A pilot study was done to calculate the sample size for each methodology. G*Power software was used. The effect size for each methodology was calculated by eta square (η2), which η2 = Sum of Squareseffect / Sum of Squarestotal. The η2 obtained were planktonic cells = 0.60; biofilm metabolic activity = 0.77; degree of conversio n = 0.89; microtensile bond strength = 0.4; demineralization at enamel-restoration margins = 0.62. For an F test and with η2 obtained the N of each methodology (and the n for each group) was calculated using the following parameters: A priori, α err prob = 0.05, number of groups = 4, power (β err prob) = 0.8.

An experimental adhesive system (Primer: Ethanol 35%, HEMA 40%, PMGDM 20%, Water 4%, Camphorquinone 0.5%, EDMAB 0.5%; Adhesive: BISGMA 70%, HEMA 29%, Camphorquinone 0.5%, EDMAB 0.5%), etch-and-rinse and 3-steps were manipulated. The adhesives compounds were mixed using a dual centrifuge (150.1 FVZ Speed Mixer DAC, Flack Tek; Herrliberg, Germany) for 2 min at 1300 rpm. Kaempferol powder (purity > 97%; Sigma-Aldrich, St. Louis, MO, USA) was added to the adhesive at concentrations of 1%, 2% and 4%, forming the following groups, respectively: K1%, K2% and K4%. Kaempferol was not added to the control group (K0). Resin composite Filtek Z250 A3 (3M ESPE Dental Products, St. Paul, MN, USA) was used as a restorative material. The light activator Bluephase (Ivoclar Vivadent, Schaan, Liechtenstein) was used for the whole study at an irradiance of 1200 mW/cm2 (± 10%).

Evaluation of the antibacterial effect of the adhesive system

The standard strain of Streptococcus mutans ATCC 25175 (American Type Culture Collection, Fiocruz, Rio de Janeiro, RJ, BRA) was used to evaluate the antibacterial activity of the adhesives without Kaempferol (K0) and with different concentrations of Kaempferol.

Evaluation of the biofilm's metabolic activity

Five specimens were made for each adhesive system. After any bubbles had been removed with a fine needle, a polyester strip and a glass coverslip were placed over the adhesive and photoactivation (Bluephase, Ivoclar Vivadent) was carried out. The experimental adhesive disks were transferred to sterile flat-bottomed 24-well plates (TPP, 24 Zellkultur Festplatte, SUI) that had been exposed to ultraviolet light in a laminar airflow chamber (aseptic environment) for 30 min for disinfection. After growth of S. mutans in BHI (Brain Heart Infusion, Difco, Sparks, USA) for 48 h at 37 °C under low oxygen tension, the bacterial suspension was adjusted for optical density of 0.5 (550 nm) according to the McFarland scale. Then, the suspension was diluted (1:100), 10.0 µl was diluted in 2 ml of BHI supplemented with 2% sucrose and added to each well along with the adhesive disk. The 24-well culture plates were incubated at 37 °C under low oxygen tension for 48 h. The metabolic activity of the S. mutans biofilms formed on the adhesive discs was analyzed using the blue demethylthiazole bromide tetrazolium bromide reduction assay (MTT, Sigma Aldrich, St. Louis, USA). The reaction using MTT is a colorimetric assay that measures the enzymatic reduction of MTT, a yellow tetrazole, to purple formazan. After gently washing the discs with sterile PBS (pH 7.2) to remove the cells not adhered to the biofilm, 1.0 ml of sterile MTT (1 mg/ml in PBS) was added to each well and incubated at 37 °C under low oxygen tension conditions for 1 h. Next, 1.0 ml of dimethyl sulfoxide (DMSO, Sigma Aldrich, St. Louis, USA) was added to each well and the plate was again incubated for 20 min at room temperature, while being protected from light and undergoing gentle shaking. Finally, the suspensions were subjected to a microplate reader with a wavelength of 540 nm. A higher absorbance indicates a higher concentration of formazan, which, in turn, indicates a higher metabolic activity of the biofilm.

