Determination of bisphenol A and related aromatic compounds released from bis-GMA-based composites and sealants by high performance liquid chromatography.

Most of the composites and sealants used in dentistry are based on bisphenol A diglycidylether methacrylate (Bis-GMA). Reports revealed that in situ polymerization is not complete and that free monomers can be detected by different analytic methods. Concerns about the estrogenicity of bisphenol A (BPA) and other aromatic components leached from commercial products have been expressed. We studied biphenolic components eluted from seven composites and one sealant before and after in vitro polymerization using HPLC and gas chromatography/mass spectrometry and we investigated how pH modifications affect the leaching of these components. We found BPA (maximal amount 1.8 microg/mg dental material), its dimethacrylate derivative (Bis-DMA, 1.15 microg/mg), bisphenol A diglycidylether (6. 1 microg/mg), Bis-GMA (2.0 microg/mg), and ethoxylate and propoxylate of bisphenol A in media in which samples of different commercial products were maintained under controlled pH and temperature conditions. Our results confirm the leaching of estrogenic monomers into the environment by Bis-GMA-based composites and sealants in concentrations at which biologic effects have been demonstrated in in vivo experimental models. The main issue with implications for patient care and dentist responsibility is to further determine the clinical relevance of this estrogenic exposure.

Polycarbonates, epoxy, and methacrylic resins are synthesized by reactions between individual chemicals, resulting in polymers with distinct physicochemical characteristics. As biomaterials they have multiple uses in human health applications; for instance, as manufactured products, for example, intraocular lenses (1), cements for trauma surgery (2,3), or bisphenol A (BPA)-based filling materials in preventive (4,5) and restorative dentistry (6). The latter requires extensive manipulation and in situ polymerization.
From a toxicologic standpoint, the migration of oligomers, monomers, and the precursors of synthetic polymers and other low-weight molecules from polymer networks must be carefully controlled because some of them can react with biologically important molecules. This is the case with BPA and bisphenol A diglycidylether (BADGE), which form adducts on DNA (7)(8)(9). BPA also binds to the estrogen receptor (10).
The estrogen receptor-mediated effects of BPA and related diphenyls make their detection an important issue. A study published in 1936 (11) demonstrated that diphenyl and diphenylmethanes had estrogenic activity in ovariectomized rats. Both compounds contained two hydroxyl groups in para positions. Later, Reid and Wilson (12) characterized further 4,4'-dihydroxy diphenylmethane-derived compounds with estrogenic activity, including molecular modifications of BPA with different hormonal activity. BPA mimics 1713-estradiol in vitro (10,(13)(14)(15)(16) as well as in estrogen target organs in animal models (17)(18)(19). BPA mimicked natural and synthetic estrogens such as diethylstilbestrol in experimental models other than breast and uterine tissue (20)(21)(22). Besides BPA, its dimethacrylate derivative Bis-DMA (16,23,24), BADGE (16), and related diphenylalkanes (25) are estrogenic in different bioassays and systems. Bis-GMA increased uterine wet weight and uterine collagen content in ovariectomized mice (26) and induced cell proliferation on MCF-7 cells (27), although previous works were unable to show any proliferative effect of Bis-GMA except under extreme pH conditions (16).
BPA was detected in a culture medium for yeasts as an estrogenic contaminant that leached from polycarbonate bottles during autoclaving (10). Also, BPA and other aromatic monomers are contaminants in wine and mineral water stored in plastic containers (28,29), in microwave susceptors (30), in food from cans coated with epoxy resin lacquers (15), and in canned oily foods (31). More recently, we found that BPA and Bis-DMA were responsible for the estrogenicity of some commercial composites and sealants used in dentistry; they leached into the saliva of treated patients (16). These results have been confirmed by Arenholt-Bindslev et al. (32) and Fung et al. (33), who found BPA in saliva samples collected immediately after the placement of Delton LC (batch no. 940218; Dentsply, York, PA). A yeast-based bioassay confirmed the estrogenic activity of these saliva samples (32). In addition to BPA (16,32,33), Bis-DMA and Bis-GMA also leached from polymerized composites and sealants, and their presence among eluted compounds was demonstrated after in vitro polymerization (34)(35)(36)(37)(38)(39). Leaching of components from composites and sealants may occur during the setting period of the resin and by the degradation of polymerized materials (40). Bis-DMA can also be a source of BPA because the enzymatic activity of saliva, esterases, extreme pH, and saliva storage can hydrolyze the dimethacrylate derivative (8,41). Figure 1 shows the chemical structure of BPA and related aromatic compounds.
