Preparation of 20% Salvadora persica aqueous extract

A fresh 1 kg of Salvadora persica roots (Miswak sticks) was selected by an expert on the plant and was obtained from a store in Jeddah, Saudi Arabia. The roots were originally collected from Almakhwah, 45 km west of Albahah, south of Saudi Arabia (1947′0″ N, 4126′0″ E). Extract of Salvadora persica was prepared according to a protocol published in previous studies [28, 35, 36]. The roots were cleaned, washed with distilled water, dried for a few days at room temperature, then cut into tiny pieces and ground in a ball mill to produce a powder. A 20 gm of Salvadora persica powder was added to 100 mL of deionized sterile water and was left at 4 °C for 48 h. The mixture was then centrifuged for 10 min at 2200× g rpm, producing a 20% Salvadora persica aqueous extract. Ultraviolet light (Camag UV-Cabinet II, Basel, Switzerland) delivering ∼ 10 mW/cm2 at 366 nm was used for 1 h to sterilize the extract, then it was stored at 4 °C and was used within 1 week from its preparation [28].

Microtensile bond strength test and nanoleakage analysis

A total of 20 extracted sound human third molars were obtained with the informed consent of the donor following a protocol (# 004–15) approved by the Research Ethics Committee, King Abdulaziz University, Jeddah, Saudi Arabia. The teeth were stored for 1 week in 0.5% chloramine T solution (Sigma-Aldrich Co., St. Louis, MO, USA), then in distilled water at 4 °C till use within 2 months following extraction. A low-speed diamond saw (Micromet AG, Munich, Germany) under water coolant was used to cut occlusal enamel and superficial dentin in a direction perpendicular to the long axis of the tooth. The exposed middle dentinal surfaces were polished with wet 320-grit silicon carbide paper (Norton Saint-Gobain Abrasives, Worcester, MA, USA) for 60 s to remove any enamel remnants and create standardized smear layers.

Restorative procedure

The smear layer-covered dentinal surfaces of the 20 teeth used were etched for 15 s using Scotchbond Universal etching gel (3 M ESPE, St. Paul, MN, USA, Lot #: 7,936,326, composition: 30–40 wt% % phosphoric acid, water, silica, polyethylene glycol, aluminium oxide), rinsed with water for 10 s, then blot-dried using cotton pellets and kept moist without pooling following the wet bonding technique. The 20 teeth, with etched dentinal surfaces, were divided into 2 groups (n = 10) according to the dentin pretreatment, which was as follows: Group 1: Control, etched dentinal surfaces were bonded by Adper Single Bond 2 (3 M ESPE, St. Paul, MN, USA, Lot #: NC68418, composition: BISGMA, HEMA, dimethacrylate, methacrylate-modified polyalkenoic acid copolymer, 10 wt% 5 nm silica particles, ethyl alcohol, water, initiators, other name: Adper Scotchbond 1 XT) according to the manufacturer’s instructions, without any dentin pretreatment as follows: two consecutive coats of the adhesive were applied to etched dentin for 15 s with gentle agitation using a fully saturated microsponge (Pearson Dental Supply, CA, USA). Gentle adhesive air-thinning for 5 s was performed then the adhesive was light-cured for 10 s (Light Emitting Diode curing unit, 3 M ESPE Elipar, Seefeld, Germany delivering 1200 mW/cm2 at 430–480 nm); Group 2: Salvadora persica extract was applied generously using a fully saturated microsponge (Pearson Dental Supply, CA, USA) to the etched dentinal surface with gentle agitation for 1 min, water rinsed for 15 s, blot-dried, then was finally bonded with Adper Single Bond 2 adhesive as mentioned in group 1.

