Review of the toxicology, human exposure and safety assessment for bisphenol A diglycidylether (BADGE).

BADGE (whose chemical names are bisphenol A diglycidylether and 2,2-bis(4-(2,3-epoxypropyl)phenyl)propane) is the lowest molecular weight oligomer in commercial epoxy resins and the major component in commercial liquid epoxy resins. The major application areas for epoxy resins are protective coatings and civil engineering. Additional applications include printed circuit boards, composites, adhesives and tooling, while a relatively small amount of epoxy resins (< 10%) finds use in protective coatings inside food and drink cans. The use of BADGE in food-contact applications was first regulated through EC Directive 2002/16/EC and amended in EC Directive 2004/13/EC with migration levels in food-contact applications being generally well below the regulatory thresholds. The paper discusses the commercial use of BADGE focusing on the current knowledge of human exposure from canned food applications. To assess the safety of this application, the exposure data are compared with no adverse effect levels (NOAEL) from various toxicological investigations with BADGE including reproductive and developmental assays, endocrine toxicity investigations, and sub-chronic and chronic assays. Consumer exposure to BADGE is almost exclusively from migration of BADGE from can coatings into food. Using a worst-case scenario that assumes BADGE migrates at the same level into all types of food, the estimated per capita daily intake for a 60-kg individual is approximately 0.16 microg kg(-1) body weight day(-1). A review of one- and two-generation reproduction studies and developmental investigations found no evidence of reproductive or endocrine toxicity, the upper ranges of dosing being determined by maternal toxicity. The lack of endocrine toxicity in the reproductive and developmental toxicological tests is supported by negative results from both in vivo and in vitro assays designed specifically to detect oestrogenic and androgenic properties of BADGE. An examination of data from sub-chronic and chronic toxicological studies support a NOAEL of 50 mg kg(-1) body weight day(-1) from the 90-day study, and a NOAEL of 15 mg kg(-1) body weight day(-1) (male rats) from the 2-year carcinogenicity study. Both NOAELS are considered appropriate for risk assessment. Comparing the estimated daily human intake of 0.16 microg kg(-1) body weight day(-1) with the NOAELS of 50 and 15 mg kg(-1) body weight day(-1) shows human exposure to BADGE from can coatings is between 250,000 and 100,000-fold lower than the NOAELs from the most sensitive toxicology tests. These large margins of safety together with lack of reproductive, developmental, endocrine and carcinogenic effects supports the continued use of BADGE for use in articles intended to come into contact with foodstuffs.

Subchronic toxicity study with ethylene-bis-(oxyethylene)-bis-(3-tert-butyl-4-hydroxy-5-methylhydro cin namate) in the cynomolgus monkey: lack of stimulation of the pituitary-thyroid-liver axis.

To evaluate the toxicological profile of the phenolic antioxidant ethylene-bis-(oxyethylene)-bis-(3-tert-butyl-4-hydroxy-5-methyl- hydrocinnamate) (EOC) in a non-human primate, male cynomolgus monkeys (Macaca fascicularis) were treated for 4 weeks by oral administration of 0, 200, or 1000 mg/kg body weight/day. Special attention was directed to parameters of the pituitary-thyroid-liver axis. Moderately increased liver weights and minimal to moderate hepatocellular hypertrophy were observed in treated animals. Otherwise, no treatment-related changes were detected in hematological, clinical chemistry, or urinalysis parameters or upon histopathological examination. Except for a slight induction of microsomal testosterone 16beta-hydroxylation, liver xenobiotic-metabolising enzyme activities and peroxisomal fatty acid beta-oxidation remained unchanged. Likewise, serum levels of thyroid stimulating hormone, thyroxine, 3,3',5-triiodothyronine and 3,3',5'-triiodothyronine as well as 5'-monodeiodinase type 1 mRNA levels in the liver, heart, cerebral cortex, and thyroid were found unchanged. The results demonstrate that, in the Cynomolgus monkey, EOC is only a very weak inducer of liver xenobiotic-metabolizing enzymes and has no effect on thyroid function. In contrast, upon feeding rats at dose levels up to 1000 ppm (equivalent to between 50 and 100 mg/kg body weight/day), EOC has been identified as a strong phenobarbital- and peroxisome proliferator-type inducer of hepatic xenobiotic-metabolizing enzymes, interfering with thyroid hormone homeostasis, causing thyroid follicular hypertrophy, and, upon chronic treatment, inducing thyroid gland follicular cell tumors (Thomas et al., 1995. In Toxicology of Industrial Compounds, pp. 319-339. Taylor and Francis). Thus, the results of this study with EOC in the cynomolgus monkey show that effects of xenobiotics on the pituitary-thyroid-liver axis as frequently observed in rodents can not necessarily be extrapolated to primates including man.

