Thyroid hormone binding proteins as novel targets for hydroxylated polyhalogenated aromatic hydrocarbons (PHAHs): possible implications for toxicity.

Some toxic effects caused by polyhalogenated aromatic hydrocarbons (PHAHs) develop through alterations in the reproductive and thyroid hormone regulatory systems, thereby affecting (brain) development, reproduction and behaviour of several species (Stone, 1995, Birnbaum, 1994, for review: Brouwer et al. , 1995, Peterson et al. , 1993). In this thesis we have focused on the effects of different classes of PHAHs, eg. polychlorinated biphenyls (PCBs), dibenzofurans (PCDFs) and dibenzo- p -dioxins (PCDDs) and their hydroxylated metabolites on thyroid hormone homeostasis. These changes seem to be partly caused by Ah-receptor mediated changes in thyroid hormone glucuronidation, and effects on the thyroid gland affecting hormone production and secretion. However, hydroxylated metabolites of PCBs, PCDFs and PCDDs may have an additional effect on thyroid hormone transport. Previous studies (Brouwer, 1987) have shown that exposure to 3,3',4,4'- tetrachlorobiphenyl (TCB) can disturb the plasma transport of thyroxine (T ₄ ) and retinol in rats through specific competition of a hydroxylated metabolite, 4-OH-3,3',4',5-tetraCB, with T ₄ for the thyroid hormone binding site of transthyretin (TTR), the major thyroid hormone transport protein in rodents. This observation raised the question if structurally related hydroxylated PHAH- metabolites could interact with TTR in the same way, as well as with other thyroxine binding proteins, like thyroxine binding globulin (TBG) and type-1-deiodinase (ID-1), subsequently disturbing thyroid hormone transport and metabolism. Special attention was also paid on the structure-activity relationships of hydroxylated PHAH metabolites for binding to TTR by using in vitro and in vivo studies and X-ray crystallographic structure analysis, In vitro studies on interactions of hydroxylated PHAH metabolites with thyroxine binding proteins In in vitro studies the interactions of several hydroxylated PHAH metabolites with 3 different T ₄ binding proteins, eg. TTR, thyroxine binding globulin (TBG) and type-1-deiodinase (ID-1), were investigated (Chapter 2, 3 and 4). The inhibition of T ₄ binding to TTR by hydroxylated PHAH metabolites was studied using in vitro T ₄ -TTR bindingstudies. These studies revealed the structural requirements for competition of T ₄ binding to TTR by hydroxylated PHAH metabolites: para- or meta-hydroxylation on one or both phenylrings, with one or more adjacent chlorine substitutions (Chapter 2). PHAH metabolites with these structural characteristics showed a remarkable resemblance to T ₄ the natural ligand for TTR, consequently displacing T ₄ from the T ₄ binding site of TTR. Both non-planar, ortho-chlorinated hydroxylated PCB metabolites and rigid, planar hydroxylated PCDF or PCDD metabolites could inhibit T ₄ -TTR binding. However, the ortho-hydroxylated PHAH metabolites and parent PHAH compounds, like TCDD, 2,3,3',4,4'-pentaCB and 3,3',4,4'-tetraCB could not inhibit T ₄ -TTR binding in vitro. In subsequent in vitro studies, a wide range of hydroxylated PHAH metabolites did not inhibit T ₄ binding to TBG, the major plasma thyroid hormone transport protein in man (Chapter 3). This indicates that ligand interactions with TTR or TBG are clearly different. Additional studies with iodothyronine derivatives, showed that tri-iodophenol and to a lesser extent di-iodotyrosine could inhibit T ₄ -TTR binding but not T ₄ -TBG binding in vitro. Finally, the enzymatic activity of hepatic ID-1, which plays a role in the (in)activation of thyroid hormones, could be competitively inhibited mainly by di-para-hydroxylated, meta-halogenated PHAH metabolites while mono-hydroxylated PHAH metabolites were 10 to 100 times less potent (Chapter 4). The differences between the structural requirements of hydroxylated PHAH metabolites for interactions with TTR, TBG and ID-1, are in line with previous studies in which related hydroxylated PHAH compounds or iodothyronine derivatives were used. In conclusion, specific hydroxylated PHAH metabolites can disturb T4-TTR interactions, or inhibit ID-1 activity in vitro, indicating that hydroxylated PHAH metabolites may