Resin-based composite (Filtek Z350 XT, 3 M ESPE, St. Paul, MN, USA, Lot #: NF30942) build-up is performed using the incremental technique where each increment was light-cured (Light Emitting Diode curing unit, 3 M ESPE Elipar, Seefeld, Germany delivering 1200 mW/cm2 at 430–480 nm) for 20 s. All samples were placed in distilled water for 24 h at 37 °C. Each tooth was then cut parallel to its long axis through the resin–dentin interface, producing 16 sticks (1 × 1 mm ± 0.1 mm) per tooth [2, 9, 10, 37], using a Techcut 4 diamond saw (Allied High Tech Products Inc., Compton, CA, USA). The sticks produced from each tooth were stored separately in distilled water at 37 °C. Each group (n = 10) was divided into 2 subgroups according to the storage interval (n = 5): 24 h and 6 m.

Microtensile bond strength test

After storage either for 24 h or 6 m, 15 sticks from each of the 5 teeth of each subgroup were used for the microtensile bond strength test. A digital caliper (Dasqua tools, Chengdu, China) was used to record the dimensions of each stick accurately to 0.01 mm. A microtensile tester (Bisco Inc., Schaumburg, IL, USA) was used to apply a tensile force at the bonded interface of each stick at a crosshead speed of 1 mm/min until failure. Microtensile bond strength (MPa) was calculated for each stick. The average of the bond strength values of the 15 sticks of each tooth was calculated to provide a single value for each tooth. Then, the values of the 5 teeth of each subgroup were averaged to provide a grand mean for each subgroup (the statistical unit is the tooth) [9, 10]. Fracture mode of all debonded specimens was examined at 50x magnification using a stereomicroscope (Meiji Techno Co., Ltd., Tokyo, Japan) and was classified into adhesive (A), cohesive in dentin (CD), cohesive in composite (CC) or mixed (M) failures.

Nanoleakage evaluation

One stick from each tooth of each subgroup was used for nanoleakage analysis (n = 5). The sticks were prepared according to the technique described by Tay et al. [38]. The sticks were completely coated by 2 layers of nail varnish (Shenzhen Meixin Industry Co., Ltd., Guandong, China), except for 1 mm at the interface, which was left uncoated. A solution of ammoniacal silver nitrate (50 wt%, pH 9.5) was used to soak the sticks for 24 h, followed by rinsing, and then placed for 8 h in a photodeveloper (Kodak Professional T-Max developer, Kodak Alaris Inc., Rochester, NY 14,615, USA) under fluorescent light. The sticks were polished with wet silicon carbide papers (Norton Saint-Gobain Abrasives, Worcester, MA, USA) of increasing fineness (600–1200-grit), followed by soft polishing cloth with 0.05 mm alumina particles suspension (Buehler, Lake Bluff, IL, USA) and then ultrasonically cleaned (Ultrasonic Cleaning System 2014, L&R Manufacturing, Kearny, NJ, USA) for 5 min. Scanning electron microscopy of resin–dentin interfaces was performed (SEM; Quanta 200 ESEM, FEI France, Mérignac, France) using backscattered mode at 1000x. The amount of silver nitrate precipitated within the interface was quantitated using image analysis software (Version 1.32 NIH Image, Scion Corp., Fredrick, MD, USA) [10, 38]. Figure 1 shows the experimental design for microtensile bond strength test and nanoleakage analysis.

Fig. 1

Experimental design for microtensile bond strength test and nanoleakage analysis. For bond strength test, there were 2 groups (control and Salvadora persica). Each group contained 10 teeth, which were divided into 2 subgroups (24 h and 6 m). Each subgroup contained 5 teeth or 75 (5 × 15) sticks. The values of the 15 sticks per tooth were averaged to provide a single value for each tooth. Then, a grand mean was obtained from the values of the 5 teeth of each subgroup (the tooth is the statistical unit). For nanoleakage test, one stick from each tooth of each subgroup was used for nanoleakage analysis (n = 5). Two images were selected from each stick, where the amount of silver nitrate was calculated in each image then averaged to give a single nanoleakage value for each stick. The nanoleakage values of the 5 sticks per subgroup were averaged to provide a mean nanoleakage % value for each time period in each group

