A clash of old and new scientific concepts in toxicity, with important implications for public health.
ABSTRACT A core assumption of current toxicologic procedures used to establish health standards for chemical exposures is that testing the safety of chemicals at high doses can be used to predict the effects of low-dose exposures, such as those common in the general population. This assumption is based on the precept that "the dose makes the poison": higher doses will cause greater effects.
We challenge the validity of assuming that high-dose testing can be used to predict low-dose effects for contaminants that behave like hormones. We review data from endocrinology and toxicology that falsify this assumption and summarize current mechanistic understanding of how low doses can lead to effects unpredictable from high-dose experiments.
Falsification of this assumption raises profound issues for regulatory toxicology. Many exposure standards are based on this assumption. Rejecting the assumption will require that these standards be reevaluated and that procedures employed to set health standards be changed. The consequences of these changes may be significant for public health because of the range of health conditions now plausibly linked to exposure to endocrine-disrupting contaminants.
We recommend that procedures to establish acceptable exposure levels for endocrine-disrupting compounds incorporate the inability for high-dose tests to predict low-dose results. Setting acceptable levels of exposure must include testing for health consequences at prevalent levels of human exposure, not extrapolations from the effects observed in high-dose experiments. Scientists trained in endocrinology must be engaged systematically in standard setting for endocrine-disrupting compounds.
- Environmental Toxicology 08/2013; · 2.71 Impact Factor
- Reproductive Toxicology 02/2013; · 3.14 Impact Factor
- International journal of circumpolar health. 01/2013; 72.
Collision of Basic and Applied
Approaches to Risk Assessment
of Thyroid Toxicants
R. THOMAS ZOELLER
Biology Department and Molecular and Cellular Biology Program, Morrill
Science Center, University of Massachusetts, Amherst, Massachusetts 01003,
ment; therefore, any environmental chemical that interferes sufficiently
with thyroid function, TH metabolism, or TH action may exert adverse
effects on brain development. Important known differences in aspects
of thyroid endocrinology between the fetus, infant, and adult allow us
to identify age-dependent vulnerabilities to thyroid toxicants with some
in the thyroid gland at different ages as well as the age-dependent sensi-
tivity to mild TH insufficiency. Several recent studies that describe risk
assessments of the environmental contaminant, ammonium perchlorate,
provide good examples of conclusions based on the selective considera-
tion of these known aspects of the thyroid system. Specifically, authors
who consider age-dependent differences in thyroid endocrinology sug-
gest that safe levels of perchlorate should be set at relatively low levels
(low parts per billion). In contrast, authors who do not consider these
known age-dependent differences in thyroid endocrinology recommend
million. Emerging evidence indicates that a variety of high production
volume chemicals can directly interact with the TH receptor. As test-
ing paradigms are designed by regulatory agencies, these age-dependent
differences in thyroid endocrinology must be considered.
KEYWORDS: thyroid hormone; brain development; perchlorate; PCB;
of hypothyroidism at birth are not readily or uniformly apparent. Before uni-
versal neonatal screening for congenital hypothyroidism (CH), these children
Address for correspondence: Professor R. Thomas Zoeller, Biology Department, Morrill Science
Center, 611 North Pleasant Street, University of Massachusetts, Amherst, MA 01003.
Ann. N.Y. Acad. Sci. 1076: 168–190 (2006). C ?2006 New York Academy of Sciences.
ZOELLER: RISK ASSESSMENT OF THYROID TOXICANTS 169
often were not identified for some months after birth.1,2The consequences
of this delay in diagnosis and treatment were disastrous. The mean full-scale
intelligence quotient (IQ) of CH infants was found to be 76 before univer-
sal screening,3but when categorized according to the timing of diagnosis,
it became evident that the older the infants were before they were diagnosed
to discover without reliance solely on clinical presentation.
