Toxic Equivalency Factors
Assumptions, variability, and uncertainty are well delineated in Part II of the Reassessment1 that addresses critical considerations in the application of the toxic equivalency factor (TEF) method (Part II, Chapter 9, section 9.2.6, p. 9-10). In addition, conclusions in Part III appear to be congruent with discussions in Part II, Chapter 9, and in the Reassessment overall. No major omissions were identified in the Reassessment, but several aspects need to be addressed or updated.
The compounds that are the focus of the Reassessment include 7 of 75 polychlorinated dibenzo-p-dioxins (PCDDs), 10 of 135 polychlorinated dibenzofurans (PCDFs), and of the total 209 polychlorinated biphenyls (PCBs), only 4 of the 122 previously defined as 2,3,7,8-tetrachlorodibenzo-
p-dioxin (TCDD)-like by the World Health Organization (WHO) (van den Berg et al. 1998). The toxic potency of each of these DLCs (their TEFs) is expressed relative to that of TCDD (also referred to as dioxin), the most potent member of this chemical class (Part III, Table 1.3, p. 1-20). These chemicals are classified as DLCs, given their similarity in chemical structure and physiochemical properties, their ability to invoke a common battery of toxic responses by a common aromatic hydrocarbon receptor (AHR)–dependent mechanism in vivo, and their ability to be persistent environmentally and to bioaccumulate. The lack of inclusion of the eight mono-ortho PCBs previously assigned TEFs by WHO in the Reassessment at the present time is due to concerns that the previously reported activity of many of these chemicals might have been primarily or partially due to a dioxin-like PCB contaminant (PCB126) present in these mono-ortho PCB preparations (DeVito et al. 2003). Although some AHR-dependent toxic effects have been observed with mono-ortho PCBs prepared by methods that should not produce the more toxic DLCs, it remains to be determined whether most of the reported toxicological effects and resulting relative potency (REP) values of these chemicals are due to contaminants, mono-ortho PCBs, or both. Given this uncertainty and the fact that reanalysis of the mono-ortho PCBs as pure compounds is currently being reexamined, they were not included in the list of relevant DLCs for consideration in the Reassessment. Once these issues are resolved, the mono-ortho PCBs should be considered in a follow-up to the Reassessment if they are documented to produce AHR-dependent toxic effects.
MAJOR ISSUES, ASSUMPTIONS, AND UNCERTAINTIES
The relative toxicological and biological potency of a complex mixture is assessed by the TEF approach. Current TEFs are “order-of-magnitude” qualitative values for dioxins, other than TCDD, and DLCs that were established by a WHO expert scientific panel that examined a large scientific database of REP estimates from in vivo and in vitro studies of the biochemical and toxic effects. In the TEF approach, the concentration of the individual compound present in the mixture (determined by instrumental analysis) are multiplied by their specific TEF value, and the sum is expressed as the TCDD toxic equivalent quotient (TEQ). Summation of the calculated TEQs for all active TCDD-related compounds in a sample extract yields the total TEQ for the specific sample extract. Numerous assumptions underlie the use of the TEF/TEQ approach; these have been well delineated, and the major aspects are discussed in detail in the Reassessment (Part II, Chapter 9, and Part III, Chapter 1, section 1.2). These assumptions and uncertainties are described and discussed below.
Role of AHR
Assumption: AHR mediates most toxicities produced by TCDD and other PCDDs, PCDFs, and coplanar PCBs that are AHR agonists. Although AHR is necessary, the ability of TCDD, other dioxins, and DLCs to produce their biochemical and toxicological effects results from downstream events regulated by AHR and AHR-dependent gene expression. The role of AHR in the toxic and biological effects of the TCDD, other dioxins, and DLCs has been supported by a substantial number of quantitative structure-activity relationship, biochemical, genetic, and targeted Ahr knockout studies.
