Polychlorinated biphenyl-induced alterations of thyroid hormone homeostasis and brain development in the rat

IntroductionThe work described in this thesis was undertaken to gain insight in the processes involved in the developmental neurotoxicity of polychlorinated biphenyls. It has been previously hypothesized that the alteration of thyroid hormone status by PCBs may be in part responsible for the developmental neurotoxicity of these compounds in humans (Rogan et al. 1986). This is a logical hypothesis, given the well-described effects of PCBs on plasma thyroid hormone levels in adult animals, and the indisputable importance of thyroid hormones in brain development.Therefore the first goal was to determine the nature and mechanism of PCBinduced decreases in circulating and brain thyroid hormone levels in fetal and neonatal rats (Chapter 2,3,4 and 5). We examined the effects of maternal PCB administration on the metabolism of thyroid hormone in the brain and liver of fetal, neonatal and adult offspring in relation to the level of thyroid hormone in the plasma and brain. Vitamin A status, which may be linked to thyroid hormone status following PCB exposure, was also examined in one reproduction study (Chapter 6).Since the kinetics and metabolism of PCBs may play a pivotal role in the alteration of thyroid hormones, a radiolabelled and easily metabolized PCB congener was used in in vivo (Chapter 2) and in vitro experiments (Chapter 3) to examine the kinetics and metabolism of a model compound in pregnant and fetal rats. The relevance of this model compound for complex PCB mixtures was ascertained following the administration of a commercial PCB mixture to pregnant rats (Chapter 5).Lastly, neurochemical analysis were conducted on the brains of the offspring following maternal PCB exposure to determine which brain regions, cell types and neurotransmitter systems are affected during brain development (Chapter 7 and 8).Biotransformation of PCBs to thyroid hormone antagonistsThe metabolism and distribution of [ 14C]-3,3',4,4'-tetrachlorobiphenyI ([ 14C]- TCB) was examined in pregnant rats and their fetuses (Chapter 2 and 3, Morse et al., 1995). The major metabolite found in adult liver and plasma, placental tissue, whole fetuses and fetal blood was 3,3',4',5-tetrachloro-4-biphenyloI (4-OH-tetraCB). While maternal tissue levels of [ 14C]-TCB derived radioactivity significantly decreased by 65-85% over a 7 day period, radioactivity in the fetus accumulated more than 100-fold over the same period. The fetal accumulation of radioactivity was due primarily to 4-OH-tetraCB, and on day 20 of gestation, fetal plasma levels of 4-OH- tetraCB were 14 times higher than maternal plasma levels (14 μM vs 1 μM).In order to determine the source of 4-OH-tetraCB in the fetus, in vitro studies were carried out by incubating [ 14C]-TCB with maternal and fetal rat microsomes and analysing the reaction products with high pressure liquid chromatography and gas chromatographic/mass spectrometric analysis. First, incubation conditions were optimized using male rat microsomes. Under optimal incubation conditions, hepatic microsomes from pregnant rats pretreated with TCB produced 4-OH-tetraCB as the major metabolite, while no metabolites were detected in incubations with microsomes from fetuses from pregnant rats pretreated with TCB. The results indicate that 4-OH-tetraCB, found in the fetal compartment is due to transplacental transport from maternally formed 4-OH-tetraCB. This is in agreement with the observation that the biotransformation of TCB is dependent on CYP1A1 induction, and no CYP1A1 activity was observed in fetal microsomes after maternal treatment with TCB.In late gestation, the high levels of 4-OH-tetraCB found in the fetal plasma were asssociated with decreases in fetal plasma thyroid hormone levels in the absence of significant decreases in maternal plasma thyroid hormones. 4-OH-tetraCB has a high affinity for transthyretin (the major plasma thyroid hormone transport protein in the rat) and competitively displaces thyroxine from this protein (Brouwer et al. 1990, Lans et al. 1993). It was therefore concluded that the accumulation of 4-OH-tetraCB in the fetus is due to the high affinity of this metabolite for transthyretin, and results in significant decreases in fetal plasma thyroxine levels.Since the model compound 3,3',4,4'-TCB is present in only very low levels in the environment, it was of interest if the exposure of pregnant rats to a commercial PCB mixture (Aroclor 1254) would also result in the accumulation of phenolic metabolites in the fetal plasma (Chapter 