Response to Hunt et al., invalid controls undermine conclusions of FDA studies.

Abstract

Dear Editor, The comments of Hunt et al. on our recent manuscripts (Delclos et al. (2014). Toxicol. Sci. 139, 174–197 and Churchwell et al. (2014). Toxicol. Sci. 139, 4–20, hereafter Delclos et al. and Churchwell et al.) raise issues related to both completed and ongoing studies on bisphenol A (BPA) at the National Center for Toxicological Research (NCTR). We address these concerns separately. Hunt et al. suggest that the results of the 90-day BPA subchronic study described in Delclos et al. are not interpretable with regard to effects <2700 μg/kg body weight (bw)/day, the study-defined “low dose” region, because of the data reported in the companion manuscript, Churchwell et al., indicating that there was unintended exposure to BPA in the negative controls. As pointed out in Delclos et al., exposure of the negative controls to low levels of the target compounds frequently occurs in the studies of ubiquitous environmental contaminants, with the critical factor being the differential exposure between negative controls and treatments. We concluded that the level of exposure in the negative control animals in our study, as evidenced by the presence of BPA-glucuronide, which can only be produced by inlife exposure, would compromise our ability to interpret any BPA-related treatment effects <8 μg/kg bw/day. However, this was not an issue in the Delclos et al.'s study because: (1) there were no effects observed for the endpoints measured in doses up to 2700 μg BPA/kg bw/day versus the negative controls; (2) the reference estrogen had clear effects, even though these animals were equally exposed to the unintended source of environmental BPA; and (3) clinical and histopathological observations in the negative controls were comparable to observations in multigenerational studies conducted at this institution using the same strain of rat fed the same base diet (NTP, 2008, 2010). Hunt et al. are concerned that the detection of BPA-glucuronide in the blood of control animals in the 90-day study indicates that the ongoing chronic toxicity study, which is also providing animals and tissues to multiple academic investigators (Consortium Linking Academic and Regulatory Insights on BPA Toxicity; CLARITY-BPA, Schug et al., 2013), will be of no use. The considerations discussed above would be applied to the interpretation of data from the chronic/CLARITY-BPA study. However, it should be made clear that the 90-day BPA subchronic study described in our two papers was conceived and implemented prior to the conception or formation of the chronic/CLARITY-BPA study. Although conduct at NCTR is common to both, and the issue of exposure of negative controls became evident only after the start of the chronic/CLARITY-BPA study, it is also the case that there are differences between the studies that significantly limit the potential for background exposure of controls in the ongoing study. The source of the reported unintended exposure in the 90-day study was proposed in Churchwell et al. to be related to the use of the broad BPA dose range, including very high doses (100,000 and 300,000 μg BPA/kg bw/day), in the same animal rooms as the control and low BPA doses (0 and 2.5 μg/kg bw/day). Consistent with our hypothesis, unpublished results of analyses of serum from animals in the chronic/CLARITY-BPA study rooms with a high dose of 25,000 μg BPA/kg bw/day indicate that the BPA-glucuronide levels in control rats, the critical measurement in distinguishing inlife exposure from postexposure sample contamination with aglycone BPA, are clearly distinguishable from levels in the lowest BPA dose group, 2.5 μg/kg bw/day. These data will be reported with the first results from the CLARITY-BPA studies.