Quantification of planktonic cells

S. mutans cell growth was also assessed by quantifying the planktonic cells in the culture medium where the bacterial biofilm was induced. After 48 h of biofilm induction on the adhesive disks, 100 μl of the cell suspension from each well was transferred to 96-well plates and these were subjected to a microplate reader at a wavelength of 660 nm to assess the turbidity of the medium and thus evaluate cell growth, since higher turbidity equals greater the number of cells [13].

Demineralization at enamel-restoration margins

Ten extracted, caries free, human third molars (Research Ethics Committee FM-UFF CAAE: 65046722.6.0000.5243) were disinfected in 0.5% chloramine-T aqueous solution for 7 days, then stored in distilled water. The roots were removed at the cementoenamel junction, using high-speed diamond burs. The crowns were sectioned in the buccolingual direction, forming 2 fragments for each tooth. The enamel surfaces were ground flat with sandpaper disks in a polishing machine (DPU 10, Struer, Midtjylland, Denmark). A preliminary Knoop hardness assessment (50 g, 10 s (s), HMV-2000, Shimadzu, Tokyo, Japan) was carried out on the specimens (2 lines, with 3 indentations each, separated 1500 µm) for distribution between the 4 groups (n = 5) according to the experimental adhesive systems tested: K0, K1%, K2%, and K4%. The variation in hardness within each group could not exceed 10%. Standardized cavities (2 mm × 2 mm × 1.5 mm deep) were prepared in the center of each fragment using #1092 diamond burs (KG Sorensen, Barueri, SP, Brazil) in a high-speed handpiece fixed in a standardization machine. In all groups, the enamel and dentin were etched with 37% phosphoric acid (Condac37, FGM; Joinville, SC, Brazil) for 15 s, rinsed with water for 30 s and blot dried with absorbent paper. The primer was actively applied for 10 s, followed by air drying for 5 s at a distance of 15 cm. The adhesive was applied, and light was activated for 20 s. The resin composite was inserted in 2 increments, each of which was light activated for 20 s. The restorations were polished with #2500 and #4000-grit SiC paper in a polishing machine.

Initial Knoop microhardness (SMH, 50 g, 10 s) of enamel surfaces was measured at 50 µm, 100 µm, 150 µm, 200 µm, and 250 µm from the restoration margin, with 5 indentations in each line (300 µm from each other).

The specimens were sterilized with ethylene oxide and fixed at the bottom of 24-well plates (24 Zellkultur Festplatte F, TPP; Trasadingen, Switzerland) with the surface of the restorations facing upwards. S. mutans isolate from the American Type Culture Collection (ATCC 25175, Fundação Oswaldo Cruz; Rio de Janeiro, RJ, Brazil) was cultured in brain heart infusion (BHI) broth (Difco; Sparks, NV, USA) supplemented with 2% sucrose at 37 °C under anaerobic conditions for 24 h. Afterwards, the bacterial suspension was adjusted to 0.5 in accordance with the McFarland scale at 550 nm using a UV/Vis spectrophotometer (Beckman Coulter DU 530, LifeScience; San Diego, CA, USA). Then, the suspension was diluted 1:100 and 10 μl of this suspension was added to each well, containing a specimen with 2 ml of BHI broth supplemented with 2% sucrose. The 24-well plates were kept for 48 h at 37 °C under microaerophilic conditions. During the experiment, the growth medium was changed every 24 h.

After the biofilm formation period, samples were removed from the plates and the nail varnish on the reference surfaces was carefully removed with acetone-soaked cotton wool. The final surface microhardness evaluation (SMH1) of enamel was performed 100 µm away from the initial indentations. The percentage of hardness change for enamel was calculated as follows: %hardness = 100 (SMH1 − SMH)/SMH [14, 15].

