Estrogenicity of resin-based composites and sealants used in dentistry.

We tested some resin-based composites used in dentistry for their estrogenic activity. A sealant based on bisphenol-A diglycidylether methacrylate (bis-GMA) increased cell yields, progesterone receptor expression, and pS2 secretion in human estrogen-target, serum-sensitive MCF7 breast cancer cells. Estrogenicity was due to bisphenol-A and bisphenol-A dimethacrylate, monomers found in the base paste of the dental sealant and identified by mass spectrometry. Samples of saliva from 18 subjects treated with 50 mg of a bis-GMA-based sealant applied on their molars were collected 1 hr before and after treatment. Bisphenol-A (range 90-931 micrograms) was identified only in saliva collected during a 1-hr period after treatment. The use of bis-GMA-based resins in dentistry, and particularly the use of sealants in children, appears to contribute to human exposure to xenoestrogens.

The impact of certain estrogenic xenobiotics on the development, health, and reproductive systems of wildlife has been clearly documented (1). As data accumulate, environmental xenobiotics are also being implicated in human infertility, genital tract malformations, and increased cancer rates in estrogen target tissues (2,3). In 1936, Dodds and Lawson reported the estrogenicity of some diphenyl compounds containing two hydroxyl groups in para positions (4). Reid and Wilson (5) subsequently confirmed the estrogenicity of 4,4'dihyroxydiphenylmethane derivatives. One such derivative, bearing two methyl groups and known as bisphenol-A, is a major component of epoxy resins. Bisphenol-A was found to leach from autoclavable polycarbonate laboratory flasks (6). Recently, we demonstrated that food packed in lacquercoated cans was active in a bioassay for estrogenicity; bisphenol-A released from the epoxy resin lining of the cans was identified as the estrogenic contaminant (7).
Bisphenol-A is a common ingredient in restorative materials used in dentistry. Since the 1960s, when bisphenol-A diglycidyl methacrylate (bis-GMA)-based restorative materials were first used in odontology, many studies have assessed the effects of resins on pulpal injury (8) and their cytotoxic properties (9)(10)(11). Little attention was paid, however, to the systemic health effects of these chemicals or their monomers (12,13). Some studies focused on the alkylating properties of the glycidaldehyde portion of bisphenol-A diglycidylether (BADGE) (14). For example, the formation of glycidaldehyde adducts in adenine residues was demonstrated in mice after cutaneous treatment with BADGE (15). Other studies examined the mutagenic an carcinogenic properties of epoxy resi monomers, with contradictory results (15, Resin-based composite restoratih materials used in dentistry ("composites' consist of two major components: a organic resin matrix and an inorganic fille Composites without inorganic fillers a. known as "sealants." Tooth-colore restorative materials are increasingly use for preventive purposes, to replace missir tooth structures and to modify tooth cob and contour. The resin matrix is initial present as a fluid monomer that is conver ed into a rigid polymer by a free radica initiated reaction of addition. The pot merization reaction (curing) can be chenm cally initiated (autocuring) or photoinitia ed using ultraviolet or visible light in tI presence of a photoinitiator.
