Tumour promotion by complex mixtures of polyhalogenated aromatic hydrocarbons (PHAHs) and the applicability of the toxic equivalency factor (TEF) concept

<p>The aim of the project described in this thesis consisted of two main objectives, first, to examine the tumour promotion potential of complex, environmentally relevant mixtures of polychlorinated biphenyls (PCBs), polychlorinated dibenzo- <em>p</em> -dioxins (PCDDs) and polychlorinated dibenzo- <em>p</em> -furans (PCDFs) and secondly, to evaluate the applicability of the Toxic Equivalency Factor (TEF) concept for the tumour promotion potential of complex mixtures of PCBs, PCDDs and PCDFs. In addition, the effect of sub-chronic exposure to these complex mixtures was determined on endocrine parameters, i.e. the vitamin A and thyroid hormone status, which play an essential role in normal tissue growth and fetal development and are possibly involved in the process of carcinogenesis.</p><p>Carcinogenicity is one of the toxic endpoints in risk assessment of PCBs, PCDDs and PCDFs (WHO, 1992). PCBs, PCDDs and PCDFs are considered as tumour promoters rather than as initiators of carcinogenicity (Safe, 1989; Silberhorn <em>et al.</em> , 1990; Whysner and Williams, 1996). So far, most studies on tumour promotion by PCBs have investigated the potency of single, mostly planar dioxin-like congeners, based on the presumption that the Ah-receptor pathway is also involved in mediating the tumour promoting effects of PHAHs (Safe, 1989; Silberhorn <em>et al.</em> , 1990). There is much less information available on the tumour promoting effects of complex mixtures of PHAHs after sub-chronic exposure.</p><p>The approach in this thesis was to focus on mixtures of polyhalogenated hydrocarbons (PHAHs), both dioxin-like and non-dioxin-like, relevant for the human intake. To reveal underlying mechanisms of possible interactions between PHAH congeners and to determine the toxic potential of the PHAH mixtures, the ethoxyresorufin- <em>O</em> -deethylase (EROD) and the AhR-dependent luciferase reporter gene (DR-CALUX) bio-assays were performed. Both assays are indicators for an Ah receptor mediated, dioxin-like toxicity. The tumour promotion potential of complex PHAH mixtures <em>in vivo</em> , was studied in female Sprague Dawley rats, using the development of altered hepatic foci (AHF) as a parameter in a two-stage initiation/promotion bio-assay introduced by Pitot <em>et al.</em> (1978).</p><h3>Chapter 2</h3><p>In chapter 2, the results of <em>in vitro</em> experiments are described. Interactions between individual mono- or di- <em>ortho</em> PCB congeners and 2,3,7,8-TCDD were studied in the EROD and the DR-CALUX bio-assay, using mouse and rat hepatoma cell lines. In addition, the dioxin-like potential of the PHAH mixtures, designed for the animal experiments, and possible interactions between congeners within the mixtures, was determined in the CALUX assay. Preliminary data are presented on the inhibition of the gap junctional intercellular communication (GJIC), which is seen as an <em>in vitro</em> parameter for tumour promotion, by the PHAH mixtures.</p><p>When individually dosed, the mono- <em>ortho</em> PCBs induced both the EROD and CALUX activity but to a lower maximum and at higher concentrations as compared to TCDD. Co-administration of mono- <em>ortho</em> PCBs and TCDD, decreased the TCDD induced EROD and CALUX activity dose-dependently, with increasing concentrations of the partially antagonistic mono- <em>ortho</em> PCBs. The residual level of the EROD and CALUX induction in case of co-administration, was equal to the maximum inducible activity level of the individual mono- <em>ortho</em> PCB congener. None of the tested di- <em>ortho</em> PCBs was capable of inducing the EROD or CALUX activity. However, all di- <em>ortho</em> PCBs antagonised the TCDD-induced EROD and CALUX activity in a dose-dependent manner and with different potencies. A couple of combined exposures were tested for the inhibition of the GJIC. The results indicated that similar non-additive interactions, as observed in the EROD and CALUX assay, were seen here.