Seventy-two, caries-free, recently extracted human third molars were stored in a 0.5% chloramines T aqueous solution at 4℃ and used within 1 month following extraction. A flat dentin surface in mid-coronal portion of tooth was prepared perpendicular to the longitudinal axis of each tooth using a slow-speed Isomet diamond saw (Buehler Ltd., Lake Bluff, IL, USA) under copious water-cooling. Each dentin surfaces were then polished with a 600-grit silicon carbide paper under running water to creat a bonding surface with a standardized smear layer in mid-coronal dentine that was devoid of enamel. The teeth were randomly divided into two groups (N=36) according to the dentin adhesives used: Single Bond 2 (3M ESPE, St. Paul, MN, USA), a water-containing, ethanol-based etch-and-rinse adhesive, and One Step plus (Bisco Inc., Schaumburg, IL, USA), a water-free, acetone-based etch-and rinse adhesive. Composition, batch number and manufacturers of the tested adhesives are listed in Table 1. Each dentin surface to be bonded was etched with a 37% phosphoric acid gel (3M ESPE, St. Paul, MN) for 15 s to create a 5-8 µm thick zone of completely demineralized dentin. The etched dentin surface was then rinsed with deionized water for 10 s and kept visibly moist to prevent collapse of the collagen matrix. Prior to the bonding procedures, the specimens of each group were randomly divided into two subgroups according to solution that was used for the pretreatment of acid-etched dentin surfaces: Half of the specimens in each group were immersed in 0.1 U/mL chondroitinase ABC (C-ABC) (Sigma chemical, St. Lois, MO.) in 0.1 M Tris/Acetate (pH 7.5) for 48 h (renewed after 24 h) at 37℃, while the other half were stored in distilled water under the same conditions. After incubation for 48 h, all specimens were thoroughly rinsed with distilled water for 5 min and bonded with the respective adhesive using three different bonding techniques (N=3 for each subgroup): I) moist dentin was blot dried; II) moist dentin was air dried for 15 s; and III) moist dentin was air dried for 15 s, re-wetted for 30 s and subsequently blot dried. The dentin adhesives were used in accordance with manufacturer's instructions. This was followed by incremental placement of a composite resin (Z-250, 3M ESPE, St. Paul, MN, USA) to a height of 5 mm. Light curing was performed with a Spectrum 800 halogen light-curing unit (Dentsply DeTrey, Konstanz, Germany) with power output of 600 mW/cm².

After water-storage at 37℃ for 24 h, each tooth was sectioned occluso-gingivally into 1 mm thick serial slabs (N=6 for each teeth) by means of the slow speed diamond saw with water-cooling. The central slab was used for SEM examination, while the others were further sectioned into 1 × 1 mm beams, using the "non-trimming"technique of the microtensile testing. The two longest beams from each slab were selected for tensile testing, yielding 30 beams per subgroup. Specimens that failed prematurely during sectioning were recorded in each subgroup, and the beam dimension of each intact specimen was measured with digital caliper to the nearest 0.01 mm to provide the bonding surface area for calculating microtensile bond strength. Each beam was fixed by their ends to the microtensile bond testing device with cyanoacrylate glue (Zapit, Dental Ventures of America, Corona, CA, USA) and tested in tension using material tester (EZ tester, Shimadzu Co., Kyoto, Japan) at a crosshead speed of 1.0 mm/min. Microtensile bond strength was determined in MPa by calculating the quotient of maximum load and the its corresponding bonding area.

Statistical analysis was performed separately for each dentin adhesive since the intention of the study was not compare bond strengths of the two adhesives. Bond strength data was normally distributed (Kolmogorov-Smirnov test) and exhibited equal variances (Levene test). Thus, the data were analyzed using a two-way analysis of variance (ANOVA) for evaluating the effects of enzyme digestion and dentin condition, followed by Duncan test for post-hoc comparisons. Statistical analysis was performed using SPSS 11.0 for window (SPSS Inc., Chicago, IL, USA) with the level of significance set at 5%.

Two representative debonded beams of the mean bond strength of each subgroup were prepared for the observation of the fracture pattern.

Moreover, the central slabs of each specimen were serially polished with 600, 1200, 1500 grit silicon carbide paper under running water, followed by a 5-sec 37% phosphoric acid gel (3M ESPE, St. Paul, MN, USA) treatment and a 10-min 5.25% sodium hypochlorite treatment. The selected beams and slabs were dried for 24 h within the drierite desiccator (Hammond Drierite Company LTD., Xenia, Ohio, USA).