Because of the low degree of polyme. ization required by the monomers in con posites, concerns have been expresse about the leaching of chemicals froi unpolymerized material, which is rapid' released after curing (16). In vitro studii have shown that most of the unpolyme ized monomers of bis-GMA had leache 24 hr after serting. In addition, cured con posites placed in the oral cavity ai attacked mechanically and chemicall, Enzymatic hydrolysis of methacrylate together with mechanical forces, coi tribute to the breakdown of composii resins (17), which are slowly and persi tently degraded (11]). A significant portion of the uncured decomposed material that is swallowed co be absorbed by the intestine. Climie et (19,20) studied the metabolic degradatic of 14C-BADGE in mice after the or administration of a single dose and found that 90% of the radioactivity was eliminated in feces and urine during the first 3 days of the experiment. Interestingly, a small k amount of BADGE (-5%) underwent F oxidative dealkylation to yield glycidaldehyde (which has alkylating properties) and V bisphenol-A, among other by-products (20). The systemic behavior of the diphenolic derivative is poorly known and needs further investigation. The purpose of this research was to determine whether bis-GMA-based restorative resins and their components have estro-Ld genic activity in an in culture assay. We in demonstrate here that the sealant and resin  A, Sigma Chemical Co., St. Louis, Missouri) with 0.5% dextran T-70 (Pharmacia-LKB, Uppsala, Sweden) was prepared. Aliquots of the charcoal-dextran suspension of a volume similar to the serum aliquot to be processed were centrifuged at 1iOOg for 10 min. Supernatants were aspirated, and serum aliquots were mixed with the charcoal pellets. This charcoal-serum mixture was maintained in suspension by rolling at 6 cycles/min at 37°C for 1 hr. The suspension was centrifuged at 10OOg for 20 min, and the supernatant was then filtered through a 0.20-pm filter (Gelman Sciences, Ann Arbor, Michigan). Charcoal dextran-treated human serum (CDHuS) was stored at -20°C until needed. Cellproliferation experiments. We used MCF7 cells in the E-screen test of estrogenicity according to a technique slightly modified (23) from that originally described by Soto et al. (24). Briefly, cells were trypsinized and plated in 24-well plates (Limbro, McLean, Virginia) at initial concentrations of 10,000 cells per well in 5% FBS in DME. Cells were allowed to attach for 24 hr, then the seeding medium was replaced with 10% CDHuS-supplemented phenol red-free DME. Different concentrations of the test compound were added, and the assay was stopped after 144 hr by removing medium from wells, fixing the cells, and staining them with sulforhodamine-B (SRB). The staining technique was modified from that described by Skehan et al. (25). Briefly, cells were treated with cold 10% trichloracetic acid and incubated at 4°C for 30 min, washed five times with tap water, and left to dry. Trichloroacetic-fixed cells were stained for 10 min with 0.4% (wt:vol) SRB dissolved in 1% acetic acid. Wells were rinsed with 1% acetic acid and air dried. Bound dye was solubilized with 10 mM Tris base (pH 10.5) in a shaker for 20 min. Finally, aliquots were transferred to a 96-well plate and read in a Titertek Multiscan apparatus (Flow, Irvine, California) at 492 nm. We verified the linearity of the SRB assay by cell number before cell growth experiments.
Results are expressed as means ± SDs. We normalized mean cell numbers from each experiment to the steroid-free control cultures to correct for differences in the initial seeding density. Differences between the xenoestrogen and estradiol-17L; groups were assessed by analysis of variance and the a posteriori Scheffe's test. A p -value < 0.05 was regarded as significant.
Estrogen andprogesterone receptor measurements. MCF7 cells were seeded in T-25 flasks in 5% FBS-supplemented DME. On the following day, the medium was changed to 10% CDHuS-supplemented phenol red-free DME medium, and estradiol-17l3 or the chemicals to be tested were added. One group of cells received vehicle alone. After 72 hr, the culture medium was discarded and cells were frozen in liquid N2. To extract receptor molecules, we incubated cells at 40C for 30 min with 1 ml extraction buffer (0.5 M KCl, 10 mM potassium phosphate, 1.5 mM EDTA, and 1 mM monothioglycerol, pH 7.4), according to a technique previously described in detail (26). After centrifugation to pellet the cell debris, estrogen and progesterone receptors were measured in a 100-jil aliquot of the supernatant by enzyme immunoassay using the Abbott ER and PgR enzyme immunoassay monoclonal kits (Abbott Diagnostic, Wiesbaden, Germany) according to the manufacturer's instructions.
Estrogen-induced cell type-specific proteins. We measured cathepsin-D and pS2 in the culture medium of MCF7 cells with the ELSA-CATH-D and ELSA-PS2 immunoradiometric assays (CIS Bio International, Gif-sur-Yvette, France). The culture medium was centrifuged at 1,200g for 10 min to eliminate floating and detached cells. Samples were kept frozen at -800C until the assays were conducted.
