deep groove was placed across the occlusal surface to allow separate surface treatments on the two halves. The three-step etch-and-rinse technique was applied for WW. For EW, the acid-etched and waterrinsed dentin was subsequently rinsed with absolute ethanol for 15 s three times with blotting in between. Before priming, half of each tooth in WW and EW was treated for 30 s with 2% aqueous chlorhexidine digluconate (Cavity Cleanser, Bisco; Schaumburg, IL, USA) or 2% chlorhexidine diacetate in ethanol (Pashley Lab; Augusta, GA, USA), respectively, creating 4 bonding protocols: EW-no-CHX, EW-with-CHX, WW-no-CHX, and WW-with-CHX. Then, primer (50% adhesive/50% ethanol) was applied for 30 s. After solvent removal, experimental nonsolvated adhesive (70% bis-GMA/28.75% TEGDMA; Bisco) was applied and light cured (Demetron 501, irradiance 700-780 mW/cm2, Kerr; Orange, CA, USA) for 20 s. Resin-based composite (Z100, 3M ESPE; St Paul, MN, USA) was placed and light cured for 40 s in three 1.5-mm-thick increments. Four dumbbell-shaped specimens with a circular crosssectional area of 0.5 mm2 were formed per tooth using the CNC specimen former. Any tooth failing to produce 4 testable specimens was replaced with a different tooth. One beam from each side of the tooth, with and without CHX, was immediately tensile tested to failure at 1 mm/ min in a passive gripping device3 (n = 32 per bonding protocol). The remaining specimens were stored at 37°C in phosphate buffer solution containing 0.1% sodium azide (n = 32 per bonding protocol), which was changed monthly until testing after 1-year storage. Fractography After the μTBS test, debonded specimens were glued to aluminum stubs using a cyanoacrylate cement (Zapit, Dental Ventures of America; Corona, CA, USA) and then gold sputter coated for evaluation with SEM (Amray 1820-D, AMRAY; Bedford, MA, USA). Fracture pathways were classified into 4 types: dentin cohesive failure, resin-based composite cohesive failure, joint failure, and mixed failure. Hybrid Layer Morphology Six additional teeth were randomly divided into WW and EW groups (n = 3/per group). Within each group, the teeth were bonded as described above, except flowable resin-based composite (Heliomolar Flow, Ivoclar Vivadent; Amherst, NY, USA) was used to facilitate sectioning for ultrastructural evaluations using TEM (JEOL 1230, 120 kV; Tokyo, Japan). Four 0.5 × 1 mm2 resindentin beams from each half of each tooth were randomly selected to evaluate nanoleakage and collagen in the short term and after 1-year storage. For the shortterm evaluations, two beams were placed overnight in Karnovsky’s fixative, one for nanoleakage and another for collagen ultrastructure. The remaining beams were The Journal of Adhesive Dentistry
Talungchit et al
stored at 37ºC in phosphate buffer solution containing 0.1% sodium azide for long-term study (1 year). For nanoleakage evaluation, specimens were stained before sectioning with 50% ammoniacal silver nitrate for 24 h, then placed in developer under fluorescent light for 8 h. Specimen coding was used for blinded evaluation. Dehydration, resin embedding, and thin sectioning were performed. For nanoleakage types, five 90-nm-thick sections from each specimen were evaluated (n = 15 per bonding protocol). After sectioning, for collagen ultrastructural evaluation, 5 non-demineralized 60-nm-thick sections from each specimen were double-stained in 1% phosphotungstic acid for 10 min and 5% uranyl acetate for 2 min. TEM images were randomly taken and measurements were performed using ImageJ (National Institutes of Health; Bethesda, MD, USA). Hybrid layer thickness was randomly measured at one spot per section (n = 15 per bonding protocol). Collagen fibril diameter and interfibrillar width within a 5 × 5 μm2 viewing area at upper, middle, and lower hybrid-layer regions were measured at 10,000X magnification in one area per section. Dentinal tubules and their branches were avoided. In tangential sections where the fibril profiles were oblong, the width of the fibril was taken to be the diameter. Within each viewing area, all cross-sectioned collagen fibrils were measured for collagen fibril diameter (3 to 10 fibrils per viewing area), and the same number of interfibrillar width measurements were performed. The mean of each measurement within each viewing area was calculated. Three-way ANOVA was performed to assess the significance of hybrid layer region on the mean collagen fibril diameter and mean interfibrillar width. Since no significant effect was found, mean collagen fibril