Human and Bovine Dentin Composition and Its Hybridization Mechanism Assessed by FT-Raman Spectroscopy

FT-Raman spectroscopy was used to study the human and bovine dentin and their interactions with adhesive systems. Ten human (H) molars and ten bovine (B) teeth were prepared exposing the dentin and then each specimen was divided into two parts. e resulted forty dentin segments were treated either with the total-etch one bottle adhesive (Prime&Bond 2.1, PB ) or with the singlestep self-etching adhesive (Xeno III, X ) and divided into four groups: HPB (control), HX, BPB, and BX. Each group was analyzed by FT-Raman spectroscopy before and aer the adhesive treatment. Six regions of the Raman spectrum were analyzed and the integrated areas of organic and inorganic peaks were calculated. Bovine untreated specimens showed higher peak area of PO4 νν2 content than in human specimens. Human untreated specimens had higher peak areas of PO4 νν4 and CO3 νν1 contents than in bovine specimens.e peak areas of amide III, CH2, and amide I contents were higher in human than in bovine specimens (before treatments). Treated dentin showed no signi�cant statistical differences between the adhesives for both inorganic and organic contents considering the same substrate. However, the differences found between human and bovine specimens aer adhesives application show a reduced accuracy of these substrates as a substitute to the human specimens.


Introduction
e characteristics of dental hard tissues are the main factor to be observed when analyzing the possibilities of the replacement of human teeth by animal teeth for in vitro studies. As a candidate for human teeth replacement, bovine permanent incisors have been employed in previous adhesion studies [1][2][3][4].
It has been reported that the adhesion to the super�cial layer of dentin showed no signi�cant differences between human and bovine dentin, and the dentin bond strength decreased with the depth of dentin because of the lower density of dentinal tubules in the bovine dentin [5]. e morphology of coronal dentin and enamel is similar when comparing bovine and human teeth. Moreover, bovine teeth provide other advantages, such as similar age of the teeth and greater availability [6]. Histochemical and comparative anatomical studies have revealed that all mammalian teeth are essentially similar. Human and bovine dentin also presented similar radiodensity [7]. ere is still some concern whether the results of experiments with animal teeth can be extended to the human teeth and to the clinical situation; however, ethics committees around the world have stimulated the replacement of human teeth by animal ones [7].
Considering these factors, investigations on the chemical composition for hybridization of adhesive systems with bovine dentin are important to validate the studies to the human dentin. Contemporary adhesive systems interact with the dentin using two different approaches: either by completely removing the smear layer (etch-and-rinse technique) or by modifying it (self-etch technique) [8].
In selecting an adhesive system for the clinical use, it is very important to evaluate its bond strength and sealing ability [9]. Bond strengths are generally tested in tension or in shear [9,10]. Even though the results are very useful regarding the effectiveness of adhesive systems, there is a lack of information about the chemical interaction of the self-etch adhesive systems with the bovine dentin as a substitute to the human dentin.
Microscopy characterization can demonstrate the morphological relationship between the dentin and the adhesive layer, whereas Raman spectroscopy can elucidate the molecular interactions between the dentin and the adhesive monomers [11]. Raman spectroscopy permits the structural analysis of samples by identifying speci�c light-induced molecular vibrations [12]. In addition, the relative intensities of bands can lead to semiquantitative estimations of sample constituents [13].
One of the distinct advantages of the Raman technique from other analytical techniques is the ability to record the molecular information of both collagen and mineral component of teeth without damaging the sample [12]. Several studies using Raman spectroscopy to analyze the resin-dentin interface have been reported [2,4,11,[14][15][16].
Based on previous studies on the application of adhesives using human or bovine dentin as a substrate, a more detailed characterization of the adhesive interaction with either human or bovine dentin components needs to be performed. Such characterization is important since detailed information on dentin structure is essential to understand the data from the investigations on dentin-adhesive materials [17].
erefore, the purpose of this study was to investigate by FT-Raman spectroscopy whether there are differences in the inorganic and organic composition of human and bovine dentin before and aer total-etch-and self-adhesive systems application.