Competitive binding assays. Cytosol from immature, female rat uteri was prepared at a protein concentration of approximately 2 mg/ml in phosphate buffer. Aliquots of this 105,000g supernatant were then incubated with various concentrations of bisphenol-A, bisphenol-A dimethacrylate, BADGE, bis-GMA, and 3 nM [3H]estradiol for 16 hr at 0-40C. The free and bound fractions were separated with the charcoal-dextran technique. We calculated the relative binding ability (RBA) of each competitor as the ratio of the concentration of radioinert estradiol/competitor required to inhibit 50% of the specific [3H]estradiol binding, with the affinity of estradiol set at 100%.
Determination ofcomponents in resinbased restorative material We analyzed four commercial bis-GMA-based composite resins using an HPLC technique. Aliquots of 50-100 mg of each sample were suspended in 1 ml ethanol and chromatographed using a Lichrocart Merck S5 ODS column (20 x 0.4 cm; Merck, Darmstadt, Germany) at a flow rate of 1 ml/min in a Perkin-Elmer 250 binary LC chromatograph with a Perkin-Elmer 235 diode array detector 235 equiped with a 20-pl loop injector (Waters 717 Plus Autosampler). The gradient was: at time = 0 min, 100% phase A (acetonitrile-water, 1:1) and 0% phase B (acetonitrile); at time = 15 min, 100% phase B and 0% phase A; and at time = 17 min, 100% phase A and 0% phase B. The elution profile was monitored at 280 and 254 nm.
Determination of bisphenol-A and related compounds in saliva. For the dinical study, 18 healthy men and women (age range 18-25 years, mean age 20 years) were treated. The experiment was conducted in accordance with the Helsinki Declaration. A sealant was applied to 12 molars according to a standard technique that included visible light as the activator of the curing agents. Samples ofwhole saliva were collected by having subjects spit for 1 hr before and 1 hr after treatment into preweighed glass flasks containing 10 ml ethanol. The amount of commercial bis-GMA-based sealant used was approximately 50 mg per subject. After collection, saliva samples were centrifuged (1OOOg, 10 min) and passed through a 0.45-pm pore filter. Aliquots of 20 pl were chromatographed by HPLC as described above and by GC/MS. Saliva samples were also tested for estrogenicity in the MCF7 breast cancer assay.
Spectroscopy. Mass spectra of extracted composite resins, filtered saliva samples, and technical grade bisphenol-A, dimethacrylate of bisphenol-A, and BADGE were obtained in a mass spectrometry system (Fisson VG Platform-II, Manchester, UK) operating at an ion source temperature of 200°C in a Fisson GC 8000 SM Autospec chromatograph (Milan, Italy). A 15-m methyl silicon column (OV-P) was used at a flow rate of 1.2 ml/min, with helium as the carrier gas. The oven temperature was 80-320°C, with a graded increase of 10°C/min. Relative retention times of all the components analyzed were 16.2 min for bisphenol-A; 20.5 min for BADGE, and 21.2 min for bisphenol-A dimethacrylate.
Steroids and chemical compounds tested Estradiol-17g was obtained from Sigma. Bisphenol-A, bisphenol-A dimeth-acrylate, BADGE, and bis-GMA were obtained from Aldrich (Aldrich-Chemie, Albuch, Germany). Chemicals were dissolved in ethanol to a final concentration of 1 mM and stored at -20°C.

Results
Estrogenic Activity of Resin-based Composites The addition of estradiol-17g to CDHuSsupplemented medium increased the numbers of MCF7 cells in culture. Maximum proliferative effect was obtained at estradiol-17f concentrations of 10 pM and higher (Fig. IA). Cell yields were sixfold greater than in control cultures after 6 days (mean ± SD, 6.67 ± 1.21; n = 15 experiments). In the absence of estradiol-17l3, cells proliferated minimally.