Dentin bond strength (μTBS) evaluation

Twenty-four extracted, caries free, human third molars (Research Ethics Committee FM-UFF CAAE: 65046722.6.0000.5243) were disinfected in 0.5% chloramine-T aqueous solution for 7 days, then stored in distilled water. The teeth were divided into 4 groups (n = 6) according to the adhesive systems evaluated (K0, K1%, K2% and K4%). The occlusal dentin of the teeth in each group was exposed with sandpaper and the peripheral enamel removed with a high-speed diamond bur. The dentin smear layer was standardized with #400 and #600-grit SiC papers (Arotec, Cotia, SP, Brazil) in a polishing machine (DPU 10, Struer, Midtjylland, Denmark) for 60 s. In all groups, the dentin was etched with 37% phosphoric acid for 15 s, rinsed with water for 30 s and dried with absorbent paper. The primer was actively applied for 10 s, gently air streamed for 5 s, and light activated for 20 s. Five 1 mm thick increments of microhybrid resin composite were horizontally added to the bonded surfaces and individually light activated for 40 s. After storage in distilled water at 37 ºC for 24 h, the teeth were longitudinally sectioned in both the mesiodistal and buccal-lingual directions, across the bonded interfaces producing beams with a cross-sectional area of approximately 1 mm2.

After storage for 24 h, the cross-sectional area of adhesive interfaces of beams was measured using a digital caliper, then the specimens were individually fixed to a microtensile device (ODMT03d, Odeme Biotechnology; Joacaba, SC, Brazil) using cyanoacrylate glue (Superbonder Gel 3M; Sao Paulo, SP, Brazil) and loaded under tension in a universal testing machine (EMIC DL 2000; Sao Jose dos Pinhais, SP, Brazil) at a crosshead speed of 1.0 mm/min until failure occurred. The µTBS (MPa) was obtained by dividing the load at failure (N) by the cross-sectional area of the beam (mm2). Each failed beam was evaluated using a stereomicroscope (SZ40; Olympus, Tokyo, Japan) at 40 × magnification, and the failure mode was classified as: adhesive (debonding at the adhesive interface), cohesive (debonding occurring mainly within dentin or resin composite), or mixed (mixture of adhesive and cohesive failure within the same fractured surface).

Degree of conversion (DC%)

The degree of monomeric conversion of adhesive systems (n = 5) was assessed using Fourier transform infrared spectroscopy (ALPHA-P FT-IR Spectrometer, Bruker Optics, Ettlingen, Germany), using the attenuated total reflectance-ATR-technique (Platinum Single Reflection Diamond Accessory (Bruker Optics, Ettlingen, Germany). Initially, the primer solvent was evaporated by leaving the bottle open, while being protected from light, on a digital scale. The solvent was considered evaporated when the weight of the primer solution stabilized. The primer and adhesive were mixed at a ratio of 1:1 and inserted into a Teflon mold (0.785 mm3) positioned on the ATR crystal of the spectrometer and the spectra between 1500 and 1800 cm−1 were recorded with the spectrometer operating with 40 scans and a resolution of 4 cm−1. Afterwards, the increments were light activated for 20 s, and the spectra were recorded exactly as was performed for the unpolymerized increments. The DC% was calculated as the ratio of the integrated area of absorption bands of the aliphatic C=C bonds (1638 cm−1) to that of aromatic C=C bond (1608 cm−1), used as an internal standard, obtained from the cured and uncured films, using the following equation:

$$}\% \, = 00 \, \times \, \left[ \, \left( }}} / \, R_}}} } \right)} \right],$$

where R = integrated area at 1638 cm−1 ̸integrated area at 1608 cm-1.

Statistical analysis

The data were analyzed using Statgraphics Centurion XVI software (STATPOINT Technologies; Warrenton, VA, USA). The normal distribution of errors and the homogeneity of variances were checked using the Shapiro–Wilk and Levene tests, respectively. Based on these preliminary analyses, biofilm metabolic activity, number of planktonic cells, DC%, and µTBS data were analyzed using ANOVA. Enamel demineralization data were evaluated using hardness loss results and analyzed using two-way ANOVA. Tukey’s HSD test was used for multiple comparisons when a statistical difference was detected. All analyses were performed at a level of significance of α = 0.05.