AHR-Independent Mechanisms Excluded
Assumption: Effects mediated by other mechanisms (AHR independent) and interactions with other chemicals are ignored. AHR-independent effects of TCDD have been previously observed, including effects on intracellular calcium levels (Puga et al. 1997), changes in gene expression (Oikawa et al. 2001), and selected toxicity in Ahr knockout mice (Fernandez-Salguero et al. 1996; Lin et al. 2001). Whether all TCDD-related compounds produce these effects is unknown. Although these mechanisms may play a role in the biochemical effects of TCDD, other dioxins, and DLCs, their significance and role in the overall toxic effects of these compounds remain to be established. However, the Reassessment should acknowledge that AHR-independent effects of TCDD occur and that future studies might demonstrate a role for these effects in the overall toxic and biological effects of TCDD, other dioxins, and DLCs.
Uncertainty of TEF Values
Considering the uncertainty in selection of the TEFs and the information presented on REPs and TEFs in the Reassessment, the 2000 EPA Science Advisory Board (SAB) Panel “questioned whether the uncertainty in the TEFs and the application of this approach to predicting risks due to current levels of exposure was adequately presented” (EPA SAB 2001, p. 29). They concluded that the Reassessment should acknowledge the need for better uncertainty analysis of the TEF values, and although no current method for doing so has been endorsed by the scientific community, several approaches were suggested, such as the use of probabilistic distributions of TEF values in TEQ evaluation (Finley et al. 2003). Available information indicates a considerable amount of variability in the REP value data that were used to derive the WHO TEF values. In addition, although the WHO TEFs were derived based on a scientific consensus evaluation of the avail-
able REP values using defined weighted criteria for individual studies, details of the quantitative basis of this weighting scheme were not clearly presented in the description publication (van den Berg et al. 1998). These issues would contribute to variability and uncertainty in the application of the WHO TEF values to health risk assessment. Application of a mathematical value or percentage of the overall range of REP values, such as those described by Finley et al. (2003), would be one way to make the process of determining the specific TEFs more transparent and to provide a standard method to develop TEFs for other TCDD-related compounds that may be added at a later date. Some members of the 2000 EPA SAB Panel also recommended “that, as a follow up to the Reassessment, EPA should establish a task force to build ‘consensus probability density functions’ for the thirty chemicals for which TEFs have been established, or to examine related approaches such as those based on fuzzy logic” (EPA SAB 2001, p. 29). The committee strongly recommends that the EPA consider inclusion of uncertainty analysis of the TEF values as a follow-up to the current Reassessment.
Consistency of DLC REP Values
Assumption: The REP of a chemical in this group is presumed to be equivalent for all end points of concern and for all exposure scenarios, and all are full agonists. Although most in vitro and in vivo studies support this assumption, the 2000 EPA SAB Panel noted in their review of the Reassessment (EPA SAB 2001) that there are reports of significant differences between the potency of some dioxins, other than TCDD, and some DLCs and specific “toxic end points” is illustrated in Table 5-4 and Table 2-4 in the Integrated Summary (SAB 2001, p. 31). For example, the panel indicated that “1,2,3,7,8- PeCDF (pentachlorodibenzofuran) has the same tumorigenicity as TCDD but was ~38 times weaker for teratogenicity; the other congener, 2,3,4,7,8-PeCDF had half the tumorigenic potency as TCDD, but is ~8 times less potent for teratogenicity” (EPA SAB 2001, p. 31). However, although it was noted that no other examples of that difference were presented in the Reassessment, the observations did raise some concerns about whether all toxic end points could be combined into a single TEF value. The 2000 EPA SAB Panel suggested that “because TEFs vary among different endpoints as well as congeners, it would also be helpful for the document to note that, as data becomes available, it may be possible to derive TEQs [and TEFs] for different endpoints” (EPA SAB 2001, p. 31). The committee agrees that end-point-specific TEFs should be used in those situations in which one is interested in assessing the effects of a sample on a specific end point; however, for general monitoring or screening approaches (that is, for TCDD-related compounds in food and environmental samples) in which all
end points should be considered, TEF values that are based on all end points should be used.