</p><p>The PHAH mixtures designed for the first animal experiment ( <em>Chapter 3,4</em> ), induced the CALUX activity up to the maximum activity level as induced by TCDD. These PHAH mixtures were also potent inhibitors of the GJIC. No interactions between individual congeners in the PHAH mixture could be observed, neither in the CALUX assay or on the inhibiton of the GJIC. Interactive effects were shown in the CALUX assay between the PCB fractions designed for the second animal experiment ( <em>Chapter 5</em> ). The 0- <em>ortho</em> and the 1- <em>ortho</em> substituted PCB fraction induced the CALUX activity up to 40% and 9% of the maximum level induced by TCDD respectively, while the 2-4 <em>ortho</em> fraction did not show any induction of the CALUX activity. Co-administration of the fractions inhibited the CALUX activity down to 3% of the maximum level induced by TCDD. The GJIC was only slightly inhibited by the 2-4 <em>ortho</em> and the reconstituted 0-4 <em>ortho</em> fractions.</p><h3>Chapter 3 & 4</h3><p>In the first animal experiment ( <em>Chapter 3,4</em> ), the development of AHF by a complex synthetic mixture of dioxin-like compounds was studied. The composition of this mixture was based on the presence of and relative ratio's between the six most relevant PHAHs in Baltic herring and covered over 90% of the TEQs present. To study possible interactive effects, PCB 153 (2,2',4,4',5,5'-HxCB) was added to the mixture as a representative of the non-dioxin-like, di- <em>ortho</em> substituted PCBs.</p><p>In chapter 3, the toxicokinetic properties of the PHAH congeners are presented. Gas-chromatography and mass-spectrometry (GC-MS) analysis of PHAH concentrations in the liver showed considerable differences in hepatic retention (as percentage of the given dose) between congeners, thereby changing the relative ratios of congeners between external and target dose in favor of the planar compounds. Further, it was shown that addition of PCB 153 to the PHAH mixture increased the hepatic retention of all dioxin-like PHAH congeners in the mixture. This observation is explained by the capacity of PCB 153 to induce Ah receptor levels in the liver, and consequently increase the hepatic level of CYP1A2, which is known to possess a high binding affinity for planar PHAHs.</p><p>In chapter 4, the AHF data are shown. The promotion of AHF was significantly increased after exposure to the PHAH mixtures, but to a lower extent than expected on the basis of the TEQs calculated from the TEF values as proposed by the WHO (Ahlborg <em>et al.,</em> 1994). A difference between the WHO TEF values (Ahlborg <em>et al.,</em> 1994) used for the calculation of the TEQ of the PHAH mixture, and the relative potency (REP) values of the individual congeners that are actually based on AHF data, may partly explain the observed differences in AHF induction between the rats exposed to the equipotent doses of TCDD and the PHAH mixtures. In addition, differences in toxicokinetic properties of the congeners and interactive effects on deposition of the congeners ( <em>Chapter 3</em> ) may have influenced the predicted toxic potency of the PHAH mixture as well. No interactive effects of PCB 153 on the AHF development or EROD induction could be observed.</p><p>It is concluded that the TEF approach predicted the tumour promotion potency of the investigated PHAH mixtures quite well, within a factor of two. An interactive effect between PCB 153 and the planar PHAHs occurred at the kinetic level.</p><h3>Chapter 5</h3><p>Chapter 5 describes the second animal experiment, in which the contribution of non-dioxin-like as well as dioxin-like PCB congeners to the total induction of AHF by a complex PCB mixture was studied. For this purpose the commercial PCB mixture Aroclor 1260 was fractionated into a 0-1 <em>ortho</em> and a 2-4 <em>ortho</em> PCB fraction, which were tested separately and as a reconstituted 0-4 <em>ortho</em> PCB mixture.