Four bis-GMA-based composite resins were assayed before polymerization using the MCF7 proliferation test. Composites and sealants were prepared in ethanol at a concentration of 100 and 50 mg/ml, respectively, and assayed at a range of dilutions (1/100 to 1/106). The cell yield obtained with 5 pg/ml sealant sample was sixfold greater than in control cultures. The magnitude of this proliferative effect was similar to that obtained with estradiol-1713. Cell toxicity was observed at concentrations of 50 pg/ml and higher (Fig. 1B).
In contrast to the sealant sample, the three resin-based composites assayed did not induce MCF7 cell proliferation at a maximal concentration of 1 mg/ml. These commercial formulations contain a large proportion of inorganic filler particles (50-85% by weight of the composite). Stock suspensions were prepared without discarding the inorganic portion. showed the corresponding peaks (Fig. 3).
To determine whether the monomer bisphenol-A can be cleaved from composites and from their main oligomers (dimethacrylate of bisphenol-A, bis-GMA, and BADGE), these were hydrolyzed in alkaline (pH = 13) and acidic media (pH = 1) after heating (100'C for 30 min), and then chromatographed as indicated. Table  1 summarizes the components identified in the chromatograms. Extensive breakdown of the oligomers into smaller compounds such as bisphenol-A was observed. Bisphenol-A and its dimethacrylate were estrogenic when assayed in the breast cancer cell proliferation assay. The concentrations required to produce maximum proliferation of MCF7 cells were 10,000-fold higher than those of estradiol-17f; (Fig. 4). MCF7 cells were grown in the presence of bisphenol-A or bisphenol-A dimethacrylate and the estrogen antagonist hydroxytamoxifen; the proliferative effect of 0.1 pM bisphenol-A and bisphenol-A dimethacrylate disappeared when cells were grown in the presence of a 10-fold higher concentration of the antagonist (data not shown).
Bis-GMA was negative in the estrogenicity test. A proliferative effect of BADGE was observed at a concentration of 10 pM (Fig. 5). Oligomers from both compounds became active in the proliferative test after hydrolysis (100°C for 30 min, in alkaline or acidic medium). Bisphenol-A and bisphenol-A dimethacrylate were identified in chromatograms when oligomers were found to be estrogenic after hydrolysis.  ER in these cells was 183 ± 29 fmol/mg of extracted protein (Fig. 6A). Estradiol-171 decreased the concentration of ER by 50% and increased that of PgR The basal PgR value was 15.5 ± 4.3 fmol/mg protein, close to the lower limit of detection of the monoclonal antibody assay (Fig. 6B). Estradiol-17l (1 nM) significantly increased PgR nearly 15-fold over the control value. Treatment of MCF7 cells with concentrations of 0.1 pM and higher of bisphenol-A dimethacrylate resulted in a significant increase in PgR levels. Bisphenol-A treatment also increased PgR levels (Fig.  6B). However, no reduction of ER levels was observed when bisphenol-A or bisphenol-A dimethacrylate induced PgR Secretion of pS2 and Cathepsin-D Cathepsin-D and pS2 accumulation in the culture medium reflected increases in cell number during the experiments. The secretion of pS2 by MCF7 cells was significantly increased by >0.1 nM estradiol-17g (-3.5-fold increase over controls). The basal concentration of pS2 (55.4 ± 15.2 ngIl6 cells) increased to 202.2 ± 19.1 ng/10 cells after treatment with 10 pM bisphenol-A dimethacrylate; this was similar to the effect of estradiol-1 7L on pS2 secretion (Fig. 7A).
The basal concentration of cathepsin-D secreted by MCF7 cells was 8.9 ± 2.9 pmol/106 cells. Estradiol-1781 treatment (1 and 10 nM) resulted in a modest 1.7-fold increase in the accumulation of cathepsin-D in the culture medium. Bisphenol-A dimethacrylate (1 and 10 pM) increased cathepsin-D secretion only 1.6-fold (Fig.  7B). These changes were not significant.