Use of TEFs for DLC Body Burdens
Perhaps the issue of greatest concern in this section of the Reassessment is whether the current WHO TEFs, which were developed to assess the relative toxic potency of a mixture to which an animal is directly exposed by dietary intake, are appropriate for the assessment of internal TEQ concentrations and potential toxic effects. Application of the equation relating body burden, half-life, and bioavailability to congeners other than TCDD to give TCDD equivalents based on intake TEF values assumes that the TEF allows adequately for any difference between the congener and TCDD for half-life and bioavailability aspects. In addition, if exposure and estimated body burden of dioxins, other than TCDD, and DLCs are based on measured tissue concentrations, then converting the tissue concentration to a TEQ with TEFs derived from external doses might not be appropriate and might introduce significant uncertainties into the total TEQ estimate. In fact, previous studies have suggested that, because of toxicokinetic differences, the REP values for three PCDFs (2,3,7,8-tetrachlorodibenzofuran [TCDF], 1,2,3,7,8-PeCDF, and octachlorodibenzofuran [OCDF]) were greater when estimated from tissue concentration than when estimated from administered dose (DeVito et al. 1997). These data would support development of body burden TEF values in which the level of toxicity is directly related to body burden concentrations of a given DLC. Questions have also been raised about including octachlorodibenzo-p-dioxin (OCDD) and OCDF in the TEF scheme. Differences in the toxicokinetics of these compounds from other chemicals complicated early studies. OCDF and OCDD were originally assigned a TEF of zero because they failed to produce effects in early toxicity studies. However, both OCDF and OCDD are poorly absorbed in the gastrointestinal tract (Birnbaum and Couture 1988; DeVito et al. 1998) and significant TCDD-like effects of each were observed only after repeated doses were given over an extended time to allow accumulation in tissue (Couture et al. 1988; DeVito et al. 1997). Whether toxicokinetic differences of dioxins, other than TCDD, and DLCs exist that would similarly affect their REP and thus their TEFs need to be determined. However, these results raise concerns about the use of intake TEFs for body burden TEQ determinations and suggest that, if possible, it would be more appropriate to generate an additional set of TEFs for body burden tissue equivalents that could be used for DLC risk evaluation purposes. In addition, the use of intake TEFs for body burden TEQ determinations questions the overall conclusion that TCDD, other dioxins, and DLC body burden in humans is currently close to levels that reportedly produce adverse effects in
animals. Would it be higher or lower depending on the specific TEFs applied? A discussion of this point could not be found in the sections on toxic equivalents and should be included.
Additivity of DLCs
Assumption: Mixtures exhibit additive toxicities based on TEFs of individual chemicals. Additivity is a particularly critical assumption for the TEF approach. Considerable discussion of this issue is provided in the Assessment, Part II, Chapter 9, and from an overall perspective, this assumption appears valid, at least in the context of risk assessment. Additivity in biochemical and toxic responses by the indicated compound has been supported by numerous controlled mixture studies in vitro and in vivo and is scientifically justifiable. That support is not the case with other non-DLC PCDDs, PCDFs, and PCBs that are reported to be partial agonists or antagonists. The presence of partial agonists or antagonists in a complex mixture or in vivo would likely reduce the overall toxic potency (TEQ) of a mixture when tested in an animal when compared with the TEQ potency calculated simply from application of TEF values to individual compounds measured by instrumental analysis of the mixture. In fact, the ability of some non-DLC PCBs and PCDFs to inhibit TCDD-induced cytochrome 4501A1 protein (CYP1A1) activity and immunotoxicity in C57BL/6J mice has been reported (Bannister et al. 1987; Davis and Safe 1988; Biegel et al. 1989; Chen and Bunce 2004), as has the ability of a lower-affinity synthetic PCDF, such as 6-methyl-1,3,8-trichlorodibenzofuran (6-MCDF), to inhibit TCDD-induced CYP1A1, teratogenicity, immunotoxicity, and porphyria in rodent models in vivo (Astroff et al. 1988; Harris et al. 1989; Bannister et al. 1989; Yao and Safe 1989). These studies indicate that persistent non-DLCs can affect the magnitude of toxic and biological effects produced by a defined amount of TEQ calculated for a given complex mixture. However, given that the presence and concentration of these chemicals in a particular extract can vary dramatically and that very few published studies demonstrate significant alterations in the additive toxicities of dioxins, other than TCDD, and DLCs by other persistent non-DLC AHR ligands in vivo, the assumption of additivity of dioxins, other than TCDD, and DLCs should be considered a valid approach at the present time. Several published papers have demonstrated synergistic activation of AHR-dependent gene-expression effects that involve cross-talk between signaling pathways even at low concentrations. However, with respect to AHR-dependent toxic effects, current data are consistent with ligand and agonist additivity, which is a key assumption of the TEF/TEQ approach. However, EPA should acknowledge the possibility that the presence of non-DLC AHR antago-
nists in a complex mixture could affect the magnitude and overall toxic effects produced by the calculated amount of TEQs present in a mixture containing such compounds.