diameter and interfibrillar width within each viewing area was assumed to be an independent statistical unit (n = 45 per bonding protocol). The presence of collagen banding was defined as 67 nm “collagen banding” if more than 25% of fibril banding was present under the TEM viewing area (n = 45 per bonding protocol). Monomer Infiltration (Micro-Raman Spectroscopy) Experimental co-monomers (Biomaterials lab, College of Dentistry, University of Iowa) with different bis-GMA/ TEG-DMA concentrations (Table 1) were prepared to construct calibration curves (Figs 1 and 2) for CM calculation. Each co-monomer was placed in 3 aluminum pans and polymerized for 60 s (Demetron Optilux 500, irradiance 700 to 800 mW/cm2, Kerr) to create resin blocks. Polymer samples were removed from aluminum pans and kept in a vacuum desiccator until same-day specimen pycnometer (Quanta Chrome, Model # MPY-2, Quantachrome Instruments; Boynton Beach, FL, USA) volumetric measurements were completed. Molar concentrations of bis-GMA and TEG-DMA (Cbis-GMA and CTEGDMA) in each co-monomer were calculated by formula (1): CM = 1000(C% w/w × mspecimen)/(100 × mM × Vspecimen) where CM is the molar concentration of a monomer (mol/L), C%w/w is the mass percentage of a monomer in a co-monomer mixture, m specimen is the specimen mass (g), mM is molecular weight of a monomer (g/ mol), and Vspecimen is the volume of a specimen (m3). Vol 16, No 5, 2014
Table 1 Resin co-monomer chemical composition (wt%) Co-monomers
Components
Co-monomer 1
98.65% TEG-DMA/ 1.1% EDMAB/ 0.25% CQ
Co-monomer 4
30.90% bis-GMA/ 67.77% TEG-DMA/ 1.08% EDMAB/ 0.25% CQ
Co-monomer 8 (Resin 2)
71.04% bis-GMA/ 27.71% TEG-DMA/ 1.01% EDMAB/ 0.24% CQ
Co-monomer 11
98.71% bis-GMA/ 1.00% EDMAB/ 0.29% CQ
Abbreviations: bis-GMA: 2,2-bis[4-(2-hydroxy-3-methacryloyloxypropoxy)]phenyl propane; TEG-DMA: triethylene glycol dimethacrylate; EDMAB: 2-ethyl dimethyl-4-aminobenzoate; CQ: camphorquinone (Sigma-Aldrich; St Louis, MO, USA).
Note that 1000 is the conversion factor between volume units “L” and “m3”, and 100 is the conversion factor between mass percentage and mass fraction. To avoid the oxygen-inhibited layer, Raman spectra were collected at 3 random spots on the bottom of each resin block using a light microscope (RamanRxnl microprobe, Leica DMPL, Leica Microsystems; Buffalo Grove, IL, USA) attached to a laser generator (Invictus 785-nm NIR Laser, #0666135 single-mode excitation fiber, Kaiser Optical System; Ann Arbor, MI, USA) and a modular research Raman spectrograph (HoloLab series 5000R, #6000131 62.5-μm collection fiber, Kaiser Optical System). Calibration was performed on a silicon standard following the manufacturer’s instruction using a calibration accessory (HCA) with a single mode excitation fiber (#6000135) and 100-μm collection fiber (#6000113; all Kaiser Optical System). The laser beam delivered through a 10X objective to the sample had an intensity of 8 to 11 mW, and a spectral resolution of 4 cm-1 in the 100 to 3450 cm-1 region. Spectra were obtained through a 100X objective with a 60-s exposure using auto-dark subtraction and cosmic-ray correction via the manufacturer’s software (HoloGRAMS, v. 4.1, Kaiser Optical System). The peak area measurement at 1610 cm-1 (aromatic ring on bis-GMA) and 605 cm-1 (C-C=O bend in methacrylate groups on bis-GMA and TEG-DMA) was performed using Raman data processing software (HoloMap, v. 7.3.0 [R2006b], Kaiser Optical System). The starting and end points of each peak on an actual spectrum were selected based on its second derivative spectrum. Mean peak areas at 605 and 1610 cm-1, bis-GMA and TEG-DMA molar concentrations (Cbis-GMA and CTEG-DMA) were used to construct calibration curves. Resin co-monomer 8 (Biomaterials lab, University of Iowa), with the same formula as resin 2 (Bisco), was used to fabricate resin-dentin bond specimens, and thus the known Cbis-GMA in this co-monomer was used as a reference to calculate short-term Cbis-GMA and CTEG-DMA across hybrid layers. Resin-Dentin Bond Specimen Preparation Eight sound human molars were prepared as described to expose coronal dentin. Sticky wax was used to build a 2-mm wall on top of the midline section to separate each 443
Talungchit et al
Normalized 1610 Peak Area
1.2 1.0
y = 0.3727x + 0.1134 R2 = 0.9939
0.8 0.6 0.4
Fig 1 Calibration curve I constructed from bis-GMA molar concentration (C bis-GMA ) and normalized peak area of experimental co-monomer resin blocks obtained by dividing the mean 1610 cm-1 peak area of each co-monomer by the mean 1610 cm-1 peak area of co-monomer 11.