play an additional role in the observed disturbances in thyroid hormone transport and metabolism after PHAH exposure in vivo . In vivo studies on effects of Aroclor 1254 and TCDD on thyroid hormone transport and metabolism Two in vivo experiments were carried out to study role that both disturbances in plasma T ₄ transport and hepatic T ₄ metabolism caused by hydroxylated PHAH metabolites play in the observed decreases in plasma T ₄ levels. Rats were exposed to Aroclor 1254, a commercial mixture containing persistent and metabolisable PCB congeners (Chapter 5) or the persistent 2,3,7,8- tetrachlorodibenzo-p-dioxin (TCDD) (Chapter 6). In adult Wistar rats, that were exposed to a high dose of Aroclor 1254, plasma T ₄ levels were decreased on day 3 and day 8 (Chapter 5). In addition, high levels of a single hydroxylated PCB metabolite, eg. 4-OH-2,3,3',4',5-pentaCB, were detected in plasma of the rats on day 8, while lower concentrations of this metabolite were present in the blood on day 3. However, the T ₄ binding capacity in plasma was decreased only in the high dosed group on day 8 but not on day 3, indicating a threshold level for the hydroxylated PCB metabolite to disturb T ₄ -TTR binding. On both day 3 and 8, hepatic cytochrome P450 1A1 levels and activity were induced, which is essential for the formation of hydroxylated metabolites. Hepatic T ₄ glucuronidation was induced simultaneously. The decreased plasma T ₄ levels found in all exposure groups could therefore be attributed to both disturbed plasma T ₄ transport and/or induced T ₄ glucuronidation. No significant changes in plasma T ₃ levels were found following Aroclor 1254 treatment. In addition, hepatic ID-1 activity was not decreased, suggesting that the in vitro inhibition by hydroxylated PCB metabolites does not occur in vivo. However in an earlier study Adams et al. (1990), in vivo exposure to the easily metabolisable 3,3',4,4'-tetraCB or persistent TCDD could inhibit ID-1 activity. The levels of PCB metabolites with the required structure for inhibition of hepatic ID-1 activity were possibly too low (Chapter 4) liver after Aroclor 1254 exposure in vivo. Another remarkable finding was the selective retention of a single specific hydroxylated PCB metabolite, 4-OH-2,3,3',4',5-pentaCB, in plasma of rats exposed to a complex mixture of PCB congeners (Fig. 1). This was the result of the presence of PCB-congeners that strongly induce cytochrome P4501AI activity and PCB-congeners that could easily form hydroxylated PCB metabolites in the Aroclor 1254 mixture, and the strict selectivity of the TTR present in plasma retaining only hydroxylated PCB metabolites that meet the structural requirements as described in Chapter 2. Surprisingly the 4-OH- 2,3,3',4',5-pentaCB metabolite which was formed and selectively retained in plasma, has a hydroxy group on the highest chlorinated ring, in contrast to the expected formation of mainly metabolites with a para - or meta -hydroxy group on the least chlorinated ring. However, the inhibition potencies of T ₄ -TTR binding for the 4-OH-2,3,3',4',5-pentaCB metabolite (Chapter 5) and the structurally related 4'-OH-2,3,3',4,5'-pentaCB metabolite (Chapter 2) were almost similar. There is no clear indication yet on the mechanism of selective retention of the 4-OH-2,3,3',4',5-pentaCB in rat plasma, although pharmacokinetics and tissue levels of the presumed parent compounds 2,3,3',4,4'- pentaCB (CB 105) or 2,3',4,4',5-pentaCB (CB 118) may play a role. No detectable levels of hydroxylated metabolites were found by GC-MS analysis of plasma extracts of rats at both day 3 and 8 following exposure to the persistent TCDD, although cytochrome P4501A1 levels and activity were markedly induced (Chapter 6). In addition no unequivocal decrease in T ₄ -TTR binding in plasma of TCDD-exposed rats was observed. However, hepatic thyroid hormone metabolism was clearly altered: T ₄ glucuronidation and brain type-2-deiodinase (ID-2) activity were increased and hepatic ID-1 activity was decreased, which may explain the observed plasma T4 reductions in the TCDD exposed rats. These changes in thyroid hormone homeostasis suggest a hypothyroxinemic state of the TCDD-exposed rats, although no decreased plasma T4 levels were found. The decrease in ID-1 