Assumption: REP of TCDD, other dioxins, and DLCs in rodent models is predictive of REP in humans, given that the rank-order potency is similar between species. Results from available in vivo, in vitro, and accidental and occupational exposure studies are generally consistent with this assumption. Numerous investigators have reported species-specific differences in AHR ligand binding affinity of TCDD, other dioxins, and DLCs. Depending on the system examined, the estimated affinity of binding of TCDD (and related compounds) to the human AHR is about 10-fold lower than that observed to the AHR from “responsive” rodent species and is comparable to that observed to the AHR from “nonresponsive” mouse strains (Roberts et al. 1990; Ema et al. 1994; Poland et al. 1994; Ramadoss and Perdew 2004). This reduced affinity appears to be at least in part due to a single amino acid substitution within the ligand binding domain of the human and “nonresponsive” mouse AHRs (Ema et al. 1994; Poland et al. 1994; Ramadoss and Perdew 2004). Although the affinity of binding of TCDD and related compounds to the human AHR is reduced compared with rodent AHRs, the qualitative and quantitative rank-order potency of these chemicals is similar. In addition to ligand binding, the REP of TCDD and related compounds to induce AHR-dependent gene expression in human cells is also reduced by up to 10-fold (Roberts et al. 1990; Harper et al. 1991; Xu et al. 2000; Zhang et al. 2003; Peters et al. 2004; Silkworth et al. 2005). Because TEFs are expressed relative to the toxicity of TCDD, the shift in TEF values of dioxins, other than TCDD, and DLCs appears to be similar between species. Several recent papers have reported that biological and toxicological responsiveness of humans to TCDD, other dioxins, and DLCs can vary up to 10-fold in vivo and in vitro and that these interindividual differences in responsiveness are not due to specific polymorphisms in AHR (Anttila et al. 2000; Harper et al. 2002; Cauchi et al. 2003). Not only do the documented species differences in AHR ligand binding and AHR responsiveness need to be addressed or taken into consideration with regard to rodent-to-human extrapolation, but the issue of interindividual variability among humans in their responses to TCDDs, other dioxins, and DLCs also needs to be considered when assessing human risk. The rank-order potency of other non-DLC AHR agonists is not necessarily similar between species, and if these chemicals are to be included in the TEF methodology in the future, species-specific TEFs would need to be developed.
Other Persistent AHR Agonists
Assumptions: Although other classes of persistent halogenated environmental chemicals that are structurally related to TCDDs, other dioxins, and DLCs have been identified, they are excluded because there are limited toxicological data and no validated TEFs for these chemicals. Another important source of uncertainty is the acknowledged likelihood that other persistent halogenated chemicals, such as brominated and mixed chloro and bromo coplanar chemicals, are present in environmental mixtures, the identities of which are just now emerging and for which TEFs have not yet been established (Reassessment, Part II, section 9.3.5; Part III, section 1.1). Many of these chemicals have been examined and observed to produce adverse AHR-dependent effects in vivo (Birnbaum et al. 1991, 2003). In fact, one mixed polychlorinated and polybrominated dibenzo-p-dioxin (2,3-dichloro-7,8-dibromo-dibenzo-p-dioxin) produced AHR-dependent toxicity in vivo (wasting and thymic involution) at concentrations up to 10 times lower than that of 2,3,7,8-TCDD (IPCS 1998a, p. 879, Table 50). Although significant information on the polybrominated dibenzo-p-dioxins and furans (PBDDs and PBDFs) is available and REP values for some of these compounds have been developed, there still are few toxicological and environmental distribution studies on these compounds. However, IPCS (1998a) suggested that development of TEFs for selected PBDDs and PBDFs is justified given their existing similarities in structure, mechanism, and potency to PCDDs and PBDFs. There are also many other classes of polyhalogenated chemicals that are known to bind to and activate AHR (polychlorinated naphthalenes, benzenes, azobenzenes, azoxybenzenes, and others), and some of these have also been shown to produce TCDD-like effects. However, the primary issue for the lack of consideration of these other TCDD-related compounds in the current assessment is that insufficient data are available on these chemicals, there are no currently determined or validated REPs and TEFs, and questions remain about the presence and persistence of these chemicals in the environment, food, and organisms. EPA should include these chemicals in the TEQ calculations when validated TEFs are developed.