Determination of dichlorobenzidine-hemoglobin adducts by GC/MS-NCI.

A method based on gas chromatography/mass spectrometry-negative ion chemical ionization detection (GC/MS-NCI) was developed for the determination of 3,3'-dichlorobenzidine (DCB)-hemoglobin adducts. Adducts were released from hemoglobin by mild alkaline hydrolysis and determined by GC/MS-NCI after extraction and derivatization with heptafluorobutyric anhydride (HFBA). 2,2'-DCB was used as internal standard and the recovery of the diarylamine derivatives in the overall procedure was 65-88%. The limit of detection attained was below 0.1 ng/g hemoglobin for DCB as well as for the metabolite N-acetyl-3,3'-dichlorobenzidine (acDCB). The method was shown to be linear up to 150 ng/g hemoglobin. In the NCI mass spectra of the HFB derivatives the dominant ion is (M-HF)-. Due to the presence of two chlorines in the diarylamines, the characteristic ratio of 1.5 for m/z 624 to 626 (for diHFB-DCB and diHFB-2,2'-DCB) and m/z 470 to 472 (for HFB-acDCB) can be observed and used for identification. The method was applied to the determination of DCB-hemoglobin adducts formed in young female Wistar rats after treatment for 4 weeks with 0.006%, 0.0012% or 0.00024% DCB via the drinking water. Two adducts were detectable by GC/MS-NCI after alkaline hydrolysis of hemoglobin samples, extraction and derivatization. The structure of these adducts could be assigned to DCB and acDCB by co-chromatography with the synthetic standards and by the presence of the characteristic ion (M-HF)-. Assessment of the time dependence of hemoglobin adduct formation during subchronic treatment with DCB revealed an increase in adduct levels during weeks 1-3. After this time adduct levels essentially remained constant. In hemoglobin samples isolated from animals treated for 4 weeks with DCB a dose-proportional increase in the total amount DCB- and acDCB-hemoglobin adducts from 8.1 ng DCB/g hemoglobin at 0.3 mg/kg body weight per day (0.00024% in drinking water) to 159.9 ng DCB/g hemoglobin at 5.8 mg/kg body weight per day (0.006% in drinking water) was observed. The ratio of the DCB adduct to the acDCB adduct was strongly dose dependent. At low DCB doses the acDCB- and DCB adducts were formed at similar levels, whereas at high DCB doses the DCB adduct was predominant.

Lack of bioavailability of dichlorobenzidine form diarylide azo pigments: molecular dosimetry for hemoglobin and DNA adducts.