</p><p>GC-MS analysis of the Aroclor 1260 fractions confirmed that there were no planar, dioxin-like compounds present in the 2-4 <em>ortho</em> PCB fraction. In addition, the 2-4 <em>ortho</em> PCB fraction did not show luciferase induction in the <em>in vitro</em> DR-CALUX bio-assay, indicating that this fraction had no dioxin-like potential ( <em>Chapter 2</em> ). A remarkable finding in the rat study was that the 2-4 <em>ortho</em> PCB fraction explained approximately 80% of the total observed effect on the development of AHF by the 0-4 <em>ortho</em> PCBs present in Aroclor 1260. In contrast to what is generally accepted, the dioxin-like PCB congeners did not significantly contribute to the effect on AHF development. No interactive effect on AHF development or the toxicokinetics was observed for the 0-1 and the 2-4 <em>ortho</em> PCB fraction. PCB 153, incorporated as additional treatment, showed a similar potential to induce AHF development as the 2-4 <em>ortho</em> PCB fraction in Aroclor 1260.</p><p>It was concluded that the TEF concept largely underestimates the tumour promotion effect of complex PCB mixtures, since the tumour promotion potential of the non-dioxin-like PCBs is not taken into account.</p><h3>Chapter 6</h3><p>In chapter 6, the results are shown on the vitamin A and the thyroid hormone status of the rats of the first and second tumour promotion experiment ( <em>Chapter 4,5</em> ).</p><p>From the first experiment it appeared that hepatic retinyl palmitate is a rather sensitive parameter for exposure to dioxin-like PHAHs, as the retinyl palmitate levels were severely decreased after treatment with the PHAH mixtures and to a similar extent as compared to TCDD treatment. However, an opposite effect was observed on the plasma retinol concentration after treatment with the PHAH mixture and TCDD, respectively. In addition, the PHAH mixture caused a relatively strong decrease of the thyroid hormone levels in plasma and decreased the ratio of total thyroxin and free thyroxine as compared to TCDD. The most likely reason for these observations is the formation of hydroxy-metabolites of PCB 118, present in the PHAH mixture, which are known to disrupt the transport-protein complex (RBP-TTR) of retinol and thyroxine and thereby drastically reducing plasma levels of both vitamin A and thyroxine. This situation does not occur in the case of TCDD exposure, which effects the vitamin A and thyroid hormone status mainly via interference with liver metabolism.</p><p>In the second experiment, the retinoid and thyroid hormone levels were not affected significantly. This indicates that in case of exposure to PCBs at environmental levels, no or at best only marginal effects can be expected on the retinoid and thyroid hormone status.</p><p>It was concluded on the basis of these observations that the effects on plasma retinol and thyroxine by complex mixtures of PHAHs are not well predicted by the TEF concept, due to involvement of several different mechanisms and mechanistic interactions depending on the composition of the PHAH mixture.</p><h3>Concluding remarks</h3><h4>Applicability of the TEF concept</h4><p>The most striking finding of this thesis work is that the non-dioxin-like PCB fraction in the commercial mixture Aroclor 1260 explained over 80% of the observed effect on AHF development ( <em>Chapter 5</em> ). On the basis of these results it was concluded that the TEF approach was inadequate in its prediction for the tumour promotion potential of a complex PCB mixture as used in the second animal experiment. However, the tumour promotion potential of the complex dioxin-like PHAH mixture used in the first animal experiment ( <em>Chapter 4</em> ) was quite well predicted by the TEF approach, e.g. within a factor of two. The observed kinetic interaction between the congeners ( <em>Chapter 3</em> ), had apparently no significant consequence for the tumour promotion potential of the PHAH mixture.</p><p>Further it was apparent that the TEF concept failed to predict the effect of the dioxin-like PHAH mixture on plasma retinol and underestimated the effect on the thyroid hormone concentrations ( <em>Chapter 6</em> ). The lack of predictability by the TEF approach for these endocrine effects is possibly due to additional toxicity of hydroxylated PCBs, formed of PCB congeners present in the dioxin-like PHAH mixture.