5). Relatively high levels of hydroxylated PCB metabolites from penta, hexa. and hepta- chlorinated biphenyls have been found in the plasma of rats exposed to Aroclor 1254 and in environmentally exposed humans (Bergman et al. 1994). A significant accumulation of 4-OH- 2,3,3',4',5-pentachlorobiphenyI (4-OH-pentaCB) was found in the plasma of late gestational fetuses from pregnant rats exposed to Aroclor 1254 (up to 4.6 μM). This PCB metabolite has a 10-fold higher binding affinity for TTR than thyroxine, thereby confirming the relevance of work with the model compound, 3,3',4,4'-tetrachlorobiphenyl. In addition, relatively large amounts of 4-OH- pentaCB were found in the fetal (0.46 μM), but not weanling rat brain, indicating that in the absence of a functional blood-brain barrier hydroxylated PCB metabolites may enter the brain. The toxicological significance of this finding deserves investigation.Effects of PCB exposure on thyroid homone levels and metabolismT 4 -Uridine-diphospho-glucuronyl transferaseDecreases in plasma thyroid hormone levels in adult rodents may also be caused by the induction of the hepatic glucuronidation of thyroxine (Bastomsky, 1974, Barter and Klaasen 1992). The effect of a single maternal dose of 3,3',4,4',5,5'-hexachlorobiphenyl (HCB) on day 1 of gestation and in combination with repeated maternal doses of TCB (day 2 to 18 of gestation) on maternal, fetal and neonatal hepatic microsomal and brain thyroxine metabolism is described in Chapter 4. The results indicated that although maternal administration of coplanar PCBs may result in the induction of fetal hepatic microsomal T 4 glucuronidation, this induction did not cause the reductions in fetal plasma T 4 levels. Only the combined dose of HCB with TCB resulted in significant decreases in fetal plasma T 4 levels. This indicates that decreased placental transport of maternally-derived T 4 and the blockage of fetal thyroid hormone transport by 4-OH-tetraCB resulted in the decrease of fetal T 4 levels. In neonates and dams, however, the induction of T 4 glucuronidation by lactational exposure to coplanar PCBs may contribute to the observed decreases in plasma thyroxine levels.Maternal exposure to the commercial PCB mixture Aroclor 1254 also induced T 4 -UDPGT activity in hepatic microsomes from pregnant and weanling rats, but not in the fetus (Chapter 5). Since only the induction of maternal hepatic microsomal T 4 -UDPGT correlated with reductions in plasma thyroid hormones, it was concluded that the induction of T4-UDPGT activity played only a minor role in the reductions of plasma thyroid hormones in fetal and weanling rats. Large reductions in plasma thyroid hormones have also been observed following dietary Aroclor 1254 exposure in homozygous Gunn rats, which are deficient in T 4 -UDPGT activity (Collins and Capen, 1980a). The only long-term effect on thyroid hormone metabolism observed following maternal PCB exposure was a significant decrease in female hepatic microsomal T 4 glucuronidation in young adult offspring.Type II thyroxine 5'-deiodinaseAs most of the biologically active hormone triiodothyronine (T 3 ) is derived from T 4 by deiodination in the brain by Type II thyroxine 5'-deiodinase 5'D-II Silva and Larsen, 1982, Kaplan et al. 1983), it was of interest to examine the effects of PCBinduced reductions in plasma T 4 levels on 5'D-II activity. Decreases in brain T 4 levels result in a slower turnover of the enzyme, yielding a higher activity per unit protein in brain homogenates (Leonard et al. 1984). This regulatory mechanism is important in maintaining brain T 3 levels. In Chapter 4, the significant decreases in fetal, neonatal and weanling rat plasma T 4 levels following coplanar PCB exposure were accompanied by significant increases in 5'D-II activity in brain homogenates. It was concluded that the increases in 5'D-II activity were in compensation for low T 4 levels in the developing rat brain, which could be detrimental for normal brain development if insufficient T 3 was formed from T 4 .Following maternal exposure to Aroclor 1254, reductions in fetal plasma T 4 were also accompanied by increases in brain 5'D-II activity. However, in contrast to the effects observed with coplanar PCBs in weanling rats, 5'D-II activity was decreased in weanling rats with normal plasma and brain T 4 levels, and equal to control values when plasma and brain T, levels were decreased. This can not be explained by the current knowledge of 5'D-II regulation.Plasma and brain thyroid hormone levelsIn the current study, the effect of maternal PCB exposure on plasma thyroid