The hypothetical release of 3,3'-dichlorobenzidine (DCB) from two insoluble azo pigments and from a soluble azo dye was investigated in female Wistar rats for a 4 week treatment with 0.2% (w/w) Colour Index Pigment 13 (PY13) or 0.2% (w/w) Colour Index Pigment Yellow 17 (PY17) in the diet or 0.06% (w/v) Colour Index Direct Red 46 (DR46) in the drinking water. Steady-state DCB-hemoglobin adduct levels were determined by GC/MS with negative chemical ionization as well as DCB-DNA adduct levels in the liver by (32)P-postlabelling and compared with the respective adduct levels obtained in animals after treatment for 4 weeks with 0.00024, 0.0012 or 0.006% (w/v) DCB in the drinking water. A dose-proportional increase in adduct levels from 8.1 ng/g hemoglobin and 2.6 ng/g DNA (relative adduct level, RAL, 3.3x10(-9)) to 160 ng/g hemoglobin and 45.4 ng/g DNA (RAL 56.1x10(-9)) was observed in the DCB-treated rats. In rats treated with DR46 total adduct levels of 17.7 ng/g hemoglobin and 5.2 ng/g DNA (RAL 6.4x10(-9))were determined. No hemoglobin of DNA adducts were found in rats treated with PY17 in the diet, at a limit of detection of 0.1 ng/g hemoglobin and 0.08 ng/g DNA (RAL 0.1x10(-9)). In animals treated with PY13 in the diet no adducts or only minimal amounts slightly above the limit of detection could be identified. Taking into consideration that PY13 was contaminated with 0.02% of the respective soluble monoazo compound, it is concluded that the small amounts of DCB detected have been released from the contaminating soluble monoazo compound and not from insoluble PY13. The results of the present study demonstrate the lack of bioavailability of DCB from the diarylide azo pigments PY17 and PY13.

Characterization of Drosophila DNA-binding protein DB-2: demonstration of its sequence-specific interaction with DNA.

A DNA-binding protein (DB-2) was isolated from unfertilized Drosophila eggs by DNA-cellulose chromatogrphy. In competition assays with DNA from other species, DB-2 preferentially binds to Drosophila DNA. This binding protein can also be isolated from pupal nuclei and comprises only a small fraction ( < 0.01%) of the total nonhistone chromosomal proteins. In order to investigate the specificity of the interaction between DB-2 and the DNA, we attempted to isolate the DNA sequences to which DB-2 binds. DB-2 was used as a probe to screen our gene bank established by inserting randomly sheared fragments of Drosophila DNA into bacterial plasmids. Groups of plasmids were tested for binding to DB-2 by a filter binding assay. The plasmids bound to the nitrocellulose filter were eluted and used for bacterial transformation. After several cycles of transformation and cloning, two plasmids, A17 and B10, were isolated that bind DB-2 specifically, as measured by filter binding and competition assays. In plasmid A17, binding of DB-2 protects two short DNA segments of approximately 13 and 30 base pairs from digestion by DNase I. By filter hybridization according to Southern, these sequences were mapped to a defined restriction fragment. Further evidence for the binding specificity was obtained by visualizing the protein-DNA complex in the electron microscope. In salivary gland giant chromosomes, A17 DNA hybridizes to a single site (95A/B) on chromosome 3.

A new method for the purification of DNA-binding proteins with sequence specificity.

We describe a technique for a rapid and efficient isolation and purification of proteins binding to defined DNA sequences. Cloned double-stranded DNA was covalently coupled to m-aminobenzyloximethylcellulose in order to purify proteins which recognize and bind to specific sequences on the DNA. The purification of two DNA-binding proteins from Drosophila melanogaster is demonstrated using the respective cloned DNA sequences.

Purification of interferon from mouse Ehrlich ascites tumor cells.

Interferon production was induced in mouse Ehrlich ascites tumor cells by infection with Newcastle disease virus. The interferon produced was purified by precipitation with ammonium sulfate, chromatography on carboxymethyl-Sephadex, treatment with blue dextran and polyethylene glycol, gel filtration on Bio-Gel P-60 and Bio-Gel P-200, chromatography on phosphocellulose, isoelectric focusing, and chromatography on octyl-Sepharose. The specific activity of the product was 1.6 x 10(9) NIH mouse interferon reference standard units/mg of protein. Electrophoresis in polyacrylamide gels in the presence of sodium dodecyl sulfate indicated that the apparent molecular weight of the interferon-active material ranged from 25,000 to 35,000. As revealed by staining the gels with Coomassie brilliant blue, the interferon activity co-migrated with the major, broad protein band. Minor, stainable bands of proteins were free of interferon activity and their apparent molecular weight was smaller than 12,000.