0.2 0.0 0.0
0.5
1.0
1.5
2.0
2.5
4.5
2.5
4.0 y = –0.2911x2 + 1.7715x – 0.3201 R2 = 1
3.5 3.0
1.5
2.5 2.0
1.0
1.5 y = –0.5096x2 – 3.0955x + 4.6958 R2 = 1
1.0
0.5
0.5 0.0
0.0 0
2
4 Ratio 1610/605
6 TEG-DMA
tooth into halves that were randomly treated with EW and WW. The same EW and WW bonding protocols were followed as previously described, with the exception that CHX treatment was not included to avoid Raman spectral interference from CHX at 1610 cm-1.18 Two 1-mm-thick resin-dentin slabs containing both EW- and WW-treated surfaces were formed per tooth and polished. On each side of the slabs, two locations with 3-μm-thick hybrid layers were randomly selected on EW- and WW-treated surfaces. At each location, Raman spectra were obtained at 5 positions (Table 4) across resin/dentin interfaces (n = 64 per group) through a 100X objective lens using 8 to 11 mW from the 785-nm laser (1.5 μm spot size) for 60 s (Raman microprobe, Kaiser Optical System). An extra resin-dentin slab was used to specify the time for re-collecting data based on degree of surface deposition and Raman signal strength. Resin-dentin slabs were stored for 7 months as described above, before re-collecting data at the same positions. Monomer Molar Concentration Calculation Absolute molar concentration of each co-monomer was determined using an adaptation of the method reported by Zou et al.37 444
2.0
Bis-GMA Molar Concentration [mol/L]
TEG-DMA Molar Concentration [mol/L]
Bis-GMA Molar Concentration [mol/L]
Bis-GMA
Fig 2 Calibration curve II constructed from bis-GMA and TEG-DMA molar concentrations (Cbis-GMA and CTEG-DMA) and 1610:605 peak area ratio obtained from resin blocks of the experimental co-monomers.
Absolute bis-GMA molar concentration For the short-term data, Cbis-GMA in the adhesive resin (AR) layer (Position 1, Table 4) was assumed to be equal to that in co-monomer 8 (1.689 mol/L) and was used as a reference. The peak area at 1610 cm-1 from each spot within a scanned location was divided by that at Position 1 in the same location to obtain a normalized peak area at 1610 cm-1. From Calibration Curve I (Fig 1), the equation for the best fit line can be written for any position in the line scan (formula 2), as well as the reference point at Position 1 (formula 3). By dividing formula 2 by formula 3, formula 4 results (see below). Since Cbis-GMA in the resin adhesive layer (Position 1) was assumed equal to 1.689 mol/L (x2), the normalized 1610 cm-1 peak area at Position 1 was equal to 1 (y2). The absolute bis-GMA molar concentration ([bis-GMA]abs) at each spot (x1) could be obtained from formula 5 using these values along with the normalized 1610 cm-1 peak area at each spot (y1). (2) y1 = 0.3727x1 + 0.1134 (3) y2 = 0.3727x2 + 0.1134 From formula (2)/(3), y1/y2 = (0.3727x1 + 0.1134)/(0.3727x2 + 0.1134) x2 = 1.689 and y2 = 1; therefore,
(4)
The Journal of Adhesive Dentistry
Talungchit et al
Table 2 Bond strength and fractography by storage time, substrate, and CHX treatment Acid-etched dentin surface condition (n = 32) WW-with-CHX
No. of teeth redone due to PTFs 17
WW-no-CHX EW-with-CHX EW-no-CHX
1
Immediate result mean (SD) (MPa) [CV]
1-year result mean (SD) (MPa) [CV]
Fractography (%) at 1 year M
J
CR
CD
32.8(11.2)a [34%]
29.8(11.8)a [40%]
3.1
90.6 (¥81.3)
0
6.3
29.9(15.1)a [51%]
26.2(11.0)a [42%]
6.3
93.8 (¥84.4)
0
0
47.5(14.0)a [30%]
45.1(12.4)a [28%]
12.5
62.5 (¥18.8)
0
25.0
51.5(10.2)a [20%]
46.0(11.3)a [49%]
31.6
46.9 (¥28.1)
3.1
18.8
Within a row, groups with the same lowercase letters are not significantly different using a paired-sample t-test. Solid brackets show significant difference using two-sample t-test. Dotted brackets show no significant difference using a paired-sample t-test. Abbreviation: CHX: chlorhexidine; CV: coefficient of variation; PTFs: pre-test failures; M: mixed failure; J: joint failure; CR: resin-based composite cohesive failure; CD: dentin cohesive failure; ¥percentage of specimens fractured solely at the top of hybrid layers (short-term fractography followed the same trend).