Natural and Synthetic Non-DLCs AHR Agonists
Assumptions: Synthetic and natural non-DLC AHR agonists with a short biological half-life and lower AHR binding affinity do not interfere with PCDD-, PCDF-, and PCB-dependent TEQ predictions. It has been recognized for several years that human and animal diets contain relatively high concentrations of naturally occurring AHR agonists and antagonists (Denison et al. 2002; Denison and Nagy 2003; Jeuken et al. 2003) and that
there are non-dioxin-like halogenated aromatic hydrocarbons (HAHs) (PCBs and PCDFs) that are relatively potent AHR antagonists (described below). From a pharmacological and receptor binding kinetics point of view, if one assumes that the binding of these non-DLC agonists or antagonists to AHR is similar to that of TCDD (that is, binding is essentially irreversible) (Farrell et al. 1987; Bradfield and Poland 1988; Henry and Gasiewicz 1993; Brown et al. 1994; Petrulis and Bunce 2000), then the presence of relatively constant and high concentrations of relatively weak non-dioxin-like agonists or antagonists in blood and tissue (e.g., from chronic consumption of relatively high levels of these chemicals) could be expected to produce AHR-dependent effects or inhibit the overall toxic and biological effects produced by a defined amount of TEQ calculated from TCDD-related compounds present in a sample extract.
In most published studies, these metabolically labile non-DLC AHR agonists do not produce AHR-dependent toxicity; however, a few studies have reported the ability of some of these chemicals to produce TCDD-like toxic effects. b-Naphthoflavone (a polycyclic aromatic hydrocarbon [PAH] AHR agonist) was reported to produce thymic involution and splenomegaly in “AHR-responsive” C57 but not “AHR-nonresponsive” DBA mice (Silkworth et al. 1984) as well as wasting and brain developmental effects in fish (Grady et al. 1992; Dong et al. 2002). Developmental exposure of rats to indole-3-carbinol (I3C), a naturally occurring AHR ligand that can be converted in acidic conditions in the stomach into potent AHR agonists, including the high-affinity AHR agonist indolo-[3,2b]-carbazole (ICZ), was reported to produce some AHR-dependent reproductive effects similar to those of TCDD, although other distinct effects of ICZ were noted (Wilker et al. 1996). In addition, inhibition of cytochrome P450-dependent metabolism of PAHs was reported to result in dioxin-like effects in developing fish embryos exposed to PAHs that are AHR agonists (Wassenberg and Di Giulio 2004a,b). Not only would inhibition of CYP-dependent metabolism increase the persistence of the PAH in fish in vivo, but this scenario could also occur in the environment where organisms are exposed to complex chemical mixtures. In contrast to the above studies, the naturally occurring AHR ligand I3C failed to produce adverse effects in rats not only in a 1-year dietary chronic exposure study (Leibelt et al. 2003) but also in a high-dose, short-term study with subcutaneously administered ICZ for up to 10 days (Pohjanvirta et al. 2002).