activity after TCDD exposure was not likely to be caused by hydroxylated metabolites, as was described in Chapter 4, but may be caused by direct effects of TCDD on ID-1 activity or by the assumed hypothyroid state of the TCDD exposed rats. While Aroclor 1254 exposure disturbed T4 plasma transport and increase T ₄ glucuronidation (Chapter 5), TCDD exposure only enhanced the hepatic elimination of T4 (Chapter 6), leading to decreased plasma T ₄ levels. Although the in vivo studies described in this thesis indicate two different mechanisms for decreases in plasma T ₄ levels after PHAH exposure, we can not exclude a third possible mechanism since several studies described changes on thyroid gland histology and thyroid hormone secretion after exposure to Aroclor 1254, TCDD and related compounds. Structural basis for interactions of hydroxy-PCB metabolites with TTR The selective retention of a specific hydroxylated PCB metabolite in vivo (Chapter 5), inspired us to look into the interactions of hydroxylated PCB metabolites with TTR in more detail, and to try to find a structural basis for the structural requirements for TTR binding as described in Chapter 2. X-ray crystallographic analysis of a complex of TTR with a hydroxylated PCB metabolite, eg. 4,4'- (OH) ₂ -3,3',5,5'-tetraCB, refined to a 2.7 A resolution, revealed a hydrogen bond formation between a para-hydroxy group of the metabolite with the paired Serine 117 amino acid residues, present in the centre of the TTR binding channel (Chapter 7). The location of this hydroxylated PCB metabolite, deep in the binding channel, and the hydrogen bond formation could explain the stronger TTR binding affinity of this metabolite than the natural ligand T ₄ . The chlorine atoms present on the meta-positions of the 4,4'-(OH) ₂ -3,3',5,5'-tetraCB, metabolite, fitted easily in the T ₄ -iodine binding pockets present in the TTR binding channel. Additional computer-assisted graphics modelling studies on the interactions of several hydroxylated PCB metabolites with TTR, showed that hydroxy groups present on meta-positions could also form hydrogen bonds with the paired Serine 117 residues in the centre of the binding channel (Chapter 7). Furthermore, the modelling studies showed no significant differences between the interactions of hydroxylated PCB metabolites and the structurally related pentachlorophenol (PCP) with TTR, although in vivo studies by others indicated the disruption of the complex of retinol binding protein (RBP) and TTR after binding of a hydroxylated PCB metabolite but not by PCP in rodents. In conclusion, the detailed structural studies described in Chapter 7 confirmed the necessity of para- or meta- hydroxylation and adjacent chlorine substituents as structural elements of hydroxylated metabolites of PCBs and related compounds for interactions with TTR (Chapter 2), leading to selective retention of a specific hydroxylated PCB metabolite in plasma of rats exposed to a complex PCB mixture in vivo (Chapter 5). <strong>Concluding remarks</strong> The outcome of the present study clearly reveals the structural requirements that are essential for interactions of PHAH metabolites and other related chemicals for interactions with TTR. Hydroxylated PHAH metabolites can structurally resemble the thyroid hormone T ₄ Overall the structural requirements for TTR interaction were hydroxysubstitution on the para - or meta positions of one or both of the phenyl rings, with adjacent chlorine substitutions, herewith confirming some of the suggestions for TTR interactions of related hydroxylated PHAH metabolites by Rickenbacher et al . (1986). Especially the observation by X-ray crystallographic structure analysis that a hydrogen bond could be formed in the TTR binding channel upon binding of a hydroxylated PCB metabolite, provides strong evidence that para - or meta -hydroxylation of the PHAH compound forms an essential prerequisite for binding to the T4 binding site of TTR. No interactions with TTR were found for the tested parent PHAH compounds, contradictory to earlier suggestions of McKinney et al. (1985) which were based mainly on computer modelling and few in vitro binding studies (Rickenbacher et al , 1986). It should be noted that graphics modelling may show that parent PHAH