hormone levels was transient, with only mild effects observed in weanling rats. Despite the significant lactational transfer of PCBs to the neonate, the effects on neonatal thyroid hormone homeostasis are less severe in neonates as in the fetus. Several mechanisms appear to be involved that may explain the difference in responses between fetuses and weanling rats: the induction of maternal hepatic T 4 glucuronidation late in gestation, the accumulation of hydroxylated PCB metabolites in the fetus, and reduced placental transfer of T 4 . Also the dilution of the tissue PCB levels during postnatal growth and the fecal and urinary excretion of PCBs may reduce the severity of plasma T 4 reductions in weanling rats following gestational PCB exposure. For example, the continuous postnatal dietary exposure of maternal rats to Aroclor 1254 results in low plasma T 4 levels throughout the weaning period (Collins and Capen, 1980b).Despite severe decreases in fetal plasma and brain T 4 levels following maternal PCB exposure, only marginal decreases were observed in fetal brain T 3 levels. This indicates that the late gestational rat fetus can maintain brain T 3 levels by an increase in 5'D-II activity, and is at little risk for PCB-induced hypothyroidism, at least in the brain.The observation that plasma TSH levels did not increase following PCB-induced decreases in plasma T 4 levels in the fetus and plasma T 3 and T 4 levels in the neonate suggests that the developing brain may have been euthyroid. However, the decreases in plasma T 4 levels themselves could be expected to result in an increase in TSH levels. Similar decreases in plasma T 4 levels in late gestational fetal Wistar rats following maternal treatment with methimazole have been shown to result in an 600% increase in plasma TSH levels (Morreale de Escobar et al. 1993), and it is likely that fetal TSH levels are modulated predominately by plasma T 4 rather than T 3 . In adult rats, significant increases in plasma TSH levels have been observed following dietary exposure to Aroclor 1254 that resulted in that decreases in plasma T 4 , but not T 3 at the same time point (Barter and Klaassen, 1994). A weak effect of PCBs on TSH secretion has been observed following a relatively high dietary exposure to Aroclor 1254, after which the rise in plasma TSH was suprisingly low in comparison to the rise in plasma TSH following dietary exposure to polychlorinated naphthalenes, which induced similar decreases in plasma T 4 levels as PCBs (Barter and Klaassen, 1994).In conclusion, maternal. PCB exposure during gestation results in a large decrease of fetal brain T 4 levels, but only marginal decreases in T 3 levels in the late gestational rat fetus. It is possible that earlier in gestation, before 5'D-II activity can compensate for decreases in brain T 4 levels, significant reductions in brain T 3 levels are induced by maternal PCB treatment.RetinoidsAnalogous to thyroid hormones, retinoids play a crucial role in brain development, although their most important effects are during early and mid-gestation (Adams, 1993). To evaluate retinoid status, plasma and hepatic retinol and retinylesters were determined following maternal Aroclor 1254 exposure. The reductions in plasma retinol levels may be caused by the accumulation of 4-OH- pentaCB; in the plasma, analogous to the disruption of the binding of retinol binding protein to transthyretin by 4-OH-tetraCB following exposure to 3,3',4,4'-tetrachlorobiphenyl. Although the effects of maternal PCB exposure on retinoid homeostasis in the fetus, neonate and young adult offspring appear to be minor, the regulation of retinoid homeostasis exhibited long-term alterations in the PCB exposed group.Alterations in neurochemistryIn Chapter 8 and 9 the effects of maternal PCB (Aroclor 1254) exposure were examined on the ontogeny of biogenic amines, a glial cell marker (glial fibrillary acidic protein, GFAP) and a neuronal cell marker (synaptophysin) in diverse brain regions.Biogenic aminesOf the biogenic amines examined, only the levels of 5-hydroxytryptamine (5-HT, serotonin) and its metabolite, 5-hydroxy-indoleacetic acid (5-HIAA) were altered by pre- and postnatal PCB exposure. It is notable that in adult animals the dopaminergic system is the most sensitive for exposure to commercial PCB mixtures, while we found no effects on the levels of dopamine or its major metabolite in the brains of PCB-exposed offspring (Seegal et al. 1985, 1986a, 1986b, 1991). Pre- and postnatal exposure to the lightly chlorinated PCB mixture Aroclor 1016 resulted in transient increases in striatal dopamine