The ability of metabolically labile phytochemicals to induce or inhibit induction of CYP1A1-dependent activities by TCDD in cell culture model systems has been reported by numerous laboratories (Williams et al. 2000; Amakura et al. 2002; Jeuken et al. 2003; Zhang et al. 2003). Moreover, while the naturally occurring AHR ligands I3C and diindolymethane have
been reported to inhibit TCDD-dependent induction of CYP1A1 in B6C3F1 mice in vivo (Chen et al. 1995, 1996), ICZ failed to interfere with the effects of TCDD in a high-dose 10-day study (Pohjanvirta et al. 2002). Lower-affinity synthetic non-dioxin-like AHR agonists, such as 6-MCDF, have been observed to inhibit TCDD-induced CYP1A1, teratogenicity, immunotoxicity, and porphyria in rodent models in vivo (Astroff et al. 1988; Bannister et al. 1989; Harris et al. 1989; Yao and Safe 1989). The ability of some non-dioxin-like PCBs and PCDFs to inhibit TCDD-induced CYP1A1 activity and immunotoxicity in C57BL/6J mice has also been reported (Bannister et al. 1987; Davis and Safe 1988; Biegel et al. 1989; Chen and Bunce 2004). In addition, administration of a synthetic flavonoid antagonist of the AHR (3'-methoxy-4' nitroflavone) to transgenic mice was observed to inhibit TCDD-inducible CYP1A1 and an AHR-responsive β-galactosidase transgene (Nazarenko et al. 2001).
In EPA’s Reassessment, a strong case is made for the distinctiveness of highly persistent AHR agonists, versus readily metabolized ones, in terms of toxicological responses and risk assessment. However, the limitation with regard to the lack of knowledge of the effects of the large number of naturally occurring and synthetic AHR ligands on the overall toxic potency of TCDD-related compounds was acknowledged in the Reassessment (Part III, p. 9-40, lines 27 to 28). Although few studies have examined the effects of non-DLC AHR agonists or antagonists on the overall toxic and biological potency of TCDD-related compounds, a few in vivo studies do provide supporting evidence that metabolically labile AHR agonists or antagonists can actually reduce the overall toxic potency of TCDD and presumably other dioxins and DLCs. On the other hand, an excellent correlation between the predicted TEQ and the magnitude of the observed response was observed in several studies examining the effects of real-world samples (soot, incinerator fly ash, sediment leachate, and fish or fish extracts) in animals exposed to these samples in vivo (DeCaprio et al. 1986; Silkworth et al. 1989; Suter-Hofmann and Schlatter 1989; Tillitt and Wright 1997; Powell et al. 1997). While the occurrence of AHR-dependent antagonism by phytochemicals and other AHR antagonists in humans has yet to be confirmed, given species similarities in the AHR and AHR signaling pathway and the relatively high concentrations of many naturally occurring dietary AHR antagonists, the possibility remains that interactions or interferences between natural AHR agonists and TCDD-related compounds might occur. Non-DLC AHR agonists could affect the TCDD-related compounds dose-response relationships for short biological responses (that is, gene induction) and contribute to an additive response for the end points. However, the metabolic lability (that is, lack of persistence) of these compounds prevent them from affecting longer-term dose-response relationships (including threshold and nonlinear assump-
tions) for toxic end points, such as cancer. That is one reason for the Reassessment to focus only on TCDD, other dioxins, and DLCs that are documented to produce AHR-dependent toxicity. Although these interactions would not affect individual TEF values or the calculation of an overall TEQ determined in controlled laboratory experiments, they could affect the magnitude and overall toxic effects produced by a defined amount of total TEQs calculated from intake or present in the body. Accordingly, EPA should acknowledge in the Reassessment the potential for non-DLCs to affect the overall biological and toxic potency of a defined amount of TEQs present in a complex mixture of chemicals and propose considering these compounds in the overall calculations when and if sufficient and appropriate in vivo data become available in the published literature to support their modulatory effect on DLC- and AHR-dependent toxicity.
KEY STUDIES AND PUBLICATIONS TO BE INCLUDED
Several relatively recent studies not included in the Reassessment support using the TEF/TEQ approach for noncancer and cancer end points; their inclusion would greatly strengthen the Reassessment.
Studies in rats with TCDD or heptachlorodibenzo-p-dioxin (HpCDD) revealed that the REP derived from acute toxicity studies were the same as that obtained in a subchronic and chronic toxicity study; both had a TEF of ~0.007 for HpCDD, although no confidence bounds were provided (WHO TEF = 0.01) (Viluksela et al. 1997a).