y1 = (0.3727x1 + 0.1134)/((0.3727x1.689)+ 0.1134) (5) x1 = (0.743y1 - 0.1134)/0.3727 Since monomer dissolution was expected during storage, the assumption of equal C bis-GMA between that in the resin block and in the resin adhesive layer of resin-dentin bond specimens was no longer valid for the long-term experiment. Therefore, a known Cbis-GMA from a resin block was used as a reference for the long term. The normalized 1610 cm-1 peak area was calculated by dividing the 1610 cm-1 peak area at each scanned spot by the mean 1610 cm-1 peak area obtained from the resin block at the beginning and the end of the experiment on that day. For the long-term experiment, x2 and y2 in formula 4 referred to Cbis-GMA in the resin block, which was equal to 1.689 mol/L, and the normalized 1610 cm-1 peak area of the resin block, which was equal to 1, respectively. Since x2 and y2 in formula 4 were the same value as those at short-term, the rest of the calculation was the same as in the short-term. The long-term absolute bis-GMA molar concentration (x1) at each scanned spot was obtained by using the normalized 1610 cm-1 peak area at each spot (y1) in formula 4. Absolute TEG-DMA molar concentration calculation For both short- and long-term data, the 1610:605 ratio at each spot (x3) was used in formulas 6 and 7 (the best fit lines from Calibration Curve II, Fig 2) to obtain relative bis-GMA molar concentration (y3 in formula 6) and relative TEG-DMA molar concentration (y4 in formula 7). (6) y3 = -0.2911(x3)2 + 1.7715x3 - 0.3201 (7) y4 = 0.5096(x3)2 - 3.0955x3 + 4.6958 Then, the absolute bis-GMA molar concentration (x 1) from formula 5 and relative bis-GMA (y3) and TEG-DMA (y4) molar concentration from formulas 6 and 7, respectively, were used in formula 8 to obtain absolute TEGDMA molar concentration (x4). (8) x4 = x1y4/y3 Vol 16, No 5, 2014
Statistical Analyses Statistical methods are described in Tables 2 to 4. Additionally, the Weibull regression model was used to determine effects of storage time, type of bonding protocol, and their interaction on μTBS. Statistical significance was set at p < 0.05, and marginal significance was defined as 0.05
Purpose: To directly test the effectiveness of ethanol-wet bonding (EW) in improving monomer infiltration into demineralized dentin through quantitative measurement of bis-GMA and TEG-DMA molar concentrations within hybrid layers, and to comprehensively evaluate the effect of EW and chlorhexidine on durability of resin-dentin bonds compared to conventional water-wet bonding (WW). Materials and Methods: A three-step etch-and-rinse adhesive (70% bis-GMA/28.75%TEG-DMA) was applied to coronal dentin using a clinically relevant ethanol-wet bonding protocol (EW) or the conventional water-wet bonding (WW) technique. Bis-GMA and TEG-DMA molar concentrations at various positions across the resin/dentin interfaces formed by EW and WW were measured using micro-Raman spectroscopy. The experiment was repeated at the same positions after 7-month storage in phosphate buffer solution containing 0.1% sodium azide. The μTBS and hybrid layer morphology (TEM) of bonding groups with and without chlorhexidine application were compared immediately and after 1-year storage in terms of nanoleakage, collagen fibril diameter, collagen interfibrillar width, and hybrid layer thickness. Results: Specimens bonded with EW showed significantly higher monomer molar concentrations and μTBS throughout the hybrid layer immediately and after storage, providing direct evidence of superior infiltration of hydrophobic monomers in EW compared to WW. Microscopically, EW maintained interfibrillar width and hybrid layer thickness for resin infiltration and retention. The application of chlorhexidine further preserved collagen integrity and limited the degree of nanoleakage in EW after 1-year storage. Conclusion: EW enhances infiltration of hydrophobic monomers into demineralized dentin. The results suggest that a more durable resin-dentin bond may be achieved with combined usage of a clinically relevant EW and chlorhexidine. Keywords: bonding, collagen(s), matrix metalloproteinases (MMPs), ultrastructure, dentin, adhesives. J Adhes Dent 2014; 16: 441–450. doi: 10.3290/j.jad.a32695
Submitted for publication: 23.10.13; accepted for publication: 17.4.14
a
Full-time Faculty Member, Department of Conservative Dentistry, Prince of Songkla University, Songkhla, Thailand. Hypotheses, experimental design, performed the experiments in partial fulfillment of requirements for PhD, wrote the manuscript, contributed substantially to discussion.
f
Emeritus Regents’ Professor of Oral Biology, Department of Oral Biology, College of Dental Medicine, Georgia Regents University, Augusta, GA, USA. Idea, consulted on ethanol-wet bonding and chlorhexidine, proofread the manuscript, contributed substantially to discussion.
b
Associate Professor, Department of Chemical and Biochemical Engineering, College of Engineering, University of Iowa, Iowa City, IA, USA. Idea, consulted on the experiment using Raman spectroscopy, proofread the manuscript, contributed substantially to discussion.
g
Professor and Chair, Department of Operative Dentistry, College of Dentistry, University of Iowa, Iowa City, IA, USA. Idea, hypotheses, experimental design, proofread the manuscript, contributed substantially to discussion.
c
Associate Professor, Department of Operative Dentistry, College of Dentistry, University of Iowa, Iowa City, IA, USA. Proofread the manuscript, contributed substantially to discussion.
Correspondence: Professor Steven R. Armstrong, Department of Operative Dentistry, College of Dentistry, University of Iowa, S229A DSB, Iowa City, IA, USA 52242. Tel: +1-319-335-7211, Fax: +1-319-335-7267. e-mail: [email protected]
d
Associate Research Scientist, Department of Preventive and Community Dentistry, College of Dentistry, University of Iowa, Iowa City, IA, USA. Performed statistical analyses, interpreted statistical findings, proofread the manuscript.
e
Clinical Associate Professor, Department of Restorative Dental Sciences, College of Dentistry, University of Florida, Gainesville, FL, USA. Idea, consulted on the experiment using TEM, proofread the manuscript, contributed substantially to discussion.
