Sodium bromate is a drinking water disinfection byproduct (DBP) formed during the ozonation of source water containing bromide. It was nominated for toxicity and carcinogenicity studies in genetically modified mouse models by the United States Environmental Protection Agency and the National Institute of Environmental Health Sciences to determine whether transgenic mouse models were effective at determining the potential hazards of DBPs. Male and female Tg.AC hemizygous and p53 haploinsufficient mice were evaluated as models to identify DBPs for carcinogenic potential. The combination of Tg.AC hemizygous mice and p53 haploinsufficient mice has been suggested as an effective means of identifying chemical carcinogens and assessing potential risk (Tennant et al., 1995). Tg.AC hemizygous mice were reported to respond to tumor promoters, mutagenic chemicals, and non-mutagenic chemicals while p53 haploinsufficient mice responded to mutagenic chemicals within 6 months, allowing the testing of more chemicals within a shorter period of time (Tennant et al., 1995). Sodium bromate was evaluated by the drinking water route because this was the most likely route of human exposure. Dermal studies were included for the Tg.AC hemizygous mice since it was reported that tumors usually occurred within 10 weeks of initiation of exposure (Tennant et al., 1995) and because humans are exposed dermally to DBPs through showering and bathing.

There were no increased incidences of neoplasms in male or female Tg.AC hemizygous or p53 haploinsufficient mice exposed to sodium bromate by the dermal or drinking water routes. Dermal administration of sodium bromate for 26 or 39 weeks did not result in significant treatment-related increases in neoplasms or nonneoplastic lesions of the skin of either male or female Tg.AC hemizygous mice. This is consistent with a study in the literature that demonstrated that potassium bromate was not a skin carcinogen or a promoter of skin cancer in Sencar mice (Kurokawa et al., 1984). Sodium bromate did not cause overt toxicity in Tg.AC hemizygous mice following dermal administration. Dermal penetration and subsequent systemic circulation of sodium bromate has been reported to be very low (CIREP, 1994). In the Tg.AC hemizygous mouse dermal study, nonneoplastic lesions attributed to sodium bromate were found in the kidneys, thyroid gland, spleen, and testes. Similar lesions were found in the kidneys, thyroid gland, and testes of Tg.AC hemizygous mice given sodium bromate in the drinking water. The systemic dose of sodium bromate was probably increased following dermal administration with oral consumption of sodium bromate during grooming, and it is likely that the systemic effects resulted from the oral exposure. The dose for the dermal administration of sodium bromate was higher than the dose for the drinking water studies based on consumption and formulation concentration data. For example, the dermal dose groups received 64, 128, and 256 mg/kg per day, whereas the exposure levels for the drinking water studies were approximately 13 to 15 mg/kg per day in the 80 mg/L group, 63 to 72 mg/kg per day in the 400 mg/L group, and 129 to 148 mg/kg per day in the 800 mg/L group. Although the exposure levels differed somewhat, the type and incidence of the effects suggested that the internal doses were similar between the two routes of exposure.

Nonneoplastic lesions attributed to sodium bromate exposure from the drinking water were found in the kidneys, thyroid and pituitary glands, and the testes/epididymides of the Tg.AC hemizygous mice. No neoplasms attributed to sodium bromate exposure were observed in male or female Tg.AC hemizygous mice. These results are in contrast to numerous traditional carcinogenicity studies that demonstrate that potassium bromate is a rodent carcinogen when administered in drinking water. Potassium bromate increased the incidence of renal tumors in male F344/N rats as well as the incidences of mesothelioma and thyroid gland tumors (DeAngelo et al., 1998). Although kidney tumors following potassium bromate exposure have been shown to be more prevalent in rats than mice, DeAngelo et al. (1998) also observed renal tumors in B6C3F₁ mice treated with potassium bromate in the drinking water at concentrations ranging from 0.08 to 0.8 g/L for up to 100 weeks. An increase in renal tumors was observed at the low concentration but not the mid or high concentrations (DeAngelo et al., 1998). Female B6C3F₁ mice treated for 78 weeks with 500 or 1,000 ppm of potassium bromate in the drinking water did not develop renal tumors (Kurokawa et al., 1986b). However, renal tumors were found in male B6C3F₁, BDF₁, and CDF₁ mice after exposure to 750 ppm potassium bromate in drinking water for 88 weeks (Kurokawa et al., 1990). Although these incidences were not statistically significant, the authors concluded that there was a potential for potassium bromate to induce renal cell tumors in mice. Nephropathy was the only renal abnormality attributed to sodium bromate exposure in the present transgenic mouse studies. A dose-related increase in the severity of chronic progressive nephropathy was associated with dysplastic foci and tumors in rats treated with potassium bromate for up to 2 years (Kurokawa et al., 1983a, 1986a). In contrast, DeAngelo et al. (1998) did not observe any association between potassium bromate treatment and nephropathy.