Specimen Preparation.
Ethical approval of the study was granted by the Ethics Committee of the University of Vale do Paraíba (L087/2005/CEP). Ten extracted erupted noncarious human third molars and ten bovine teeth were used in this study. e human teeth were obtained from patients whose extractions were part of the dental treatment, and the bovine incisors teeth were obtained from bovine jaws. All specimens were stored in saline solution (Aster Produtos Médicos LTDA, Sorocaba, SP, Brazil) at 9 ∘ C until use. Aer the extraction, the remaining so tissue was removed from the tooth surface with a dental scaler (7/8� Du�ex, Rio de Janeiro, RJ, Brazil). e teeth were polished with a paste of pumice (S.S. White, Rio de Janeiro, RJ, Brazil) and �ltered water using a Robinson brush (Viking-KG Sorensen, Barureri, SP, Brazil) in a low speed handpiece (KaVo do Brasil SA, Joinvile, SC, Brazil). Aer the cleaning procedure, the teeth were stored in 0.1% thymol aqueous solution at 9 ∘ C for one week long. To prepare the dentin specimens the teeth were washed for 24 h with �ltered water to eliminate thymol residues [18].
e occlusal one-third of the human teeth crowns were sectioned perpendicularly to the long axis using a watercooled low-speed diamond disc at 250 rpm with a 100 g load (Isomet 1000-BUEHLER, Lake Bluff, IL, USA). e dentin surface was grinded on wet 600-grit silicon carbide paper (Norton, São Paulo, SP, Brazil) at 150 rpm (Knuth Rotor-Struers, Brazil) for 1 min, under constant water cooling to produce a standardized smear layer [18]. Roots were removed with a water-cooled low-speed diamond disc producing a dentin slab for each tooth. Each specimen was then divided into two parts.
Similarly, ten bovine teeth were cleaned aer the extraction and stored in 0.1% thymol aqueous solution like the human teeth. e buccal enamel surface was removed using a water-cooled low-speed diamond disc at 250 rpm with a 100 g load and grinded on wet 600-grit silicon carbide paper at 150 rpm to expose the dentin layer. e dentin surface was polished for 1 min, under constant water cooling to produce a standard smear layer. e specimens were sectioned parallel to the long axis and the pulps were removed resulting into two parts. Ultrasonic cleaning (Maxiclean 1450, Merse, Campinas, SP, Brazil) with distillated water was performed for human and bovine teeth for 5 min in order to remove the excess of debris. e specimens were then stored in saline solution in a refrigerator at 9 ∘ C for one week. e chemical formulations of the two adhesives are listed in Table 1. All adhesives were applied onto dentin surfaces, in accordance with the manufacturers' instructions. Each adhesive was cured with a halogen light-curing unit (LCU) (Degulux so-start, Degussa AG, Hanau, Germany) with a power density of 745 mW/cm 2 . e light intensity of the LCU was measured by a power meter (Field Master GS, Coherent Inc., Auburn, CA, USA).

FT-Raman Spectroscopy Analysis.
FT-Raman spectroscopy analyzed the top surface of dentin slabs before and aer the treatments. One spectrum for each specimen was collected. e FT-Raman spectrometer (RFS 100/S-Bruker Inc., Karlsruhe, Germany) with a Ge diode detector cooled by liquid N 2 was used to collect the data. e samples were excited by an air-cooled Nd : YAG laser ( = 1064.1 nm). e power of the incident Nd : YAG laser on the sample was 100 mW. e spectral resolution was set to 4 cm −1 and for each measurement one spectrum was accumulated with 100 scans [18]. Aer the dentin treatment with each adhesive system, the same procedure was repeated and one spectrum for each specimen was accumulated on the top surface of dentin slabs, adding 80 spectra. e spectra of uncured adhesives were also collected using capillary tubes.
For the qualitative and semiquantitative spectral analysis, the spectra were baseline corrected and then normalized to the 960 cm −1 peak [19]. In the dentin Raman spectrum, six regions were evaluated: mineral component at 365-488 cm −1 (p1), 520-650 cm −1 (p2), and 995-1120 cm −1 . e statistical analysis of the Raman results was performed by the one-way ANOVA at a 95% level of con�dence. e Dunn's Multiple Comparisons post hoc test was performed using the Instat soware (GraphPad Soware, San Diego, CA, USA) to assess the signi�cance of the relative area evaluation of the treated dentin data. e Mann-Whitney test was also performed to study the signi�cance between the average of twenty untreated bovine and the average of 20 untreated human spectra integrated areas.