Vol 16, No 5, 2014
Parts of this study were presented at the IADR 86th General Session and Exhibition, Toronto, Canada, 2008, Abstract #0356 (Poster session), IADR 87th General Session and Exhibition, Miami, FL, USA, 2009, Abstract #1517 (Oral presentation session), and IADR 91st General Session and Exhibition, Seattle, WA, USA, 2013, Abstract #1668 (Oral presentation session).
441
Talungchit et al
B
onding effectiveness to dentin relies on resin infiltration and encapsulation of the demineralized collagen fibril network, creating a hybrid layer.23 Deterioration of the hybrid layer is a consequence of hydrolysis of the polymer matrix and degradation of unencapsulated demineralized collagen.2,20 Therefore, enhanced resin infiltration and sustained integrity of both resin matrix and collagen are crucial for a durable resin-dentin bond. Hydrophobic monomers are preferable to hydrophilic monomers to prevent water sorption and hydrolytic degradation of the resin matrix.16 However, hydrophobic monomers fail to infiltrate hydrophilic water-saturated demineralized dentin using the conventional water-wet bonding technique (WW).1 The ethanol-wet bonding technique (EW) replaces water in demineralized dentin with ethanol, thus increasing miscibility and solubility of subsequently applied hydrophobic resin adhesives, yielding strong and durable bonds.28 Ethanol, a transitional solvent, maintains interfibrillar width by preventing interpeptide H-bonding in collagen, thereby preventing the collapse of the collagen matrix during solvent evaporation and resin adhesive application.25 Various EW protocols have been used in recent studies.15,26,29 However, the results were non-uniform, and there is no study which quantitatively measured and compared the amount of dimethacrylate monomers, ie, bis-GMA and TEGDMA, in hybrid layers formed by EW and WW. Once activated, host-derived matrix metalloproteinases (MMPs)22 and cysteine cathepsins24 slowly degrade unencapsulated collagen within hybrid layers, compromising resin-dentin bond durability. Chlorhexidine (CHX) was introduced in bonding to inhibit proteolytic degradation.10,32 The use of CHX with WW produced a more stable collagen morphology and better bond strength compared to no CHX both in vivo13 and in vitro.5,7 Although limited amounts of denuded collagen are expected when using EW, CHX may still be beneficial in preserving collagen integrity. The purpose of this study was to evaluate the shortand long-term (1-year) effects of EW and CHX on resindentin microtensile bond strength (μTBS) and hybrid layer quality as measured by nanoleakage, collagen fibril diameter, interfibrillar width, hybrid layer thickness, and the presence of collagen banding using TEM. Micro-Raman spectroscopy was used to evaluate monomer infiltration in both EW and WW specimens by measuring the bis-GMA and TEG-DMA molar concentrations (CM) across hybrid layers immediately and after 7-month storage in phosphate buffer solution. The null hypotheses were that EW and CHX had no effect on resin-dentin μTBS nor on hybrid layer quality and that EW had no effect on CM within hybrid layers compared to WW.
MATERIALS AND METHODS Microtensile Bond Strength Test Institutional Review Board (IRB)-approved, sound human molars were stored in 0.5% chloramine-T at 4°C and used within 6 months of extraction. Teeth were rinsed free of chloramine-T and stored in PBS before being randomly divided into WW and EW groups (n = 32 per 442
group) and ground with a carbide bur (#55, Brasseler; Savannah, GA, USA) in a computer numerical control (CNC) specimen former (University of Iowa, Iowa City, IA, USA) to expose a flat surface of coronal dentin. A 0.5-mm-deep groove was placed across the occlusal surface to allow separate surface treatments on the two halves. The three-step etch-and-rinse technique was applied for WW. For EW, the acid-etched and waterrinsed dentin was subsequently rinsed with absolute ethanol for 15 s three times with blotting in between. Before priming, half of each tooth in WW and EW was treated for 30 s with 2% aqueous chlorhexidine digluconate (Cavity Cleanser, Bisco; Schaumburg, IL, USA) or 2% chlorhexidine diacetate in ethanol (Pashley Lab; Augusta, GA, USA), respectively, creating 4 bonding protocols: EW-no-CHX, EW-with-CHX, WW-no-CHX, and WW-with-CHX. Then, primer (50% adhesive/50% ethanol) was applied for 30 s. After solvent removal, experimental nonsolvated adhesive (70% bis-GMA/28.75% TEGDMA; Bisco) was applied and light cured (Demetron 501, irradiance 700-780 mW/cm2, Kerr; Orange, CA, USA) for 20 s. Resin-based composite (Z100, 3M ESPE; St Paul, MN, USA) was placed and light cured for 40 s in three 1.5-mm-thick increments. Four dumbbell-shaped specimens with a circular crosssectional area of 0.5 mm2 were formed per tooth using the CNC specimen former. Any tooth failing to produce 4 testable specimens was replaced with a different tooth. One beam from each side of the tooth, with and without CHX, was immediately tensile tested to failure at 1 mm/ min in a passive gripping device3 (n = 32 per bonding protocol). The remaining