There have been several studies on the mechanism of potassium bromate induced renal carcinogenesis. The development of tumors from chronic exposure to potassium bromate is thought to result from oxidative DNA damage following bromate metabolism and subsequent lipid peroxidation (Kurokawa et al., 1990; Sai et al., 1991, 1992b; Umemura et al., 1998). Data in the literature indicate that the species differences observed in the induction of renal cell tumors are correlated with different levels of lipid peroxidation. Lipid peroxide levels were significantly increased in a dose- and time-dependent manner in the kidneys of male F344 rats after intravenous administration of potassium bromate but not in CDF₁, B6C3F₁, or BDF₁ mice (Kurokawa et al., 1987b). Bromate causes oxidative DNA damage as evidenced by the formation of 8-hydroxydeoxyguanosine (8-OH-dG) in kidney DNA. Kurokawa et al. (1990) have demonstrated a positive correlation between the formation of 8-OH-dG in rat kidney DNA and potassium bromate induced carcinogenesis. Incubation of bromate directly with DNA in vitro did not result in 8-OH-dG production indicating that bromate needs to undergo cellular metabolism to cause DNA damage (Kurokawa et al., 1990).

It is possible that the duration of treatment in the present studies was not long enough to detect renal tumors in the Tg.AC hemizygous mouse. Other chemicals that cause renal cancer in 2-year rodent studies have also failed to cause tumors in the Tg.AC hemizygous model. Three chemicals that caused kidney tumors in male B6C3F₁ mice and another chemical that produced kidney tumors in male F344 rats did not produce tumors in Tg.AC mice when administered dermally (Spalding et al., 2000). Chloroprene exposure caused increased incidences of renal tubule adenomas in male and female F344/N rats and male B6C3F₁ mice but failed to cause tumors in Tg.AC hemizygous mice after inhalation exposure (NTP, 1998; Pritchard et al., 2003). Male B6C3F₁ mice given bromodichloromethane, another drinking water disinfection by-product, by gavage for 2 years had increased incidences of kidney adenomas and adenocarcinomas (NTP, 1987). However, the Tg.AC hemizygous mouse failed to respond to doses of bromodichloromethane that exceeded those in the 2-year study (NTP, 2006a). These data suggest that the Tg.AC hemizygous model is not responsive to rodent renal carcinogens. In a review of 38 chemicals evaluated by NIEHS/NTP in genetically modified mice, 11 (20%) of the chemicals that produced tumors in the 2-year assays were not detected as carcinogens in the transgenic mouse models (Bucher, 1998). A larger review found that the Tg.AC hemizygous model has about 77% accuracy for detecting potential carcinogens (Pritchard et al., 2003).

Nonneoplastic lesions in the thyroid gland of Tg.AC hemizygous mice included follicular secretory depletion, follicular cell hypertrophy, and lymphocytic infiltration. These findings were observed in male and female Tg.AC hemizygous mice exposed to sodium bromate dermally or in drinking water. While no thyroid gland lesions were found in the 2-year potassium bromate drinking water study using B6C3F₁ mice, potassium bromate is carcinogenic in the F344/N rat thyroid, producing thyroid gland follicular adenomas and carcinomas at water concentrations as low as 0.02 g/L (DeAngelo et al., 1998). Bromate is similar in structure to chlorate and perchlorate. Both are thyroid gland toxicants and induce follicular cell tumors in rats (Hooth et al., 2001; NTP, 2005a). Most chemically induced thyroid gland neoplasms appear to result from direct interference with the synthesis of thyroid hormones, resulting in a decreased circulating level of T₃ or T₄ with subsequent elevated thyroid stimulating hormone (TSH) secretion and stimulation of thyroid cell proliferation. Follicular cell proliferation often follows a progression from hyperplasia to neoplasia (Hardisty and Boorman, 1990). The mechanism by which potassium bromate induced thyroid gland tumors is not known. Bromide binds to the sodium-iodide symporter of the thyroid gland with low affinity (Van Sande et al., 2003). High levels of bromide result in a decrease in iodide accumulation in the thyroid gland and a rise in iodide excretion by the kidneys, leading to decreased thyroid hormone synthesis and stimulation of thyroid gland cell proliferation (Pavelka, 2004).

The effect of bromate on the thyroid gland may explain the increased incidence of pars distalis hypertrophy in female Tg.AC hemizygous mice treated with 800 mg/L sodium bromate in the drinking water for 27 or 43 weeks. Treatment-related lesions of the pituitary gland in mice are uncommon and are usually a secondary effect of toxicity involving a target endocrine organ (Mahler and Elwell, 1999). Induction of par distalis lesions secondary to proliferative lesions of the thyroid follicular cells have been observed in previous chronic rodent studies (Mahler and Elwell, 1999). These findings are attributed to a reduction of circulating T₃ and/or T₄ which stimulates cells by a feedback mechanism in the hypothalamus and pars distalis of the pituitary gland to increase the production of corresponding releasing factors and stimulating hormones (Mahler and Elwell, 1999). Thus, the effects of sodium bromate on the pituitary gland may be secondary to the changes that occurred in the thyroid gland.