A mixture of four PCDDs or individual PCDDs at equipotent doses (based on TEFs) to rats produced comparable biochemical changes after single as well as multiple doses. The authors concluded that TEFs from acute toxicity studies can accurately predict the toxicity of dioxins, other than TCDD, and DLC mixtures regardless of whether they are administered as single compounds or as a mixture, the results supporting additive toxicity for those compounds (Stahl et al. 1992; Viluksela et al. 1998a,b).
Rats given a mixture of two PCDDs, four PCDFs, and two PCBs (in a ratio found in foodstuffs) at a concentration of 2.0 µg TEQ/kg of body weight produced adverse reproductive and developmental effects comparable to those at a TCDD concentration of 1 µg/kg (Hamm et al. 2003). The authors concluded that the TEQ approach was a reasonable predictor of the reproductive effects studied.
Application of TEFs adequately predicted the increased incidence of liver tumors in rats (hepatocellular carcinoma and cholangiocarcinoma) induced by exposure to a mixture of TCDD, 3,3',4,4',5-PCB, and 2,3,4,7,8-
PeCDF compared with an equivalent concentration of TCDD (Walker et al. 2005).
CONCLUSION AND RECOMMENDATIONS
Overall, even given the inherent uncertainties and limitations, the TEF method, when applied correctly, is a reasonable, scientifically justifiable, and widely accepted method to estimate the relative toxic potency of dioxins, other than TCDD, and DLCs on human and animal health.
Specific Conclusions and Recommendations
AHR-independent mechanisms excluded. AHR-independent effects of TCDD have been reported, and although their significance and role in the overall toxic effects remain to be established, the Reassessment should acknowledge the existence of these AHR-independent effects because future studies may demonstrate that they play some role in the overall toxic and biological effects of TCDD, other dioxins, and DLCs.
Uncertainty of TEF values. A significant degree of uncertainty exists in the current consensus TEFs, and the quantitative weighting considerations that have gone into their establishment are not clear. While the Reassessment should acknowledge the need for better uncertainty analysis of the TEF values, extensive and appropriate uncertainty analysis would take considerable time and effort. Accordingly, the committee endorses the recommendation of some members of the 2000 EPA SAB Panel “that, as a follow up to the Reassessment, the EPA should establish a task force to build ‘consensus probability density functions’ for the thirty chemicals for which TEFs have been established, or to examine related approaches such as those based on fuzzy logic” (EPA SAB 2001, p. 29).
Consistency of REP values. Most in vitro and in vivo studies support the assumption that the indicated dioxins, other than TCDD, and DLCs are not only full agonists but that their REP is similar for all end points of concern and exposure scenarios. However, significant end-point-specific differences in the REP of some dioxins, other than TCDD, and DLCs have been reported and whether other differences exist remains to be determined. Consistent with the recommendations of the 2000 EPA SAB Panel, this committee also suggests that it would be appropriate for the Reassessment to note that end-point-specific TEFs/TEQs might be derived as data become available and that those specific values be used when that end point is being considered. It should also be made clear that general monitoring or screening approaches (that is, for TCDD-related compounds in food and
environmental samples) should use TEF values that are based on REPs values of all end points.
Use of TEFs for DLC body burdens. This is perhaps the greatest issue of concern in this section of the Reassessment because it remains to be determined whether the current WHO TEFs, which were developed to assess the relative toxic potency of a mixture to which an animal is directly exposed by dietary intake, are appropriate for the assessment of internal TEQ concentrations and potential toxic effects. The issue was not well described or well justified in the Reassessment and might be incorrect. It is further complicated by an EPA paper (DeVito et al. 1997) suggesting that use of TEFs for DLC body burdens might not be appropriate for some PCDFs. The issue would be further complicated if toxicokinetic differences of other DLCs similarly affect their REP. Overall, it remains to be determined whether intake TEFs are appropriate for body burden TEQ determinations. If body burdens are going to be used as the dose metric, the committee recommends that a separate set of body burden TEFs be developed and applied for this evaluation or that the appropriateness of intake TEFs for body burden TEQs be scientifically justified. Without these corrected values, the overall TEQs estimated by use of intake TEFs could be inaccurate.