specimens were stored at 37°C in phosphate buffer solution containing 0.1% sodium azide (n = 32 per bonding protocol), which was changed monthly until testing after 1-year storage. Fractography After the μTBS test, debonded specimens were glued to aluminum stubs using a cyanoacrylate cement (Zapit, Dental Ventures of America; Corona, CA, USA) and then gold sputter coated for evaluation with SEM (Amray 1820-D, AMRAY; Bedford, MA, USA). Fracture pathways were classified into 4 types: dentin cohesive failure, resin-based composite cohesive failure, joint failure, and mixed failure. Hybrid Layer Morphology Six additional teeth were randomly divided into WW and EW groups (n = 3/per group). Within each group, the teeth were bonded as described above, except flowable resin-based composite (Heliomolar Flow, Ivoclar Vivadent; Amherst, NY, USA) was used to facilitate sectioning for ultrastructural evaluations using TEM (JEOL 1230, 120 kV; Tokyo, Japan). Four 0.5 × 1 mm2 resindentin beams from each half of each tooth were randomly selected to evaluate nanoleakage and collagen in the short term and after 1-year storage. For the shortterm evaluations, two beams were placed overnight in Karnovsky’s fixative, one for nanoleakage and another for collagen ultrastructure. The remaining beams were The Journal of Adhesive Dentistry
Talungchit et al
stored at 37ºC in phosphate buffer solution containing 0.1% sodium azide for long-term study (1 year). For nanoleakage evaluation, specimens were stained before sectioning with 50% ammoniacal silver nitrate for 24 h, then placed in developer under fluorescent light for 8 h. Specimen coding was used for blinded evaluation. Dehydration, resin embedding, and thin sectioning were performed. For nanoleakage types, five 90-nm-thick sections from each specimen were evaluated (n = 15 per bonding protocol). After sectioning, for collagen ultrastructural evaluation, 5 non-demineralized 60-nm-thick sections from each specimen were double-stained in 1% phosphotungstic acid for 10 min and 5% uranyl acetate for 2 min. TEM images were randomly taken and measurements were performed using ImageJ (National Institutes of Health; Bethesda, MD, USA). Hybrid layer thickness was randomly measured at one spot per section (n = 15 per bonding protocol). Collagen fibril diameter and interfibrillar width within a 5 × 5 μm2 viewing area at upper, middle, and lower hybrid-layer regions were measured at 10,000X magnification in one area per section. Dentinal tubules and their branches were avoided. In tangential sections where the fibril profiles were oblong, the width of the fibril was taken to be the diameter. Within each viewing area, all cross-sectioned collagen fibrils were measured for collagen fibril diameter (3 to 10 fibrils per viewing area), and the same number of interfibrillar width measurements were performed. The mean of each measurement within each viewing area was calculated. Three-way ANOVA was performed to assess the significance of hybrid layer region on the mean collagen fibril diameter and mean interfibrillar width. Since no significant effect was found, mean collagen fibril diameter and interfibrillar width within each viewing area was assumed to be an independent statistical unit (n = 45 per bonding protocol). The presence of collagen banding was defined as 67 nm “collagen banding” if more than 25% of fibril banding was present under the TEM viewing area (n = 45 per bonding protocol). Monomer Infiltration (Micro-Raman Spectroscopy) Experimental co-monomers (Biomaterials lab, College of Dentistry, University of Iowa) with different bis-GMA/ TEG-DMA concentrations (Table 1) were prepared to construct calibration curves (Figs 1 and 2) for CM calculation. Each co-monomer was placed in 3 aluminum pans and polymerized for 60 s (Demetron Optilux 500, irradiance 700 to 800 mW/cm2, Kerr) to create resin blocks. Polymer samples were removed from aluminum pans and kept in a vacuum desiccator until same-day specimen pycnometer (Quanta Chrome, Model # MPY-2, Quantachrome Instruments; Boynton Beach, FL, USA) volumetric measurements were completed. Molar concentrations of bis-GMA and TEG-DMA (Cbis-GMA and CTEGDMA) in each co-monomer were calculated by formula (1): CM = 1000(C% w/w × mspecimen)/(100 × mM × Vspecimen) where CM is the molar concentration of a monomer (mol/L), C%w/w is the mass percentage of a monomer in a co-monomer mixture, m specimen is the specimen mass (g), mM is molecular weight of a monomer (g/ mol), and Vspecimen is the volume of a specimen (m3). Vol 16, No 5, 2014
Table 1 Resin co-monomer chemical composition (wt%) Co-monomers
Components
Co-monomer 1
98.65% TEG-DMA/ 1.1% EDMAB/ 0.25% CQ
Co-monomer 4
30.90% bis-GMA/ 67.77% TEG-DMA/ 1.08% EDMAB/ 0.25% CQ
Co-monomer 8 (Resin 2)
71.04% bis-GMA/ 27.71% TEG-DMA/ 1.01% EDMAB/ 0.24% CQ
Co-monomer 11
98.71% bis-GMA/ 1.00% EDMAB/ 0.29% CQ
Abbreviations: bis-GMA: 2,2-bis[4-(2-hydroxy-3-methacryloyloxypropoxy)]phenyl propane; TEG-DMA: triethylene glycol dimethacrylate; EDMAB: 2-ethyl dimethyl-4-aminobenzoate; CQ: camphorquinone (Sigma-Aldrich; St Louis, MO, USA).