Another eight caries-free, human third molars were used to evaluate change of the collagen architecture after removal of the chondroitin sulfate in proteoglycan. Specimen preparation was carried out according to the method proposed by Perdigao et al.22) in order to avoid artifacts formation due to dehydration shrinkage of specimens and their observation under a high to ultra-high vacuum.21) Fourteen 0.9 mm-thick dentin disks parallel to the occlusal surface were obtained from mid-coronal dentin that was devoid of enamel by low speed diamond saw. A transverse groove was made in the bottom surface of each dentin disk to facilitate splitting after specimen preparation, while its upper surface was etched with a 37% phosphoric acid gel for 15 s, rinsed with water for 10 s and blot dried. All the specimens were randomly distributed into two groups (N=4): one group was then treated with chondroitinase-ABC (C-ABC) and the other group was stored in distilled water as the same way described in the microtensile test section. After C-ABC/distilled water storage, specimens of each group were subdivided into two subgroups (N=2) according to dentin moisture, namely wet, and rewet. The specimens were immediately immersed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer at pH 7.4 for 4 hours at 4℃. After fixation, the disks were rinsed with 0.1 M sodium cacodylate buffer at pH 7.4 for 1 h, with three changes, followed by distilled water for 5 min. After dehydration in ascending ethanol series (50-100%), the specimens were HMDS-dried. Specimens for examination of sagittal surface of acid-etched dentin were split with forceps before mounting the stubs.

After dried, all of SEM specimens were mounted in aluminum stubs with carbon tape. They were sputter-coated with platinum by means of sputter-coating Unit (E-1045, Polaron, Germany) and examined with a FE-SEM (SUPRA 55, Zeiss, Germany) at an acceleration voltage of 5-10 kV.

Premature failure (PF) was observed in all groups of air-dried dentin and Single Bond 2 groups pretreated with C-ABC. The data for microtensile bond strength and percentage of premature failure of each experimental group are summarized in Table 2. For One Step plus, the bond strengths in descending order were wed-bonding subgroups, subgroups using rewetting technique, and dry-bonding subgroups. The same was true when the Single Bond 2 group was taken into consideration. Two-way ANOVA indicated a significant difference in the bond strengths of each dentin adhesives among different moisture conditions of dentin (p < 0.05), whilst there was no significant difference between dentin with and without C-ABC treatment (p > 0.05), and showed that the interaction between dentin condition and C-ABC Treatment was significant for Single Bond 2 only (p = 0.002), not for One Step plus (p > 0.05).

For the subgroups using wet-bonding or dry-bonding technique, microtensile bond strength showed no significant difference after C-ABC treatment (p > 0.05). Nevertheless, the subgroup using rewetting technique after air dry in the Single Bond 2 group demonstrated a significant decrease of microtensile bond strength after C-ABC treatment compared with its counterpart without C-ABC treatment (p < 0.05).

Premature failure (PF) was observed in groups of dry-bonding technique in both tested dentin adhesives and C-ABC pretreated groups in Single Bond 2 (Figure 1).

Representative FE-SEM images of fractured specimens were selected and shown in Figure 2. Images of resin-dentin interface were shown in Figure 3 and image of collagen architecture were shown in Figure 4, 5. For Single Bond 2, failure occurred within the adhesive resin to the bottom of the hybrid layer and dentinal tubules were occluded with the resin tag (Figure 2a). For One Step Plus, failure was occurred within the densely formed hybrid layer (Figure 2c). After C-ABC treatment, failures were occurred at the bottom of hybrid layer with resin tag exclusion from dentinal tubules at both dentin adhesives (Figure 2b, 2d).

Regardless of dentin adhesive, thickness of hybrid layer was increased in C-ABC treated group compared with not treated group when adhesives were used to wet dentin (Figure 3).

Regarding FE-SEM observation with high magnification of the collagen architecture, the collagen of normal acid-etched wet dentin is uniformly arranged and compact interfibrillar space and fully expanded state (Figure 4a, 4b, Figure 5a). After C-ABC pretreatment, collagen architecture is loosely packed and some fibrils are aggregated together and relatively collapsed compared with normal acid-etched wet dentin (Figure 4c, 4d, Figure 5b, 5c).