Binding Affinity of Composite Resin Compounds for the Estrogen Receptor
The binding affinity of bisphenol-A, bisphenol-A dimethacrylate, BADGE, and bis-GMA for the cytosol ER was determined by competitive binding analysis. The receptor relative binding affinities (RBA) were 0.012 and 0.0033 for bisphenol-A and bisphenol-A dimethacrylate, respectively, versus 100 for estradiol-1711.
Bis-GMA and BADGE did not compete with estradiol for binding to the ER, even at concentrations a million-fold higher than estradiol-17f (Fig. 8).

Analysis and Quantification of Estrogenic Substances in Saliva
Eighteen saliva samples taken before and after sealant treatment were chromatographed using the technique described in Methods. Figure 9 shows the chromato-   Figure 5. Cell proliferation in MCF7 cells. Cells growing in 10% charcoal dextran-treated human serumsupplemented medium were exposed for 144 hr to (A) bis-GMA and (B) BADGE and to products obtained by hydrolytic treatments of these compounds in alkaline (pH = 13) and acidic medium (pH = 1  (Table 2). After treatment, all saliva samples contained variable amounts of bisphenol-A ranging from 90 to 931 pg; this was confirmed by mass spectrometry. Subjects were treated with a fixed amount of approximately 50 mg of sealant, and the proportion of free, unpolymerized material collected during 1 hr after treatment never exceeded 2% of the total  Figure 7. Accumulation of (A) pS2 and (B) cathepsin-D in culture medium. MCF7 cells were grown in 10% charcoal dextran-treated human serum-supplemented medium and exposed for 144 hr to 1 and 10 nM estradiol-17B. Parallel cultures were exposed to 1 and 10 pM bisphenol-A or bisphenol-A dimethacrylate. Controls received the vehicle alone. *Values significantly different from control, p < 0.001.
her saliva before the second treatment. The results from this subject were excluded from analysis. Samples containing the highest amounts of bisphenol-A and bisphenol-A dimethacrylate were estrogenic in the proliferation test. Figure 10 shows the proliferative pattern of a sample containing 231 jig bisphenol-A in 27 ml of collected saliva plus 10 ml of ethanol. An aliquot of 100 pi of this sample diluted in 1 ml of the culture medium was positive in the estrogenicity assay. The estimated amount of bisphenol-A assayed was 0.62 pg/ml. Because the presence of endogenous estrogens could not be ruled out, saliva samples taken before the sealant was applied were also assayed; none showed estrogenicity.

Discussion
Epoxy resins of the bisphenol-A type have many applications such as lacquer coating in food cans, in dental, surgical, and prosthetic devices, and as additives in a variety of other plastic materials (27). Here we demonstrate that bisphenol-A and bisphenol-A dimethacrylate, components of commercial resin-based composites and sealants used in dentistry, are estrogenic and may represent additional sources of human exposure to xenoestrogens.
Commercial smooth-surface composites vary in monomer composition and thus have different physical properties and biological behaviors. We found that the amount of bisphenol-A and bisphenol-A dimethacrylate varied between commercial   GMA-based resins and found bis-GMA in all samples at concentrations ranging from 19 to 51% of the total composite weight. Iso-bis-GMA was also present and represented 40% of the total percentage of bis-GMA. Bisphenol-A dimethacrylate was found in four commercial mixtures in amounts that ranged from 6 to 10% of the total composite weight. In addition to these monomers, other compounds such as light activators, phthalates, and benzoates, as well as a large amount of inorganic fillers, are commom ingredients of composites (28). Once applied to tooth cavities, composites and sealants are polymerized in situ. The degree of conversion of oligomers into polymers varies depending on the composition of the resin and its distance from the tooth surface. Conversion of 60-75% is expected with most common composites (16), although levels of curing as low as 30% are found in the bottom of fillings. Lower levels of conversion are thought to be associated with greater elution of free components from the composites (16). A strong inverse relationship has been found between the leaching of resin from bis-GMA-based composites and monomer conversion. Monomer leaching reached a plateau at conversion levels of 60% (29). In nine commercial resins, the amount of residual bis-GMA measured at the end of the setting time was in the range of 0.4-1.2% of the original weight after curing. Amounts of the residual monomers of up to one-tenth the initial release were eluted into water over a period of 14 days (29). Higher values were described by Ferracane and Condon (16), who found differences in percentage elution in relation to curing and double illumination. Our data confirm these figures: residual bisphenol-A in saliva after curing ranged from 0.1 to 2% of the 50 mg initially applied to the tooth surface in samples taken during 1 hr after treatment.