compounds may fit the TTR binding site, but gives little information on binding affinity to TTR. In an affirmative in vitro T ₄ -TTR binding assay, several parent PCB congeners (3,3',4,4'-tetraCB, 3,3',4,4',5-pentaCB, 3,3',4,4',5,5'-hexaCB, 2,3,3',4,4'-pentaCB, 2,2',5,5'- tetraCB), TCDD and Aroclor 1254, a commercial PCB mixture, were tested at high concentrations and exhibited no inhibition of T ₄ -TTR binding (unpublished data). <img alt="Fig. 1." src="http://library.wur.nl/wda/abstracts/i1990_1.gif"/> PHAH metabolites that are predicted to have high binding affinities for TTR are indeed detected in plasma of rodents experimentally exposed to PCBs or PCDFs (Morse et al ., 1995b, Koga et al ., 1990, Kuroki et al. , 1993). For instance, exposure to a complex mixture of PCBs, Aroclor 1254, led to the selective retention in blood of a single PCB metabolite, 4-OH-2,3,3',4',5- pentaCB, which met the structural requirements for TTR binding (Bergman et al ., 1994). On the basis of the proposed structural requirements we can give an explanation for the interactions of related compounds with TTR as found by others, for instance pentachlorophenol (Van den Berg, 1990, Van Raay et al ., 1994, Den Besten et al ., 1991), natural compounds like flavones and halogenated aurones (Cody, 1989; Ciszak et al ., 1992) and certain drugs like milrinone (Wojtczak et al. , 1993). Moreover, these structural insights make it possible to predict whether other related classes of environmental contaminants can interact with TTR. In addition, due to combined exposure in the environment, one may expect additivity of binding to TTR of hydroxylated PHAH metabolites with other structurally related environmental or natural compounds. An important question is whether these experimental data can be extrapolated to other species. Two factors are essential for the TTR mediated selective retention of hydroxylated PHAH metabolites namely the presence of TTR in blood and the formation of relevant PHAH metabolites. It is most likely that hydroxylated. metabolites can be formed in many species other than rodent. Recently hydroxylated PCB metabolites were identified in human and seal plasma, which were environmentally exposed to background PCB levels (Bergman et al. , 1994). The major metabolites were again the 4-OH-2,3,3',4',5-pentaCB metabolite and to a lesser extent 4-OH-2',3,3',4',5-pentaCB in seal and human plasma and 4-OH-2,2',3,4',5,5',6-heptaCB in human plasma. Thus the hydroxylated PCB metabolites detected in vivo completely matched the structural requirements for TTR binding. The PHAH metabolite patterns in plasma of both experimentally and environmentally exposed animals and humans are species specific and depend not only on the structural requirements for binding to TTR, but also on the exposure situation and the capacity of biotransformation of PHAHs of the species. The species-specific metabolism of PHAHs decreases in the order: terrestrial mammals&gt;aquatic mammals&gt;birds&gt;fish (Safe, 1989). Several mammalian and avian species, like rats, seals, porpoises and eiderducks could form hydroxylated metabolites in in vitro microsomal incubations with the model substrate 3,3',4,4'-tetraCB (TCB). However fish, like trout and flounder, could not metabolise TCB, although cytochrome P4501A-like activity, responsible for biotransformation of planar PHAHs, can be induced (Murk et al., 1994, Morse et al., 1995c, Ishida et al., 1991). The selective retention of specific hydroxylated PHAH metabolites in plasma through binding to TTR is expected in species that both can metabolise PHAHs and posses TTR as a plasma thyroid hormone binding protein. TTR is a evolutionary conservative protein present in plasma of not only rodents but also other placental mammals, birds and to a lesser extent in reptiles. No TTR was detected in the lower species like fish and amphibians. In higher mammals like man, however, not only TTR but also thyroxine binding globulin (TBG), is present in blood as a primary thyroid hormone transport protein. In conclusion, hydroxylated PHAH metabolites can be formed and selectively retained by TTR in blood of a wide variety of species. The toxicological consequences of the TTR mediated selective retention of hydroxylated PHAH metabolites in plasma