levels (Seegal, 1994). Therefore the effects of PCB exposure on regional brain monoamine metabolism during development do not resemble the effects in adult animals.In general, the effects can be characterized by an increase in 5-HIAA concentrations and the 5-HIAA/5-HT ratio in the lateral olfactory tract and prefrontal cortex, and an increase in 5-HIAA levels in the striatum and hippocampus on postnatal day 90. Since the effects on the serotonergic system are almost absent on day 21 postpartum when exposure to PCBs via lactation ceased, there appears to be a delayed effect on the ontogeny of serotonin metabolism.GFAP and Synaptophysin levelsThe most consistent effects of maternal PCB exposure on GFAP levels were observed in the lateral olfactory tract and the brainstem. Increases in GFAP concentrations were observed in both male and female offspring 21 and 90 days after birth. Increases were also observed in cerebellar GFAP levels on 21 and 90 days postpartum. In the brainstem of male and female offspring maternal PCB exposure prevented the increase in GFAP concentrations that was observed in control offspring, indicating a delay in the ontogeny of brainstem GFAP expression.Synaptophysin levels in the brain of the offspring were affected in a more complex manner than GFAP following maternal PCB exposure. Following maternal PCB exposure, the most sensitive brain regions from both sexes for decreases in synaptophysin concentrations on postnatal day 21 were the lateral olfactory tract and the brainstem. However, in young adult animals brainstem synaptophysin levels were significantly decreased in males and significantly increased in females. Synaptophysin levels were also significantly decreased in the striatum and hypothalamus of female, but not male offspring following maternal PCB exposure.The mechanisms involved in the alterations of GFAP and synaptophysin levels in the brains of the PCB-exposed offspring are not yet been elucidated. Increases in GFAP levels accompanied by decreases in synaptophysin levels in the lateral olfactory tract and prefrontal cortex are characteristic of reactive gliosis following neuronal loss (O'Callaghan and Miller, 1989). The decreases in GFAP and synaptophysin levels in the brainstem in weanling rats may be indicative of a developmental delay in brainstem maturation. The raphe nuclei in the brainstem contain serotonergic neurons which project to the lateral olfactory tract and the prefrontal cortex (Kosofsky and Molliver, 1987). It is therfore conceivable that the alterations in serotonergic matabolism as well as GFAP and synaptophysin levels in the lateral olfactory tract and prefrontal cortex result from a developmental delay in the serotonergic innervation of these brain regions.Relevance of the conducted research for human development and future toxicological researchThyroid hormonesIn a recently published study of 105 mother-infant pairs, elevated maternal body burdens of polychlorinated dibenzo-p-dioxins, dibenzofurans and biphenyls (levels in milk fat) were shown to be associated with alterations of human thyroid hormone status (Koopman-Esseboom et al. 1994). The effects are characterized as negative correlation of PCDD, PCDF and PCB congeners with maternal plasma TT 3 before delivery and maternal plasma TT 4 and TT 3 after delivery, and a positive correlation with plasma TSH levels in the infants in the second week and third month after birth. In infants with a higher exposure, plasma TT 4 levels were significantly lower (10%) and TSH levels were significantly higher (37%). Maternal body burden of three PCB congeners (CB 118, CB 138 and CB 153) was positively correlated with umbilical plasma TSH levels. In a similar study with 38 mother-infant pairs an increase in infant plasma TSH levels and plasma TT 4 levels was observed in infants with a higher exposure to the total toxic equivalents of PCDDs and PCDFs (Pluim et al., 1992).The relative effects of PCBs, PCDFs and PCBs on plasma thyroid hormone levels observed by Koopman-Esseboom et al. (1994) in mother-infant pairs are generally the same as observed in adult mice, rats and monkeys at much higher doses. In contrast to the decreases observed in late gestational fetal rat plasma TT 4 and FT 4 following maternal exposure to Aroclor 1254, no effect was observed of maternal body burden on umbilical cord TT 4 and FT 4 levels.It is unlikely that the alterations in thyroid hormone homeostasis associated with maternal PCB and PCDD levels observed in human newborns and infants will result in developmental alterations of the nervous system. Compensatory mechanisms, such