</p><h4>Interactive effects</h4><p>The <em>in vitro</em> experiments ( <em>Chapter 2</em> ) indicated the possibility of interactive effects between dioxin-like PHAHs and mono- and di- <em>ortho</em> PCBs, at the level of Ah receptor binding. However, competition between compounds is only likely to occur under conditions of Ah receptor saturation and a large concentration difference between the dioxin-like PHAH and the mono- and/or di- <em>ortho</em> PCBs. These conditions can be easily reached <em>in vitro</em> but will be seldomly observed <em>in vivo</em> . In the animal experiments the major interaction was observed at the kinetic level, namely, PCB 153 enhanced the hepatic deposition of the dioxin-like PHAHs ( <em>Chapter 3</em> ), most likely by induction of CYP1A2 which is known to have a high binding affinity for dioxin-like PHAHs. No interactive effects, at any level, were observed between the dioxin-like and the non-dioxin-like PCB fraction of Aroclor 1260. This may be explained by the low hepatic retention of the congeners, possibly due to the low TEQ level of the dioxin-like fraction in Aroclor 1260, i.e. no or <strong>a marginal induction of hepatic binding proteins</strong> . At such a low level of hepatic retention interactive effects will not be seen. In terms of TEQs the dioxin-like fraction was certainly more close to environmental levels of exposure as occurs in wildlife and human, leading to the conclusion that kinetic interactions do not play a role at environmental exposure levels.</p><h4>Implications for risk assessment</h4><p>A remaining question is if, on the basis of these results, it can be concluded whether the current approach for risk estimation of complex mixtures of PCDDs, PCDFs and PCBs is appropriate or not. On the basis of a chronic carcinogenicity study performed by Kociba <em>et al.</em> (1978) for TCDD a no-observed-adverse-effect level (NOAEL) of 1 ng/kg/day was derived. A NOAEL for the synthetic dioxin-like PHAH mixture might be close to the NOAEL of TCDD ( <em>Chapter 3,4</em> ), if it is assumed that the dose-response curves for the carcinogenic potential of the PHAH mixtures and TCDD have a similar shape. This indicates that the TEF approach sufficiently predicts the potential risk of exposure to dioxin-like PHAHs.</p><p>For the non-dioxin-like PCBs it is more complicated, since the TEF concept is not applicable for the risk estimation of non-dioxin-like PCBs nor is there another tool available for this purpose. In addition, the total PCB intake is not well known since no congener specific analysis of the occurrence of PCBs in foodstuff is available as well as any information about differences in congener patterns between food items. An extensive survey was done in the Netherlands by Liem and Theelen (1997), who reported an intake of 20 ng/kg/day (sum of 29 PCBs) in 1994, based on a Dutch diet. Given the assumption that a level of 20 ng/kg/day in foodstuff covers approximately 20-30% of the total dietary intake, a daily intake of PCBs of 60-100 ng/kg/day can be estimated of which >90% consists of 2-4 <em>ortho</em> PCBs. In the tumour promotion experiment, a NOAEL for the 2-4 <em>ortho</em> PCBs in Aroclor 1260 was not achieved; the lowest experimental dose of 1 mg/kg bw/week (~140 µg/kg bw/day, <em>Chapter 5</em> ) still enhanced the development of AHF two-fold. Using a factor of 5 for extrapolation from the lowest-observed-adverse-effect-level (LOAEL) to NOAEL, a NOAEL of approximately 30 µg/kg/day can be deduced. For the calculation of a Tolerable Daily Intake (TDI) for dioxin-like compounds, the WHO uses a safety margin of 100. When the same margin is used for the non-dioxin-like compounds a TDI of 300 ng/kg/day can be calculated, which is a factor 3-5 above the presumable daily intake. Although there are many uncertainties in this calculation, there is probably no reason for immediate concern as large safety margins were applied. However, the results of this thesis work demonstrate the necessity for risk assessment to look at both the dioxin-like and non-dioxin-like PCBs.</p><h3>Overall conclusions</h3><p>Overall, the most important conclusion which can be drawn from this thesis work is that the majority of the effect on tumour promotion by PCBs is caused by a non-dioxin-like mechanism of action. Therefore the TEF approach, although useful to predict effects of dioxins and similar compounds, does not predict the tumour promotion potential of complex mixtures of PCBs as being present in the environment. This may have important implications for the risk assessment of complex mixtures of PHAHs as occur in e.g. foodstuff.</p&gt