are not yet fully understood. The TTR protein plays a primary role in the transport of thyroid hormones in the blood of many species. Although TTR binds less T ₄ than TBG in human serum, TTR may be responsible for much of the immediate delivery of T ₄ and T ₃ to cells due to the lower binding T ₄ affinity. Furthermore TTR is important for the transport of retinol in blood by forming a complex with retinol-binding protein (Robbins, 1991). Disturbances in thyroid hormone plasma levels are found in several species experimentally or environmentally exposed to PHAHs, like rodents, seal (Brouwer et al., 1989) and man (Koopman Esseboom et al., 1994), species in which hydroxylated PCB metabolites were also present in plasma. Disruption of thyroid hormone homeostasis after exposure to PHAHs can however be caused by at least two mechanisms, eg. the disturbed plasma T ₄ transport through competitive binding of hydroxylated PHAH metabolites to TTR, but also the Ah-receptor mediated induction of T ₄ glucuronidation by parent compounds. The possible disruption of the TTR-RBP complex upon binding of a hydroxylated PHAH metabolite can also markedly decrease plasma retinol levels in rodents (Brouwer, 1987). It was suggested that seals exposed to PHAHs in the environment, have an impaired function of the immune system (de Swart, 1995), possibly resulting from disturbed retinoid levels (Brouwer, 1991, Brouwer et al. 1989). Hydroxylated PHAH metabolites may attribute to these effects on thyroid hormone and retinoid homeostasis through interactions with TTR in plasma. It is not known whether similar effects occur in man. TTR is the major thyroid hormone binding protein in cerebro-spinal fluid (CSF), suggesting a role in distribution of thyroid hormones in the central nervous system. This TTR is produced in the choroid plexus and is present in high concentrations in CSF of rats and humans, even at a very early stage in development. Moreover, in all species where TTR is present in blood, TTR has also been detected in brain. Because TTR is an important carrier for T ₄ to target tissues, for instance brain, one may expect that it may also act as a facilitated transport system for hydroxylated PHAH metabolites. This is in accordance with observations of a strong accumulation of hydroxylated PCB metabolites of maternal origin in the plasma and brain of late gestational fetuses from pregnant rats or mice exposed to PCBs (Morse et al., 1995b,d, Darnerud et al., 1995). In rat fetuses perinatally exposed to Aroclor 1254 the selective accumulation of the 4-OH-2,3,3',4',5-pentaCB metabolite in maternal plasma and fetal plasma and brain led to decreases in brain T ₄ levels, while brain T ₃ levels were only lightly changed. In addition plasma and hepatic retinoid concentrations were decreased in fetal and neonatal offspring (Morse et al., 1995a). The hydroxylated metabolites accumulated to high levels in fetal rat brain and may themselves attributeto observed neurochemical changes (Morse, 1995). Hydroxylated PHAH metabolites have been shown to possess biological activity in vitro (Brouwer, 1994). Hydroxylated PCB metabolites can interfere with mitochondrial structure and function in vivo and in vitro (Lans et al., 1990, Narashimhan et al., 1991). Moreover, they can bind to the Ah-receptor and weakly induce EROD activity. In addition, an in vitro marker of tumor promoting potential, the gap-junctional intercellular communication, could be weakly inhibited. Hydroxylated PCB metabolites can also exert (anti)-estrogenic activities in vivo (Bergeron et al., 1994) and in vitro (Kramer et al., 1994). No clear structure activity relationships for (anti)- estrogenicity could be found for the tested hydroxylated PCB metabolites. However, the hydroxylated PCB metabolites selectively retained in fetal plasma and brain (Morse et al., 1995d) do have a weak (anti)-estrogenic activity. The intrinsic capacity to disrupt endocrine systems, eg. thyroid and estrogen status, and the relatively large accumulating levels of hydroxylated PCB metabolites in late gestational rat fetuses, suggests there is a potential risk for adverse developmental effects by these hydroxylated PHAHs. This possible hydroxy PCBmediated route of developmental toxicity should be investigated in