Absolute testis weights were significantly decreased in Tg.AC hemizygous males exposed to sodium bromate dermally for 39 weeks or in the drinking water for 43 weeks. The incidences of germinal epithelium degeneration increased with increasing exposure concentration. At 43 weeks, the incidences of epididymal tubule degeneration were significantly increased in the 800 mg/L males. The tubules with degeneration had reduced numbers of spermatozoa. These observations are consistent with previous NTP reproductive toxicity studies demonstrating decreases in sperm density in male Sprague-Dawley rats exposed to sodium bromate in the drinking water (NTP, 1996; NTP, 2001). The mechanism of this effect is not known.

Clinical pathology parameters indicated that sodium bromate had minimal treatment-related effects on the red blood cells. Tg.AC hemizygous mice exposed to sodium bromate dermally or in drinking water exhibited decreases in erythron (hemoglobin, hematocrit, and red cell counts), mean cell hemoglobin (MCH), and mean cell hemoglobin concentration (MCHC), which may indicate a weak depressive effect on bone marrow red cell production. An apparent regenerative-type response was indicated by an increase in reticulocyte counts. Increased hematopoiesis was also indicated by significant increases in the percentage of polychromatic erythrocytes in Tg.AC hemizygous mice treated with sodium bromate. The incidences of splenic hematopoietic cell proliferation increased with increasing dose, and significant increases occurred in female Tg.AC hemizygous mice dosed dermally with 128 and 256 mg/kg sodium bromate for 26 weeks. With the exception of a small decrease in MCH, hematology parameters were not generally altered in p53 haploinsufficient mice. For the Tg.AC hemizygous mice, there appeared to be a very mild treatment-related anemia that was accompanied by a regenerative response. The mechanism for the anemia is unknown; however, the bone marrow and the spleen were able to respond to the lowered erythron.

Significant dose-related increases in micronucleated erythrocytes, indicative of induced chromosomal damage in the form of breaks or chromosome loss, were observed in all three experimental groups of mice (Appendix D). Increases in the percentage of polychromatic erythrocytes (PCEs) in peripheral blood, paralleling the observations of regenerative responses to the erythron decreases in Tg.AC hemizygous mice, were also observed. A significant increase in the percentage of PCEs indicates a stimulation of erythropoiesis and this may be important in interpreting experimental results, particularly weak positive responses in the micronucleus assay, as an enriched PCE population may artificially inflate a weak response. Increased erythropoiesis has been shown to slightly elevate the frequency of micronucleated PCEs as a consequence of accelerated cell division producing an increase in mitotic errors (Hirai et al.,1991; Suzuki et al., 1989). In this study with sodium bromate, significant increases in micronucleated normochromatic erythrocytes (MN-NCE) were seen in Tg.AC hemizygous mice at doses lower than those that resulted in significant increases in the percent PCEs, and significant increases in MN-NCE were observed in p53 haploinsufficient mice in the absence of any significant alteration in PCE/NCE ratios at any dose level. Thus, it does not appear that stimulation of erythropoiesis alone is responsible for the observed increases in the frequency of micronucleated erythrocytes in Tg.AC hemizygous and p53 haploinsufficient mice.

No neoplastic or nonneoplastic findings were observed in the p53 haploinsufficient mice exposed to sodium bromate in the drinking water. The absence of a carcinogenic response in p53 haploinsufficient mice, in light of the positive micronucleus test results, is surprising because a strong correlation has been reported between positive results in subchronic peripheral blood micronucleus tests and rodent carcinogenicity (Witt et al., 2000). Although the number of positive studies from which this correlation derives is small, additional evidence was provided by Morita et al. (1997), who reported a 90.5% correlation between carcinogenic activity in humans and positive results in the rodent micronucleus test when data were corrected for known structure-activity considerations with regard to micronucleus assay sensitivity. In addition, Zeiger (1998) reported a 70% correlation between rodent carcinogenicity and positive results in the mouse bone marrow micronucleus test in an unadjusted dataset of 83 NTP chemicals. Other chemicals, including diisopropylcarbodiimide (NTP, 2006b) and aspartame (NTP, 2005b), have produced significant increases in MN-NCE in the absence of evidence of carcinogenicity in p53 haploinsufficient mice. It is unlikely that the negative response was due to low exposure levels. Based on earlier studies, the concentrations used for exposure in the present study produced renal tumors in rodents, including mice. In addition, the actual exposure levels were similar to the doses used for the Tg.AC mice where minimal toxicity was observed. Therefore, the data indicate that the p53 haploinsufficient mouse was not responsive to sodium bromate exposure even though sodium bromate is genotoxic.

In conclusion, no neoplastic lesions were attributed to sodium bromate, a known rodent carcinogen, in Tg.AC hemizygous and p53 haploinsufficient transgenic mice exposed dermally or through the drinking water. These studies provide evidence that these transgenic mouse models are not a sensitive and rapid means of assessing potential toxicity and carcinogenicity of sodium bromate.

Footnotes

*

Explanation of Levels of Evidence of Carcinogenic Activity is on page 11. A summary of the Technical Reports Review Subcommittee comments and the public discussion on this Report appears on page 13.