as the induction of Type II 5'- thyroxine deiodinase in the brain, should offset decreases in plasma and brain T 4 levels. In the late gestational rat, near normal brain T 3 levels were maintained despite severe decreases in plasma and brain T 4 levels following maternal PCB exposure. However, the potential exists for significant reductions in fetal thyroid hormone levels earlier in gestation before compensatory mechanisms have fully developed or in specific brain regions not examined in this study.There are several important aspects in which the thyroid hormone transport differs between humans and rats with possible consequences for the effects of PCBs. The main thyroid hormone transport protein in humans is TBG, while in rats TTR is the major protein (Robbins, 1991). Although under certain circumstances TBG may be present in rats, it has a low affinity for T 4 (Rouaze-Romet et al. 1992). Hydroxylated PCB metabolites bind only very weakly to TBG, so it is possible that the impact of hydroxylated PCB metabolites on plasma T 4 levels in humans may be minor. However, the fetal mouse has both TTR and TBG as transport proteins, and mouse TBG has similar binding properties to human TBG (Vrancks et al. 1990), so the mouse may be a better model than the rat for studying the effects of PCBs on thyroid hormone transport. Recent research has shown that following maternal TCB exposure 4-OH-tetraCB accumulates in the fetal mouse, binds to TTR, resulting in the decrease of fetal plasma T 4 levels (unpublished results, D.C. Morse and P.O. Darnerud). Therefore, it is possible that transplacental. transport of hydroxylated PCBs in humans results in the decrease plasma and brain thyroid hormone levels before the rise of fetal hypothalamic-pituitary function in mid-gestation.While it is generally accepted that thyroid hormone deficiency in neonates and in late gestation has a negative effect on brain development in the rat as well as humans, the effects of thyroid hormone deficiency earlier in gestation are not clearly understood (Morreale de Escobar et al. 1993, review, Porterfield and Hendrich, 1993, review). Thyroid hormone and their receptors have been found in human fetuses by 10 weeks of gestation (Fisher, 1985), although the functional significance of these observations is currently unclear. The finding of Pharoah et al. (1972) that neurological damage of endemic cretinism could be prevented if iodized oil was given to the mother before the second trimester of pregnancy supports a role of thyroid hormones in brain development in this period.Therefore several questions remain to be answered: does maternal PCB exposure result in significant decreases in brain thyroid hormones in early and mid-gestation in rodents and humans, and whether such decreases are relevant for brain development.Effects on neurochemical developmentGestational and lactational exposure to a commercial PCB mixture, Aroclor 1254 resulted in long-term effects on the neurochemical development of the progeny of rats (Chapter 7 and 8). The study does not support the hypothesis that PCB-induced pre- or postnatal hypothyroidism was the cause of the neurochemical alterations (Chapter 5). The study does give an indication which neurotransmitter systems, which cell types and which brain areas may be affected by in utero and lactational exposure to a higher chlorinated PCB mixture, providing a solid base for further research.One of the questions which has interested researchers in PCB-induced toxicity for nearly 20 years is which PCB congeners are responsible for the toxicity of these compounds. This question has only been adequately answered for the immunotoxicity and some developmental endpoints (teratogenesis and fetotoxicity) in which the interaction of the PCB congeners with the Ah-receptor plays an important role.To date, reports have been published on the behavioral neurotoxicity of only two individual PCB congeners, 3,3',4,4'-tetrachlorobiphenyl and 3,3',4,4',5-pentachlorobiphenyl, both of which are coplanar PCBs with a high affinity for the Alireceptor. 