Untreated
Dentin. e typical Raman spectra for the untreated human and bovine dentin in the regions of 300-1200 cm −1 and 1200-1800 cm −1 are shown in Figures 1  and 2, respectively. e spectra have been vertically shied for clarity. Bovine and human spectra showed the same vibrational bands of mineral and organic components. In Figure 1, the intense peak at 960 cm −1 is associated with the phosphate (PO 4 3− 1 ) stretching vibration in the mineral apatite component of dentin, and the peak at 1071 cm −1 is attributed to carbonate (CO 3 2− 1 ) vibrations [13]. e peaks at 431 and 590 cm −1 are related to PO modes of phosphate, respectively [13]. In Figure 2, the bands at 1245, 1452, and 1667 cm −1 are attributed to the organic components of dentin, that is, type III collagen, CH 2 vibrations, and type I collagen (C=O), respectively [13]. Figure 3 shows the representative Raman spectra of the pure dental adhesives Prime & Bond 2.1 (Figure 3(a)) and Xeno III (Figure 3(b)), where the spectral features associated with both the aromatic and aliphatic components within both adhesive systems at 1608 and 1638 cm −1 , respectively [1], can be identi�ed. By a careful examination of this �gure, aer the adhesive treatment, the appearance of a shoulder at 1638 cm −1 in the dentin spectra (Figures 3(c) and 3(d)) for both substrates is noted.

Integrated Area Evaluation.
Based on the calculated integrated areas of the Raman peaks, the results of the Mann-Whitney test showed statistical signi�cant di�erence between the average of untreated human ( 2 ) and bovine ( 2 ) spectra related to the inorganic and organic dentin contents ( 1) ( Table 2). Bovine specimens showed higher peak area of PO 4 3− 2 contents than in human specimens. However, human specimens had higher peak areas of PO 4 3− 4 and CO 3 2− 1 contents than in bovine specimens. e peak areas of amide III, CH 2 , and amide I contents were higher in human than in bovine specimens.
Comparing the normal and treated spectra, it was observed that the adhesive treatment maintained or changed without statistical signi�cance ( ) the integrated area of the peaks related to inorganic (p1, p2, and p3) and organic (p4, p5, and p6) components of dentin for both adhesive systems ( Table 2). Statistical comparisons of the integrated areas of the Raman peaks between the adhesives and considering the same substrate (HPB versus HX; BPB versus BX) of treated dentin showed no signi�cant statistical differences between the adhesives for both inorganic and organic contents ( 0.05) (Tables 2 and 3).
Comparisons of treated dentin between the two substrates and considering the same adhesive (HPB versus BPB; HX versus BX) showed statistical signi�cant differences of the integrated areas of the Raman peaks, except for the peak at 590 cm −1 ( 0.05) (Tables 2 and 3). e integrated area of the p1 peak related to the PO 4 3− 2 content in BX specimens was higher than in HX ( 0.05). e integrated area of the p3 peak related to the vibrational mode of carbonate in HX and HPB specimens was higher than in BX and BPB ( 0.001). e integrated area of the p4 peak related to the type III collagen was higher in HPB than in BPB specimens ( 0.05). e integrated areas of the p5 and p6 peaks related to the CH 2 bonds and type I collagen, respectively, were higher in HPB and HX than in BPB and BX specimens ( 0.05) (Tables 2 and 3).