Note that 1000 is the conversion factor between volume units “L” and “m3”, and 100 is the conversion factor between mass percentage and mass fraction. To avoid the oxygen-inhibited layer, Raman spectra were collected at 3 random spots on the bottom of each resin block using a light microscope (RamanRxnl microprobe, Leica DMPL, Leica Microsystems; Buffalo Grove, IL, USA) attached to a laser generator (Invictus 785-nm NIR Laser, #0666135 single-mode excitation fiber, Kaiser Optical System; Ann Arbor, MI, USA) and a modular research Raman spectrograph (HoloLab series 5000R, #6000131 62.5-μm collection fiber, Kaiser Optical System). Calibration was performed on a silicon standard following the manufacturer’s instruction using a calibration accessory (HCA) with a single mode excitation fiber (#6000135) and 100-μm collection fiber (#6000113; all Kaiser Optical System). The laser beam delivered through a 10X objective to the sample had an intensity of 8 to 11 mW, and a spectral resolution of 4 cm-1 in the 100 to 3450 cm-1 region. Spectra were obtained through a 100X objective with a 60-s exposure using auto-dark subtraction and cosmic-ray correction via the manufacturer’s software (HoloGRAMS, v. 4.1, Kaiser Optical System). The peak area measurement at 1610 cm-1 (aromatic ring on bis-GMA) and 605 cm-1 (C-C=O bend in methacrylate groups on bis-GMA and TEG-DMA) was performed using Raman data processing software (HoloMap, v. 7.3.0 [R2006b], Kaiser Optical System). The starting and end points of each peak on an actual spectrum were selected based on its second derivative spectrum. Mean peak areas at 605 and 1610 cm-1, bis-GMA and TEG-DMA molar concentrations (Cbis-GMA and CTEG-DMA) were used to construct calibration curves. Resin co-monomer 8 (Biomaterials lab, University of Iowa), with the same formula as resin 2 (Bisco), was used to fabricate resin-dentin bond specimens, and thus the known Cbis-GMA in this co-monomer was used as a reference to calculate short-term Cbis-GMA and CTEG-DMA across hybrid layers. Resin-Dentin Bond Specimen Preparation Eight sound human molars were prepared as described to expose coronal dentin. Sticky wax was used to build a 2-mm wall on top of the midline section to separate each 443
Talungchit et al
Normalized 1610 Peak Area
1.2 1.0
y = 0.3727x + 0.1134 R2 = 0.9939
0.8 0.6 0.4
Fig 1 Calibration curve I constructed from bis-GMA molar concentration (C bis-GMA ) and normalized peak area of experimental co-monomer resin blocks obtained by dividing the mean 1610 cm-1 peak area of each co-monomer by the mean 1610 cm-1 peak area of co-monomer 11.
0.2 0.0 0.0
0.5
1.0
1.5
2.0
2.5
4.5
2.5
4.0 y = –0.2911x2 + 1.7715x – 0.3201 R2 = 1
3.5 3.0
1.5
2.5 2.0
1.0
1.5 y = –0.5096x2 – 3.0955x + 4.6958 R2 = 1
1.0
0.5
0.5 0.0
0.0 0
2
4 Ratio 1610/605
6 TEG-DMA
tooth into halves that were randomly treated with EW and WW. The same EW and WW bonding protocols were followed as previously described, with the exception that CHX treatment was not included to avoid Raman spectral interference from CHX at 1610 cm-1.18 Two 1-mm-thick resin-dentin slabs containing both EW- and WW-treated surfaces were formed per tooth and polished. On each side of the slabs, two locations with 3-μm-thick hybrid layers were randomly selected on EW- and WW-treated surfaces. At each location, Raman spectra were obtained at 5 positions (Table 4) across resin/dentin interfaces (n = 64 per group) through a 100X objective lens using 8 to 11 mW from the 785-nm laser (1.5 μm spot size) for 60 s (Raman microprobe, Kaiser Optical System). An extra resin-dentin slab was used to specify the time for re-collecting data based on degree of surface deposition and Raman signal strength. Resin-dentin slabs were stored for 7 months as described above, before re-collecting data at the same positions. Monomer Molar Concentration Calculation Absolute molar concentration of each co-monomer was determined using an adaptation of the method reported by Zou et al.37 444
2.0
Bis-GMA Molar Concentration [mol/L]
TEG-DMA Molar Concentration [mol/L]
Bis-GMA Molar Concentration [mol/L]
Bis-GMA
Fig 2 Calibration curve II constructed from bis-GMA and TEG-DMA molar concentrations (Cbis-GMA and CTEG-DMA) and 1610:605 peak area ratio obtained from resin blocks of the experimental co-monomers.