Role of AHR and additivity of DLCs. These aspects are well described and well supported by extensive numbers of scientific studies. However, EPA should acknowledge the possibility that AHR antagonists present in a complex mixture could affect the magnitude and overall toxic effects produced by a calculated amount of total TEQs present in a given sample even if they do not affect the TEQ calculations. This issue was not addressed in the Reassessment.
Rodent-to-human prediction. Although the REP of dioxins, other than TCDD, and DLCs in rodent models is predictive of REP in humans from a qualitative rank-order potency point of view, some species-specific differences in AHR ligand binding affinity of TCDD, other dioxins, and DLCs have been observed. However, because TEF values are expressed relative to that of TCDD in the individual species, the TEF values for dioxins, other than TCDD, and DLCs appear to be similar between species. If significant differences in the REP of dioxins, other than TCDD, and DLCs are found between humans and other species, then adjustments should be made in the TEFs, and these should be acknowledged in the Reassessment.
Other AHR agonists.
Related HAH DLCs. Lack of consideration of other persistent halogenated chemicals, such as brominated, chlorinated, and mixed chloro and bromo coplanar chemicals, which clearly exert their toxic and biological effects in an AHR-dependent manner could result in underestimation of
the overall TEQ for a given sample. Although REP values and TEFs have been developed for some of these chemicals, few studies have been carried out with most of them, and their relative toxic potency is unknown. Given the structural similarities and mechanism of action of these chemicals in vivo and in vitro with the established compounds, as validated REP values become available, TEFs should be assigned, and these chemicals should be included in the TEF/TEQ approach. This course of action should be noted in the Reassessment.
Synthetic and naturally occurring non-DLC AHR ligands. A large number of synthetic and naturally occurring non-DLC AHR ligands have been identified and are present in human diets and presumably in blood and tissues. The assumption that non-DLC AHR agonists with a short biological half-life do not interfere with DLC-dependent TEQ predictions for mixtures is controversial and remains to be confirmed. Although receptor binding kinetic evaluations suggest that these chemicals could interfere with TCDD, other dioxins, and DLCs if at high concentrations in blood and tissue, few of these metabolically labile non-DLC AHR agonists have been observed to directly produce AHR-dependent toxicity. The Reassessment makes a strong case for the ability of only highly persistent AHR agonists to produce toxicity, but the lack of knowledge of the effects exerted by the large number of naturally occurring dietary and synthetic AHR ligands on the overall toxic potency of TCDD, other dioxins, and DLCs still leaves the question open, particularly with regard to humans. Although these AHR ligands would not affect TEQ calculations, they could affect the magnitude of the toxic and biological effect of a defined amount of TEQ. This point should at least be made clear in the Reassessment, and when a sufficient number of published studies demonstrate the ability of non-DLC AHR agonists or antagonists to modulate the overall effects of DLCs, then EPA should consider how these chemicals would affect the current TEF/ TEQ approach for potency estimates.
WHO’s plan to reexamine DLC TEFs in 2006. The major issues of concern described above for the TEF approach will also be the focus of a meeting of the International Programme on Chemical Safety (announcement in IPCS 2004). The issues include (1) considering methods and approaches for deriving TEFs, including quantitative (statistical) methods, such as establishing an uncertainty range of available REP data and application of a specified cut-off value to derive TEF values, application of weighting factors to existing data, and related issues; (2) determining whether to continue to include mono-ortho PCBs in the present TEF concept; (3) considering whether other compounds should be considered for inclusion in the TEF concept, taking into account the prerequisites for inclusion outlined by Van den Berg et al. (1998); and (4) determining the applicability of the use of TEFs to estimate intake versus internal concentra-
tions and to what extent could or should internal WHO TEF factors be established in the future? EPA should consider the outcome of the IPCS TEF update meeting and incorporate the issues and changes into the Reassessment.
Updating the Reassessment. Although the Reassessment clearly states that the WHO TEFs of 1998 will be used for assessment and calculation, if or when TEF values are changed or new chemical TEFs are added by the current or future WHO TEF panels (such as the 2005 panel), EPA should consider incorporating the new TEF values and methods for TEQ determination.