Discussion and Conclusion
Despite previous histochemical, anatomical, morphological, and mechanical comparative studies considering human and bovine teeth [1-3, 7, 9-11, 14], no studies have evaluated and characterized each vibrational mode of the dentin components of both substrates or established the relationship between dentin chemistry, specimens variability, and a possible in�uence on the adhesion. is relationship is important owing to the increase in the replacement of human teeth by animal teeth for in vitro studies. FT-Raman analysis revealed that untreated human and bovine dentin had signi�cant differences in the inorganic and organic contents in super�cial dentin (Table 2). ose differences found between substrates are probably due to the fact that the dentin in bovine incisors presents larger dentinal tubules and are more porous in the intertubular dentin than human molars [20]. Mature human dentin contains more carbonate than enamel (5-8 wt%) [13]. e mineral in human dentin is a carbonated apatite and is located either in the gaps between the collagen molecules (intra�brillar) or attached to the collagen �brils (extra�brillar) [21]. is fact is important because the collagen matrix is the key component of acidetched dentin.
When analyzing the results of the Mann-Whitney test, to verify chemical differences between untreated human and bovine dentin, our study showed higher peak areas related to the inorganic (p2, p3) and organic contents (p4-p6) in human specimens than in bovine specimens ( Table 2). Bovine specimens showed higher inorganic peak area related to the PO 4 3− 2 vibrational mode than in human specimens. ose differences in inorganic content related to phosphate and carbonate vibrational modes could be explained by the differences of arrangement, density, and diameter of dentin tubules between human and bovine dentin. Tanaka et al. [22] observed that the radiodensity of bovine coronal dentin was, on average, lower than that of the human coronal dentin. Fonseca et al. [7] also observed that bovine dentin is less dense than human dentin. ose mentioned differences could in�uence the etching process by phosphoric acid or by acidic primers where the amounts of minerals remained would be different between substrates. Pashley et al. [23] showed better in�ltration of resin in human dentin because of the higher percentage of dentin tubule surface in human than in bovine specimens. is higher percentage of dentin tubules could result in a better penetration of an acid into human dentin allowing a higher amount of dentin dissolution in human specimens at the same time as compared to bovine specimens [24].
e elevated organic content found in human untreated dentin could be another factor which in�uences the adhesion process since the exposed and stabilized collagen layer of dentin is a key factor to an adequate adhesion. Micromechanical retention is considered the most important mechanism for bonding resin to dentin. Such retention can occur when resin completely in�ltrates dentin surfaces and creates a resin-reinforced dentin layer [25]. A hybrid layer can be produced by etching the dentin prior to priming and bonding the dentin surface. During priming, hydrophilic monomers that diffuse across the demineralized dentin stabilize the hydrated collagen network and displace water with polymerizable monomers. Finally, the adhesive resins are applied to the primed dentin and polymerized [25]. e strength of the adhesive bonds between restorative materials and dentin is affected by the number of dentinal tubules per mm 2 and by their diameter, as well as the relative amount of intratubular and intertubular dentin [3]. According to Nakamichi et al. [5], a sufficient area of substrate could be obtained in different dentin depths of bovine incisors, but only the super�cial layer could be considered a substitute to human dentin. e dentin in bovine incisors presents larger dentinal tubules and more porous on intertubular dentin than human molars [3]. e statistical data presented in Table 3 showed that when the same substrate was considered, both adhesives interacted similarly with the dentin components despite their difference in the composition and mechanism of action. However, when the same adhesive was considered but with different substrates, signi�cant statistical differences were found among the components of treated dentin, except for the phosphate peak at 590 cm −1 . Previous studies comparing the total-etch one bottle adhesive with the self-etching adhesive reported higher bond strength values for the self-etching primer �lear�l; however this difference was not statistically signi�cant [9,10,26]. �ne explanation for these �ndings was the difference in the adhesive composition. e total-etch one bottle adhesives studied contained acetone and the selfetching adhesive contained water. Products that use acetone as the solvent may require a moist dentin substrate to produce adequate bonding. ese products may be extremely sensitive to the amount of water on the dentin surface, and even a small amount of drying may have a signi�cant role in reducing the bond strength [9,10,26]. e integrated area from the Raman peaks of human and bovine dentin aer adhesives application, shows statistical signi�cant differences between those two specimens (Tables  2 and 3), indicating a reduced accuracy of these substrates as a substitute to the human specimens. Since adhesive systems are developed according to the characteristics of human teeth, this fact should be considered when interpreting data from studies, in which bovine teeth were used as substitutes in the adhesion research. Regarding the analysis of adhesive interaction with dentin, some authors have found, by FT-Raman spectroscopy, piesces of evidence suggesting that ionic bonds and hydrogen bonding may form through the interaction of the ester function of adhesive ligand with the amide groups of collagen or through the formation of hydrogen bonds between the ligand and the collagen receptor [11]. In the present study a shoulder band at 1640 cm − was observed in the spectra, aer the adhesive treatment (Figure 3), and this band is related to the C=C bond of the adhesive as observed in other studies [1,11,14]. is aspect indicated the chemical interaction of the adhesive with dentin.
In summary, we presented direct information regarding the differences in the chemical composition of the human and bovine dentin interaction with one single-step selfetching adhesive and one total-etch one bottle adhesive by using FT-Raman spectroscopy. e major contribution of this study was the chemical characterization of the differences between the mineral and organic components of untreated and adhesive-treated human and bovine dentin as substrates. Untreated bovine dentin showed lower organic content than the human dentin and this is a possible limitation for in vitro studies of adhesion since adhesion mechanism in dentin is based on adhesive in�ltration in the exposed collagen layer aer demineralization by acid etching or acidic primers.
Further investigation is necessary to better understand how those differences in organic and inorganic content between these substrates in�uence the bond strength. In view of the results obtained in the present study, the authors believe that additional research must be conducted on the subject of the micromorphology and chemical composition of bovine dentin, using analytical tools as scanning electron microscopy, energy dispersive X-ray �uorescence, and micro-Raman among others, with the aim of acquiring more data about bovine teeth to the better use of bovine specimens in research conducted in this �eld.