Subchronic and chronic reference values (RfVs) were derived for 1,3-butadiene (BD) based upon its ability to cause reproductive and developmental effects observed in laboratory mice and rats. Metabolism has been well-established as an important determinant of the toxicity of BD. A major challenge to human health risk assessment is presented by large quantitative species differences in the metabolism of BD, differences that should be accounted for when the rodent toxicity responses are extrapolated to humans. The methods of Fred et al. (2008)/Motwani and Törnqvist (2014) were extended and applied here to the noncancer risk assessment of using data-derived extrapolation factors to account for species differences in metabolism, as well as differences in cytotoxic potency of three BD metabolites. This approach made use of biomarker data (hemoglobin adducts) to quantify species differences in the internal doses of BD metabolites experienced in mice, rats and humans. Using these methods, the dose-response relationships in mice and rats exhibit improved concordance, and result in subchronic and chronic inhalation reference values of 29 and 10 ppm, respectively, for BD. Confidence in these reference values is considered high, based on high confidence in the key studies, medium-to-high confidence in the toxicity database, high confidence in the estimates of internal dose, and high confidence in the dose-response modeling.
Bibliographical noteFunding Information:
This work was funded by the Olefins Panel of the American Chemistry Council (contract number 10003558 ). Representatives from member companies of this panel, including author CMN, were provided an opportunity to review the draft manuscript to ensure clarity and completeness, but had no influence on the design, analysis, interpretation, or conclusions of this work. Authors CRK and SMH are owners of Summit Toxicology, and therefore have a financial interest in this work.
This work was funded by the Olefins Panel of the American Chemistry Council (contract number 10003558); 700 2nd Street NE | Washington , DC | 20002.
Metabolism of BD to multiple reactive metabolites (2,3-epoxy-1-butene or EB, 1,2,3,4-diepoxybutane or DEB, 3,4-epoxybutane1,2-diol or EBD) plays an important role in the mode of action (MOA) for its noncancer endpoints. For example, repeated intraperitoneal injection of mice and rats to BD metabolites results in ovarian atrophy, with DEB producing effects in both species and EB producing effects only in mice; this is presumably due to greater metabolic conversion of EB to DEB in this species (Doerr et al., 1995, 1996). The weight of evidence of species differences in the metabolic activation of BD is robust (see Section 2.2). Due to large species differences in metabolism, mice, rats, and humans are exposed to vastly different internal doses of reactive metabolites, which underlies observed species differences to sensitivity to BD toxicity. Motwani and Törnqvist (2014) relied upon metabolite-specific biomarkers (e.g., hemoglobin adducts), second order rate constants for adduct formation, and species-specific erythrocyte lifespans to quantify species differences in internal doses of BD metabolites (e.g., AUC for BD metabolites in blood) in mice, rats, and humans. This approach reflects the best available science for BD, and was extended and applied here to support the application of data-derived extrapolation factors (DDEFs; USEPA, 2014) to replace default uncertainty factors to account for species differences in BD metabolic activation. Benchmark dose (BMD) methods were used to model robust data sets for the reproductive and developmental effects of BD in terms of human equivalent concentrations (HECs), as an improvement over the use of NOAEL or LOAEL values to derive of noncancer reference values (RfVs) for BD (Haber et al., 2018).This section provides background information on previous dose-response assessments, metabolism, and mode of action (MOA) studies for BD. MOA judgements are used to support key decisions in the noncancer dose-response assessments conducted for BD. In so doing, this information reduces the uncertainty associated with using relevant rodent toxicity data for BD human health risk assessment.Previous noncancer risk assessments conducted by regulatory agencies and risk assessors are summarized in Table 1. Nearly all of these assessments make use of BMD modeling, which has many advantages over the use of NOAEL or LOAEL values (USEPA, 2012). There is a high degree of consistency across agencies with respect to the endpoints of interest and the data sets serving as the basis for noncancer risk assessment. Nearly all acute and short-term noncancer assessments are based on the effects of BD on fetal body weights in exposed mice (Hackett et al., 1987a; Green, 2003). Similarly, nearly all chronic noncancer assessments are based on the effects of BD on ovarian atrophy in mice (NTP, 1993). The noncancer risk assessments, however, do differ in several areas. First, there are differences in the benchmark response rates (BMR: ranging from 1% to 10%) used to define the point of departure (POD), some of which may reflect programmatic differences across agencies for this decision point (Holman et al., 2017). Second, uncertainty factors (UFs) utilized in the risk assessments based on fetal body weight and ovarian effects of BD ranged broadly from a net value of 10 to 1000 (Table 