Stiffness and hydroxyproline release tests

A total of 40 third molar teeth were selected as discussed before and were sectioned to provide 0.5 mm thick discs of mid coronal dentin, which were further cut into 40 dentin sticks (width = 3 mm, length = 6 mm). The sticks were placed in 10% phosphoric acid in vibrating sealed bottles for 18 h at 4 °C until they were completely demineralized and then rinsed with distilled water for 2 h. The modulus of elasticity of each stick was measured and was set at 5 MPa to ensure complete dem ineralization [39]. The 40 completely demineralized dentin sticks were equally and randomly divided into 2 groups (20 each) according to the test evaluated: stiffness and hydroxyproline tests.

Stiffness evaluation

A three-point flexure test was used to measure the initial stiffness of 20 completely demineralized dentin sticks. An aluminum testing jig with a 2.5 mm support span was used and specimens were centrally subjected to compression while immersed in distilled water, using a 1000 g load cell testing machine (Vitrodyne 1000, Burlington, VA, USA) at 1 mm/min crosshead speed. The specimens were deformed to a maximum strain of 15%. Load-displacement curves were produced and were then converted to stress-strain curves. Elastic modulus (E) was calculated at the steepest, most linear portion of stress–strain curves using the following formula:

where m = slope (N/mm); L = support span (mm); d = thickness of stick (mm); b = width of stick (mm).

Since specimen displacement was not measured with an extensometer or strain gauge but was an estimate from cross-head displacement, and the specimens’ thickness was not one-sixteenth of the length [40], the calculated elastic moduli were approximate. We were more interested in changes in modulus of elasticity rather than their absolute values.

After measuring the initial stiffness, the 20 demineralized sticks were immersed in distilled water (control) or Salvadora persica extract for 1 min (n = 10) and then immediately subjected again to 3-point flexure testing, under same parameters, to calculate the new stiffness values [13].

Hydroxyproline release test

Hydroxyproline is a major constituent of dentinal collagen. Degradation of collagen can be detected by the amount of hydroxyproline released [41]. When demineralized dentin is placed in a buffer solution at 37 °C, dentinal MMPs degrade the collagen network, which reduces its stiffness and results in loss of dry mass containing hydroxyproline [13, 42]. However, crosslinking the demineralized dentin resulted in a significant reduction in the lost dry mass, thus reducing the amount of hydroxyproline release [11, 42].

For this test, 20 dentin sticks (width = 3 mm, length 6 mm, thickness = 0.5 mm) were prepared and completely demineralized, as mentioned before. The demineralized sticks were placed in distilled water or Salvadora persica extract for 1 min (n = 10). Hydroxyproline release was measured using a colorimetric method [13, 43]. In brief, the treated sticks were immersed in a buffer solution (0.05 M HEPES solution, pH 7.4, Mallinckrodt Baker, Inc. Phillipsburg, NJ, USA) at 37 °C for 1 week. Hydrochloric acid (6 N HCL, Arcos Organics, Phillipsburg, NJ, USA) was used to hydrolyze the dissolved collagen into amino acids. Then, the HCL was allowed to evaporate, leaving dry remnants, which were then examined for the amount of hydroxyproline released. Figure 2 shows the experimental design for stiffness and hydroxyproline release tests.

Fig. 2

Experimental design for stiffness and hydroxyproline release tests, (n = 10)

Statistical analysis

Data were collected and entered using the statistical package for the Social Sciences (SPSS) version 28 (IBM Corp., Armonk, NY, USA). Normality was checked for by the Shapiro–Wilk test and proved to not deviate from normal distribution (p-value > 0.05). Data were summarized using means and standard deviations. For the microtensile bond strength test and nanoleakage assessment, two-way ANOVA was used to compare treatment and time interval groups, followed by the post hoc Bonferroni test. For stiffness and hydroxyproline release tests, a comparison between groups was performed using an unpaired t-test. p-values < 0.05 were considered as statistically significant.