The aim of the present study was to determine aromatic components eluted by in vitro polymerized Bis-GMA-based composites and sealants, to investigate how pH modifications affect the leaching of these components from the initial commercial product, and to assess their presence prior to polymerization. We found that BPA, Bis-DMA, BADGE, and Bis-GMA, among other aromatic components, leached from composites and sealants both before and after polymerization; their presence was always confirmed by gas chromatography/ mass spectrometry (GC/MS). Dental material. We used an Optilux VCL-300 polymerization lamp from Demetron Research Corporation (Danbury, CT). Its luminescence was tested and gave mean values of 435 mW/cm2. Polymerized composite samples were formed using 0.5cm diameter glass cylinders 1 cm high so that all the samples presented the same geometric shape.
Composites and sealants. Dental materials tested in this study were selected from among commercial products available in the Spanish market. They were selected with the exclusive condition that the producer indicated that they were Bis-GMA-based resins. The composites and sealant studied are as follows: * Charisma (batch 054; Heraeus Kulzer, Wehrheim, Germany) HPLC analysis. For HPLC analysis, we used a C-18 column (flow of 1 mL/min) and a UV detector at 280 nm. We used the following solvents: acetonitrile/water 1/1 (v/v) for phase A and acetonitrile for phase B. We used helium gas, a temperature of 25°C, and an injection volume of 20 pL. The gradient was 100% at 0 min (phase A), 100% at 15 min (phase B), and 100% at 17 min (phase A). The total duration of the chromatogram was 17 min. GCIMS analysis. For the GC/MS analysis, we used a 15-m methyl silica capillary column (OV-P); an injector temperature of 240°C; and an ion source temperature of 200°C. The temperature gradient was as follows: initial temperature of 80°C (2 min); final temperature of 320°C; and rate of temperature increase of 10°C/min. The carrier gas was helium, the flow was 1.2 mL/min, and the injection volume was 0.2 pL.
Qualitative analysis ofBPA and related aromatic compounds. We analyzed solutions of reference standard products in ethanol by HPLC at concentrations between 10-7 and 10-3 mol/L. Mean retention times of the products are listed in Table 1. In the GC/MS chromatography, mean retention times ± SD (minutes) for each of the reference standards were 21.5 ± 0.2 for BPA; 22.8 ± 0.2 for Bis-GMA; 23.9 ± 0.1 for EBPA; 26.8 ± 0.2 for Bis-DMA; 27.7 ± 0.2 for PBPA, and 27.9 ± 0.3 for BADGE. Dental methodology. Nonpolymerized samples were weighed (100 mg of each composite and 50 mg of the sealant), and samples were carefully decanted into the appropriate topaz glass vial to which 1 mL of distilled water was then added. A glass distillator and glass containers were used for water treatment and storage. After vigorous agitation, samples were left to settle for 24 hr.
Polymerized samples (100 mg) of each composite studied were compacted into a small glass cylinder to take on its geometric form. The cylinder was placed on a petri plate for compacting. The composite was polymerized to a 2-mm depth. Fifty milligrams of the sealant product was placed on a petri dish. Polymerization was immediately performed for 40 sec by situating the front of the lamp in close contact with the opposite side of the glass to that on which the cylinder was placed. Samples were extracted from the cylinder by applying pressure and were moved to a vial to continue the analysis. In all cases 1 mL distilled water was then added, and after vigorous agitation was left to settle for 24 hr.
The pH of the medium and the polymerized/nonpolymerized status were the paired variables. The pH values of 1, 7, 9, and 12 were regulated with 1 M HCI, distilled water, saturated solution of NaHCO3, and 1 M NaOH, respectively. The temperature used for each of these pH values was 37°C. Three samples of each commercial product were analyzed under these eight physicochemical conditions.

Results
Quantitative analysis ofBPA and related compounds. The detection and quantification limits were calculated in accordance with 10 concordant measurements of standard solutions for each of the products analyzed (BPA, EBPA, PBPA, BADGE, Bis-GMA, and Bis-DMA). Table 1 shows the detection limits  Articles * Bisphenol A in composites and sealants that corresponded to the value 30 (42) deduced from experiments performed to establish the linearity of the detector. Peak areas were determined by using the integrator HP-3394. Calculations were made according to International Union of Pure and Applied Chemistry recommendations. Detection limits ranged from 0.20 pg/mL for BPA to 1.70 pg/mL for PBPA.
Calibration curve parameters and linearity of the detector responses are shown in Table 1 together with the estimated quantification limits. Standard solutions of variable concentration between 10-7 and 10-3 mol/L were prepared and 20 pL of each was injected into the HPLC. Calibration curves were constructed from 10 concordant measurements and were used for the subsequent quantification in the analyses.