Absolute bis-GMA molar concentration For the short-term data, Cbis-GMA in the adhesive resin (AR) layer (Position 1, Table 4) was assumed to be equal to that in co-monomer 8 (1.689 mol/L) and was used as a reference. The peak area at 1610 cm-1 from each spot within a scanned location was divided by that at Position 1 in the same location to obtain a normalized peak area at 1610 cm-1. From Calibration Curve I (Fig 1), the equation for the best fit line can be written for any position in the line scan (formula 2), as well as the reference point at Position 1 (formula 3). By dividing formula 2 by formula 3, formula 4 results (see below). Since Cbis-GMA in the resin adhesive layer (Position 1) was assumed equal to 1.689 mol/L (x2), the normalized 1610 cm-1 peak area at Position 1 was equal to 1 (y2). The absolute bis-GMA molar concentration ([bis-GMA]abs) at each spot (x1) could be obtained from formula 5 using these values along with the normalized 1610 cm-1 peak area at each spot (y1). (2) y1 = 0.3727x1 + 0.1134 (3) y2 = 0.3727x2 + 0.1134 From formula (2)/(3), y1/y2 = (0.3727x1 + 0.1134)/(0.3727x2 + 0.1134) x2 = 1.689 and y2 = 1; therefore,
(4)
The Journal of Adhesive Dentistry
Talungchit et al
Table 2 Bond strength and fractography by storage time, substrate, and CHX treatment Acid-etched dentin surface condition (n = 32) WW-with-CHX
No. of teeth redone due to PTFs 17
WW-no-CHX EW-with-CHX EW-no-CHX
1
Immediate result mean (SD) (MPa) [CV]
1-year result mean (SD) (MPa) [CV]
Fractography (%) at 1 year M
J
CR
CD
32.8(11.2)a [34%]
29.8(11.8)a [40%]
3.1
90.6 (¥81.3)
0
6.3
29.9(15.1)a [51%]
26.2(11.0)a [42%]
6.3
93.8 (¥84.4)
0
0
47.5(14.0)a [30%]
45.1(12.4)a [28%]
12.5
62.5 (¥18.8)
0
25.0
51.5(10.2)a [20%]
46.0(11.3)a [49%]
31.6
46.9 (¥28.1)
3.1
18.8
Within a row, groups with the same lowercase letters are not significantly different using a paired-sample t-test. Solid brackets show significant difference using two-sample t-test. Dotted brackets show no significant difference using a paired-sample t-test. Abbreviation: CHX: chlorhexidine; CV: coefficient of variation; PTFs: pre-test failures; M: mixed failure; J: joint failure; CR: resin-based composite cohesive failure; CD: dentin cohesive failure; ¥percentage of specimens fractured solely at the top of hybrid layers (short-term fractography followed the same trend).
y1 = (0.3727x1 + 0.1134)/((0.3727x1.689)+ 0.1134) (5) x1 = (0.743y1 - 0.1134)/0.3727 Since monomer dissolution was expected during storage, the assumption of equal C bis-GMA between that in the resin block and in the resin adhesive layer of resin-dentin bond specimens was no longer valid for the long-term experiment. Therefore, a known Cbis-GMA from a resin block was used as a reference for the long term. The normalized 1610 cm-1 peak area was calculated by dividing the 1610 cm-1 peak area at each scanned spot by the mean 1610 cm-1 peak area obtained from the resin block at the beginning and the end of the experiment on that day. For the long-term experiment, x2 and y2 in formula 4 referred to Cbis-GMA in the resin block, which was equal to 1.689 mol/L, and the normalized 1610 cm-1 peak area of the resin block, which was equal to 1, respectively. Since x2 and y2 in formula 4 were the same value as those at short-term, the rest of the calculation was the same as in the short-term. The long-term absolute bis-GMA molar concentration (x1) at each scanned spot was obtained by using the normalized 1610 cm-1 peak area at each spot (y1) in formula 4. Absolute TEG-DMA molar concentration calculation For both short- and long-term data, the 1610:605 ratio at each spot (x3) was used in formulas 6 and 7 (the best fit lines from Calibration Curve II, Fig 2) to obtain relative bis-GMA molar concentration (y3 in formula 6) and relative TEG-DMA molar concentration (y4 in formula 7). (6) y3 = -0.2911(x3)2 + 1.7715x3 - 0.3201 (7) y4 = 0.5096(x3)2 - 3.0955x3 + 4.6958 Then, the absolute bis-GMA molar concentration (x 1) from formula 5 and relative bis-GMA (y3) and TEG-DMA (y4) molar concentration from formulas 6 and 7, respectively, were used in formula 8 to obtain absolute TEGDMA molar concentration (x4). (8) x4 = x1y4/y3 Vol 16, No 5, 2014
Statistical Analyses Statistical methods are described in Tables 2 to 4. Additionally, the Weibull regression model was used to determine effects of storage time, type of bonding protocol, and their interaction on μTBS. Statistical significance was set at p < 0.05, and marginal significance was defined as 0.05