3,3',4,4'-tetrachlorbiphenyl was shown to be a developmental neurotoxin in mice following high dose gestational exposure, reducing striatal dopamine and dopamine receptor levels in mice, delaying advoidance behavior and inducing neuropathological alterations in the cranial roots (Tilson et al. 1979, Chou et al. 1979, Agrawal et al. 1981). Postnatal exposure of mice to 3,3',4,4'-tetrachlorobiphenyl has also been shown to affect hippocampal muscarinic receptor levels and alter spontaneous activity (Eriksson, 1988, Eriksson et al. 1991). Due to the rapid metabolism of 3,3',4,4'-tetrachlorobiphenyl and accumulation of hydroxylated metabolites in the fetus no conclusions can be drawn as to the role of the parent compound or its metabolites in the developmental neurotoxicity of this PCB congener (Morse et al. 1995).On the other hand, exposure of pregnant rats to poorly metabolizable 3,3',4,4',5- pentachlorobiphenyl, delayed the onset of spontaneous activity and neuromuscular maturation in the offspring, which was related to delay in body weight gain. However, the development of reflexes and visual discrimination was not affected by maternal exposure to 3,3',4,4',5-pentachlorobiphenyl (Bernhoft et al. 1994). Taken together, these data may indicate that highly toxic coplanar PCBs may not be direct developmental neurotoxins in rodents. If this is the case, it is questionable if the use of individual PCB congeners in studies on the effects on neurochemical and behavioral development will resolve novel structure-activity relationships within the same time-frame as the structureactivity relationships for immunotoxicity or CYP1A1 induction in adult animals. First, the reproductive studies involved are much more lengthy and costly than acute studies with adult animals. Secondly, there is no general agreement on experimental protocols (timing and length of PCB administration, neurochemical and behavioral endpoints) between researchers working in this field, frustrating the comparison of data. Thirdly, it is very likely that complex interactions of Ah-receptor binding, PCB metabolism, fetal accumuation of metabolites and hormonal alterations affect brain development in vivo, so that the structure-activity relationships for individual congeners will not predict the effects of complex mixtures.Therefore it may be more useful to characterize the effects of environmentally relevant mixtures in terms of dose-response studies, neurochemical and behavioral endpoints and species sensitivity. Environmentally relevant mixtures can be obtained by extracting contaminated foodstuffs or constructing mixtures using synthetic standards. Although mixtures will vary somewhat in their composition, the results of such studies may be more relevant for regulatory purposes than data based on studies with individual congeners. Cell culture techniques using glial cell or dissociated neural cell cultures may be useful in investigating structure-activity relationships of the direct effects of individual PCB congeners on brain development. Parameters that have been first demonstrated to be affected in vivo should be analysed in vitro.Main conclusions:1) PCB congeners (3,3',4,4'-tetrachlorobiphenyl, 2,3,3',4,4'-pentachlorobiphenyl and 2,3',4',4,5-pentachlorobiphenyl) can be metabolized to hydroxylated metabolites which accumulate in the fetal plasma and brain and cause severe reductions in late gestational fetal plasma and brain thyroxine levels in rats.2) The reductions in brain T 4 levels in late gestational fetal rats are effectively compensated by increases in Type II thyroxine 5'-deiodinase activity, so that only marginal decreases in brain T 3 levels are observed following maternal exposure to a commercial PCB mixture. This is an indication that PCB-induced decreases in plasma T 4 levels are not responsible for alterations in the development of the central nervous system.3) maternal exposure to the commercial PCB mixture (Aroclor 1254) specifically alters the development of serotonin metabolism in the brain of the offspring in rats. Since the dopamine metabolism exhibits a greater sensitivity and persistency for the administration of Aroclor 1254 in adult rodents and macaques than serotonin metabolism, the mechanism of PCB-induced developmental neurotoxicity is distinct from the mechanism of alterations in biogenic amine metabolism in adult animals.4) The development of both neuronal and glial cells is affected in the brains of offspring from pregnant rats treated with Aroclor 1254. The alteration in astrocyte development in the brainstem of PCB-exposed offspring is not a response to neuronal death, for levels of glial fibrillary acidic protein (GFAP) are decreased, while increased neuronal death is generally accompanied by increases in GFAP expression. It is therefore likely that PCBs affect brain development by altering cell differentiation and proliferation.5) Since the brainstem is one of the first structures to develop in the brain, the observed alterations in brainstem development following pre- and postnatal PCB exposure probably have a negative effect on the subsequent development of other brain structures.