The accuracy of the chromatographic method was studied in the same conditions and was expressed as a percentage of recovery. The concentrations were calculated from the area of the peak and the mean response factor for each product ( Table 2). Recovery percentages were always above 95% and ranged from 95.7% for BPA to 100.2% for EBPA.
HPLC and GC/MS qualitative analysis. The compounds were first identified in HPLC from the retention times of each product. Figure 2 shows the chromatograms of polymerized Pekalux and Silux composites and the Delton sealant at pH 7. Peaks corresponding to the retention times of BPA, Bis-GMA, BADGE, PBPBA, EBPA, and Bis-DMA were observed. These findings were confirmed by GC/MS in all of the samples for which HPLC chromatograms showed the corresponding peaks. Figure 3 shows chromatograms (GC/MS) of compounds detected in the Delton sealant.
HPLC quantitative analysis. To determine how pH modifications affected the leaching of components both from the initial nonpolymerized commercial product and from the in vitro polymerized product, 20-,L aliquots of the 1-mL water suspension in the different physicochemical conditions defined were analyzed by HPLC. Tables 3-10 list the mean ± SD (in micrograms per milliliter) of the concentrations of the aromatic components (BPA, EBPA, PBPA, BADGE, Bis-GMA, and Bis-DMA) quantified in each commercial sample before (nonpolymerized) and after polymerization (polymerized) in four pH conditions (acid, pH 1; neutral, pH 7; and alkaline, pH 9 and pH 12). Data indicate the amount of compounds leached for a 24-hr period, expressed in micrograms per milliliter medium; because the volume of the sample was 1 mL, the data shown represent the total amount of leached components for 100 mg composite and 50 mg sealant. BPA was detected in all of the commercial samples studied both before and after polymerization. The maximal amount of BPA (1.8 pg/mg) was found in nonpolymerized Charisma independent of the pH. The lowest levels of BPA were found in nonpolymerized Polofil. In addition, polymerized Tetric and nonpolymerized Brillant were negative for this monomer at neutral pH. BADGE was found in all the samples (maximal 6.1 pg/mg in polymerized Delton) except nonpolymerized Charisma. The oligomer Bis-GMA was found in all polymerized and nonpolymerized samples except Silux samples, which were also negative for Bis-DMA and PBPA. However, Silux samples showed the presence of EBPA both before and after polymerization.
With respect to the presence of Bis-DMA, this monomer was found in all of the samples studied except the Brillant and Silux products. All others showed variable amounts of Bis-DMA (maximal 1.15 pg/mg for polymerized Pekalux at the highest alkaline pH) after polymerization (Charisma, Pekalux, Polofil, Tetric, Z-100, and Delton), and, in Polofil and Delton, also before polymerization. Finally, the presence of PBPA and EBPA was less frequent. PBPA was exclusively detected in nonpolymerized samples of Pekalux and in polymerized Delton. EBPA was detected in nonpolymerized Pekalux and Silux and in polymerized Delton, Silux, and Tetric.

Discussion
The leaching of aromatic components from composites and sealants polymerized in vitro poses a clinical problem because of the demonstrated estrogenicity of some biphenolic compounds. We studied seven composites and one sealant and detected BPA in all of the commercial products analyzed, although some acid and neutral conditions were less conducive to its release. Regarding the release of other biphenolic compounds (25), all samples that showed detectable levels of Bis-DMA were also positive for the presence of Bis-GMA, BADGE, and BPA. The elution of leachable compounds from samples greatly depends on the polymerization conditions (43), and standardization of these conditions was thus an essential first step in our study design. The polymerized and nonpolymerized resins were left undisturbed at rest and were submitted only to variable pH at a controlled temperature, so that the components migrated to the liquid phase without the use of organic solvents. Methanol (35,44,45) and water/ ethanol combinations (46,47), a tonic drink (48), and food-simulating chemicals (47) have been reported in the literature as appropriate media for the extraction of polymerized composites. We used water in this study in an attempt to reproduce in vitro the conditions to which Bis-GMA-based resins are submitted when utilized in dentistry, as in our earlier work (16). Nevertheless, the use of water may result in lower levels of monomer elution than those found when organic solvents are applied, and this would account for quantitative differences between our findings and those of other authors.