In addition to bisphenol-A, present in 100% of the saliva samples, its methacrylate derivative was present in 3 out of 18 samples. Both compounds were identified by chromatography and estrogenic activity as measured in the MCF7 breast cancer cell proliferation test, the E-screen assay.
The oligomer bis-GMA was not estrogenic in the E-screen bioassay. Hydrolyzed BADGE and bis-GMA in alkaline or acidic media were estrogenic. Chromatographic analyses of hydrolyzed samples detected the presence of bisphenol-A and bisphenol-A dimethacrylate after treatment. The metabolic routes of systemic degradation of some composite resins have been studied in mice (15,19,20). For example, BADGE is metabolized by oxidative dealkylation to give a phenol diol derivative. Much attention has been given to the smaller fragment of the cleavage process because of the mutagenic properties of the glycidylaldehyde portion. However, the estrogenic activity of these derivatives has not been evaluated; their chemical similarity to some bisphenol-A derivatives with this property suggests the need for closer examination. Moreover, little information is available about the potential estrogenicity of byproducts of the environmental breakdown of biphenolic compounds. Lobos et al. (30) isolated a novel bacterium from the sludge of a wastewater treatment plant of a plastics manufacturing facility; the organism was able to degrade bisphenol-A to 4-hydroxybenzoic acid and 4-hydroxyacetophenone (both monophenolic by-products) and to 2,2-bis(4-hydroxyphenyl)-1-propanol, and finally to 2,3-bis(4-hydroxyphenyl)-1,2propanediol. So far, the estrogenicity of these bisphenol-A derivatives has not been explored, although the similarity of the latter compound to some stilbenes suggests that it may display considerable activity.
Bisphenol-A is toxic to fish and invertebrates at concentrations of 1.1-10 mg/l ( > 4 pmol/l) (31). In a mouse model, bisphenol-A induced photoallergic contact dermatitis (32). If administered during organogenesis, bisphenol-A is fetotoxic when administered to pregnant mothers. A study of the effects of bisphenol-A on reproduction and fertility showed a significant reduction in seminal vesicle weight and sperm motility in treated male mice (33). These effects were present in parents and in a second generation when offspring were maintained at the same exposure level as their parents (4.37-17.5 mg/kg body weight). In addition, bisphenol-A treatment resulted in a significant reduction in the number of litters per pair and in live pups per litter.
Composites were discovered to be toxic in humans because lesions developed in unlined cavities (34,35). The concentrations to which cells and tissues are exposed in humans are unknown (11), but it was estimated that seven components of composites, including bisphenol-A, were toxic in the 10-100 pmol/l concentration range in an in culture assay (11). The sealant we tested was toxic to MCF7 breast cancer cells at concentrations > 50 jig/ml ( -200 pmol/l of bisphenol-A). Lower concentrations were not toxic but were fully estrogenic in the bioassay (Fig. 1B).
Human exposure to bisphenol-A and dimethacrylate of bisphenol-A has increased during the last 30 years because of the extensive use of bisphenol-A-based resins. The use of sealants as an effective therapy and preventive treatment for dental pits and fissures in adults and children has been endorsed by European and U.S. agencies (18,36,34). In view of the documented exposure to bis-GMA-based composites and sealants used in dental treatments for adults and children, the use of these xenoestrogens should be reevaluated.