1). Lastly, the assessments differ in their approach to quantify species differences in metabolism of BD. Species differences in the internal dose of the parent chemical were addressed in the OEHHA assessment for BD, which relied upon PBPK modeling to derive a data derived extrapolation factor (DDEF) of 1.68 (calculated as the ratio of predicted human blood concentration:predicted animal blood concentration for parent chemical for equivalent exposures to BD in air) (OEHHA, 2013). Kirman and Grant (2012) utilized hemoglobin adduct efficiencies (i.e., adduct burden per ppm*hour of BD exposure) to quantify large differences among mice, rats, and humans with respect to the internal dose of BD metabolite, diepoxybutane (DEB). Notably, all other assessments relied upon a default approach of relying on the concentration of BD in air, to which additional uncertainty factors were applied to account for interspecies differences. Accounting for quantitative differences in BD metabolism between species represents an important challenge for human health risk assessment, including those planned for BD under TSCA regulations (USEPA, 2020). This challenge was recognized by ATSDR in their toxicological profile for BD (ATSDR, 2012), which states that noncancer reference values (minimal risk values or MRLs) were not derived for BD “due to the large species differences in the metabolism of 1,3-butadiene and the lack of chemical-specific data to adjust for these differences, which may result in the MRL overestimating the risk to humans.” More recent data and methods published since the time of these assessments now allow for these important differences to be addressed quantitatively in human health noncancer risks assessments for BD.Oxidation reactions for BD generally leads to metabolic activation, resulting in the formation of reactive epoxides (e.g., EB, DEB, EBD), which can alkylate cellular macromolecules (see Section 2.3 below). Oxidation reactions are catalyzed by cytochrome P450, particularly CYP2E1, although other isozymes (e.g., CYP2A6) and enzyme systems can contribute (Deuscher and Elfarra, 1994). Oxidation reactions follow Michaelis-Menten kinetics (i.e., saturable at high concentrations), and cofactors include a source of reducing equivalents (NADPH) and molecular oxygen. There are substantial species differences in oxidative metabolism (mouse > rat > human), which is consistently supported by data from studies quantifying BD metabolism in cell fractions (Csanády et al., 1992; Schmidt and Loeser, 1985), whole cells (Krause and Elfarra, 1997; Seaton et al., 1995), by organ perfusion (Filser et al., 2001, 2010), and in whole animals (Himmelstein et al., 1994, 1995; Thornton-Manning et al., 1995a,b; Filser et al., 2007). Because available evidence strongly supports a conclusion that metabolic activation of BD is lower in humans than it is in rodents (Swenberg et al., 2011; Boysen et al., 2012; Motwani and Törnqvist, 2014), humans are generally expected to be at lower risk of metabolite-mediated toxicity for a given external exposure to BD.To support decisions made in deriving RfVs, data regarding the MOA for the noncancer effects of BD are briefly summarized below.An MOA for BD-mediated ovarian toxicity, for which ovarian atrophy is identified as a late key event, is described in more detail in Kirman and Grant (2012). Briefly, there is strong evidence that the ovarian atrophy effects of BD are mediated by the formation of its diepoxide metabolite, DEB (Doerr et al., 1995, 1996). Ovarian toxicity was observed following exposure to diepoxides (DEB; vinylcyclohexene diepoxide, a metabolite of BD-dimer vinylcyclohexene) and diepoxide precursors (EB, 4-vinylcyclohexene, vinylcyclohexene epoxide, isoprene). Ovarian atrophy was absent following exposure to structural analogues that do not form diepoxides (ethylcyclohexene oxide, vinylcyclohexane oxide, cyclohexene oxide) (Doerr et al., 1995, 1996). Although the molecular mechanism is not clearly understood, these diepoxides act as bifunctional alkylating agents (i.e., capable of forming two adducts, including DNA-DNA crosslinks, DNA-protein crosslinks; whereas monofunctional alkylating agents such as EB and EBD are capable of forming a single adducts) and appear to selectively destroy the primordial and primary ovarian follicles via programmed cell death or apoptosis, thereby accelerating the normal process of atresia (Springer et al., 1996; Hoyer and Sipes, 2007). Accelerated oocyte depletion leads eventually to premature ovarian failure and cessation of the estrous cycle. Based upon a review of the available information, there is sufficient information to establish a MOA for the ovarian effects of BD in rodents.Reports of ovotoxicity (i.e., depletion of primordial and primary follicles, a key event in the MOA that precedes ovarian atrophy; Kirman and Grant, 2012) have been made in nonhuman primates exposed to a structurally similar diepoxide (vinylcyclohexene diepoxide; Appt et al., 2006). Observations made in a surrogate species that is more closely related to humans suggests that qualitatively the endpoint of rodent ovarian toxicity following BD exposure is relevant to human health. The detection of low levels of pyr-Val in humans (Swenberg et al., 2011; Boysen et al., 2012) indicates that humans are capable of producing DEB, albeit at much lower levels than in rodents. Overall, there are clear