The effect of pH variations on the leaching of monomers in dental microenvironments was reported previously (49). Extreme pH has some influence on the extraction of components from polymerized and nonpolymerized samples. In addition, leached components may be broken down in the stomach. The role of pH in the elution of compounds is evident in the results of the present study, especially for certain monomers in polymerized samples. For example, elution of BPA increased as the pH become more alkaline and was greater at pH 12 than at any other pH value. BPA values ranged between 0.014 (1.4 pg/mL) and 0.195 pg/mg (19.5 pg/mL) for the Charisma composite, between undetectable and 0.113 pg/mg (11.3 pg/mL) for Brillant, and between 0.129 (12.9 pg/mL) and 1.161 pg/mg (116.1 pg/mL) for the Tetric product, at pH 7 and 12, respectively. Although in the present study we highlight the data for BPA, these enhanced elution values at extreme pH were similar to those of other components found in most of the composites studied. Interestingly, when Bis-DMA was subjected to pH 11 for 30 min at 50°C, an almost 100% conversion of Bis-DMA to BPA was reported by Schmalz (41), who also showed the effect of saliva and esterase treatment on the hydrolysis of the Bis-DMA to BPA. Another recent study (50) confirmed that Bis-DMA was readily hydrolyzed to BPA when stored for 4 months in saliva at -20°C, whereas BPA was stable under the same storage conditions. We stress that the composites and sealants are unstable and that to a greater or lesser degree, depending on the aggressiveness of the medium, it is always possible to detect the elution of monomers, olygomers, and precursors. These findings confirm reports that in vitro polymerization is not complete and that free monomers can be detected by different analytic methods (38,39,51,52). Several publications have appeared since our first report on BPA and Bis-DMA leaching from the Delton sealant (16) that may help to define a position regarding monomer leachability and their hormonal activity in vivo. For instance, Hamid and Hume (38) and Nathanson et al. (39)   Articles * Bisphenol A in composites and sealants leaching from unpolymerized dental materials and Arenholt-Bindslev et al. (32) and Fung et al. (33) found BPA in the saliva of patients treated with Delton. Hamid and Hume (38) concluded that earlier concern about the possible adverse effects of BPA was questionable, whereas Nathanson et al. (39) claimed that there was no current need to use different sealants or restrict their use. Recently, we refuted point by point (54) the observations of Nathanson (39) and called for further research to address this issue. In the current study we used a reverse-phase HPLC method that substantially differs from the methods used by other authors. We seti the detector at 280 nm, molecules absorb and ir other nonaromatic compc Wavelengths of 210-227 n previously for the determir ene glycol dimethacrylat (34,38,39,55). BPA and pounds of interest to us (i DMA, and Bis-GMA) sho2 27 nm but also absorbed out interference from line dimethacrylate derivatives, trum with a single peak Finally, chromatographi pointed out that derivation with tri-Dunds is avoided.
methylsilyl was necessary to detect BPA in m have been used dental material (53). nation of triethyl-In fact, Bis-DMA was found by te and Bis-GMA Nathanson et al. (39) in appreciable amounts the other com-(1.23 pg/mg in Delton and 0.39 pg/mg in .e., BADGE, Bis-Defender sealants, respectively), confirming wed absorbance at previous reports of the presence of this comat 280 nm withpound leaching from composites and al ethylene glycol sealants (16,35,56). Bis-GMA was also which had a spec-found in the eluates of polymerized sealant at 218 nm (48).
With respect to the implications for patient care and dentist responsibility, more data must be gathered before a complacent attitude toward this hazard is adopted. As Soderholm and Mariotti (58) proposed, the main issue is to determine whether the estrogenic effects of dental sealant and composite monomers have any real clinical consequences. This concern should be addressed by new studies that focus on the toxicityestrogenicity-of leached monomers, 352 oligomers, and precursors, similar to the , ,| -, , sz paper recently published by our group (25).
It will be difficult to give an a priori prediction of the risk to human health posed by Scan El+ 349 4.196 such exposures, but new data may help assess this exposure within the general risks attributable to xenoestrogens and especially BPA (59). Four recent observations have raised concern about the estrogenicity of bisphenols. First, a more potent in vivo BPA effect has been demonstrated as compared to previ-364 ous in vitro assays (19,21). For instance, 3 days of exposure to microgram levels of BPA (60-100 pg/rat/day) released from capsules 350 promoted cellular proliferation in rat uterus and vagina, which showed molecular and morphologic alterations nearly identical to those induced by estradiol (60). The biologic significance of the 179 pg BPA eluted from a 100-mg polymerized sample of commercial l365 composite in our study should be considered from this standpoint. Second, BPA seems to 351 act on other target organs as well as the obvi-_ i3,6 ous organs [breast and uterus (21,22,49,61)]. ko46a WO m80 j400 Third, genetic differences in susceptibility to the estrogenic effect of BPA have raised concerns about subpopulations with a higher sensitivity to this estrogen (18,21 (63).
We confirm the leaching of BPA and other aromatic compounds from one sealant and we present new data on biphenolic monomers leaching from seven other