quantitative differences across species with respect to the internal dose of DEB that need to be considered when extrapolating rodent ovarian atrophy data to humans for risk assessment purposes.The ovarian effects of BD have been well studied in mice exposed via inhalation to 6.25–625 ppm BD (NTP, 1993), and have been used as basis for noncancer RfVs by multiple agencies and risk assessors (USEPA, 2002a; OEHHA, 2013; TCEQ, 2015, Table 1). Data are also available for similarly exposed rats, which have consistently shown a lack of ovarian effects following exposures ranging from 300–8000 ppm BD (Owen et al., 1987; Bevan et al., 1996; Marty et al., 2021). Although ovarian atrophy has not been reported in rats exposed to BD (Owen et al., 1987; Bevan et al., 1996), ovarian toxicity is clearly observed when rats are directly injected with DEB for 30 days (Doerr et al., 1996); thus rats are also considered to be a responsive species for this endpoint. Quantal data sets (incidence for ovarian atrophy) for both species were used separately and combined to support RfV derivation for BD. The combined dose-response data set includes observations in 995 rodents, spanning durations ranging from approximately 9 to 105 weeks of exposure (Table 2). As noted above, differences in sensitivity between mice and rats to the ovarian effects of DEB have been observed (mice > rats) (Doerr et al., 1995, 1996). To support characterizing ovarian atrophy data from both species with a single dose-response relationship, an EFAD value of 0.088 (i.e., rats are estimated to be approximately 11-fold less sensitive to DEB than mice based on toxicodynamic differences) was applied to rat data points based on upon potency differences for DEB using the data of Doerr et al. (1995) (see Appendix A). This approach represents a small departure for USEPA guidelines to calculating HEC values (USEPA, 2014) and may be viewed as conservative since it adopts an assumption that humans are as sensitive as mice rather than rats (i.e., HEC values for mice were not adjusted dividing by an EFAD of 0.088 to express them in terms of rat sensitivity).Overall, confidence in the subchronic and chronic RfVs we derived for BD is high. The key data sets are defined by well-conducted studies that have been consistently selected by regulatory agencies to support noncancer risk assessments for BD. Furthermore, we based both subchronic and chronic RfVs upon multiple studies utilizing data from both rats and mice, which improves the overall confidence of the assessment in predicting responses across mammalian species (e.g., in humans). Confidence in the dosimetry of the assessment is also considered high, as doses were calculated from excellent biomarker data that are metabolite-specific and have been quantified in all three species of interest (mice, rats, and humans). Confidence in the toxicity database is considered medium-to-high, since the toxicity of BD has been well-studied and the database is considered robust, albeit caveated by some sources of uncertainty (as discussed below).Although an acute RfV was not specifically derived here for assessing single day exposures to BD, the subchronic RfV can also be used as a health-protective surrogate to assess acute exposures to BD. This practice is consistent with the use of fetal body weight effects to derive acute RfVs for BD by other agencies (Table 1), and it is considered health-protective due to differences in exposure duration (e.g., a single day exposure that reflects a small fraction of the human gestation period vs. a 10-day exposure that reflects a large fraction of the rodent gestation period). Alternatively, USEPA's Acute Exposure Guideline Levels (AEGLs), which describe the human health effects from rare exposure to airborne chemicals (e.g., chemical spills), could be considered. AEGL values derived by USEPA for BD are considerably higher than the RfVs derived here (AEGL1 = 670 ppm; AEGL2 = 2700 ppm; AEGL3 = 6800 ppm) and could also be used to support acute risk assessments for single-day exposures to BD (USEPA, 2009).This work was funded by the Olefins Panel of the American Chemistry Council (contract number 10003558); 700 2nd Street NE | Washington, DC | 20002.This work was funded by the Olefins Panel of the American Chemistry Council (contract number 10003558). Representatives from member companies of this panel, including author CMN, HS, and NE were provided an opportunity to review the draft manuscript to ensure clarity and completeness, but had no influence on the design, analysis, interpretation, or conclusions of this work. Authors CRK and SMH are owners of Summit Toxicology, and therefore have a financial interest in this work.This work was funded by the Olefins Panel of the American Chemistry Council (contract number 10003558). Representatives from member companies of this panel, including author CMN, were provided an opportunity to review the draft manuscript to ensure clarity and completeness, but had no influence on the design, analysis, interpretation, or conclusions of this work. Authors CRK and SMH are owners of Summit Toxicology, and therefore have a financial interest in this work.
© 2022 The Authors
- Benchmark dose
- Data derived extrapolation factors
- Hemoglobin adducts
- Internal dose
- Mixtures assessment
- Noncancer reference concentration
- Pooled analysis
- Species differences
PubMed: MeSH publication types
- Journal Article