Taken together, these studies demonstrate that TH is essential for brain
development from early fetal development (first trimester) to at least 2 years
after birth. Studies of the human fetal and neonatal brain demonstrate that
thyroid hormone receptors (TRs) are expressed early in development,5–8and
pattern, much like that described for the developing rat brain.9Thus, there
are two critical issues to address when evaluating the mechanisms by which
environmental chemicals might interfere with TH signaling in the developing
brain. The first is that if a chemical can cause a reduction in circulating levels
of TH, it is important to determine empirically whether the reduction in TH is
with TH signaling at the receptor, it is important to determine empirically
whether the effect is isoform specific. These two topics will be addressed in
THYROID HORMONE INSUFFICIENCY—TO WHAT EXTENT
CAN COMPENSATORY MECHANISMS AMELIORATE
THE CONSEQUENCES OF LOW THYROID HORMONE?
The thyroid endocrine system is governed by mechanisms that are in many
ways different from those governing the sex hormones, despite the fact that
both are controlled by classic neuroendocrine systems. These mechanisms
appear to allow a constant supply of TH to cells despite fluctuations in TH
synthesis or in circulating levels of TH.
The thyroid system is a classic neuroendocrine axis (FIG. 1); the hypotha-
lamus controls the pituitary gland, which in turn controls the thyroid, and
feedback mechanisms between thyroid secretions and the hypothalamus and
pituitary maintain the activity of this axis within relatively narrow limits.10
The active THs, thyroxine (T4), and triiodothyronine (T3) are formed in the
thyroid gland. These hormones are synthesized in an unusual way in that they
are derived from coupling two iodinated tyrosyl residues that make up the
larger hormone precursor, thyroglobulin (Tg). Tg is a large glycoprotein con-
taining two identical subunits each of nearly 3000 amino acids, creating a
660 kDa mature protein.11Following iodination, the protein is stored in the
170 ANNALS NEW YORK ACADEMY OF SCIENCES
ZOELLER: RISK ASSESSMENT OF THYROID TOXICANTS 171
← − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − −
FIGURE 1. The hypothalamic-pituitary-thyroid (HPT) Axis. Neurons whose cell bodies reside in the
hypothalamic paraventricular nucleus (PVN) synthesize the tripeptide TRH.123,124Although TRH-containing
neurons are widely distributed throughout the brain,125,126TRH neurons in the PVN project uniformly to the
median eminence,127,128a neurohemal organ connected to the anterior pituitary gland by the hypothalamic-
pituitary-portal vessels,129and are the only TRH neurons to regulate the pituitary-thyroid axis.21,130TRH
is delivered by the pituitary-portal vasculature to the anterior pituitary gland to stimulate the synthesis and
release of TSH.131TRH stimulates the synthesis of the TSH beta subunit.131However, TRH also affects the
post-translational glycosylation of TSH which affects its biological activity.132−137Interestingly, a recent
report by Nikrodhanond et al.138demonstrates that the role of TRH in regulating the pituitary-thyroid axis is
stronger than the role of TH negative feedback.
a beta subunit.139All three pituitary glycoproteins (luteinizing hormone, LH; follicle stimulating hormone,
FSH; and TSH) share the same alpha subunit.140Pituitary TSH binds to receptors on the surface of thyroid
follicle cells stimulating adenylate cyclise.141,142The effect of increased cAMP is to increase the uptake
of iodide into thyroid cells, increase iodination of tyrosyl residues on TG by thyroperoxidase, increase the
synthesis and oxidation of TG, TG uptake from thyroid colloid, and production of the iodothyronines T4and
T3. T4is by far the major product released from the thyroid gland.142Recent anatomical studies have shown
that human pituitary thyrotropes express the mRNA encoding the TSH receptor,143,144which may represent a
negative feedback loop accounting for the fact that serum TSH is reduced in some Gravesatients with normal
levels of TH.145
globulin (TBG), 15% is bound to transthyretin (TTR) and the remainder is bound to albumin.146TBG, the
least abundant but most avid T4binder, is a member of a class of proteins that includes cortisol binding
protein (CBP) and is cleaved by serine proteases in serum.147These enzymes are secreted into blood during
inflammatory responses and, in the case of CBP, can induce the release of cortisol at the site of inflammation.
among the vertebrates and may be developmentally regulated in a generalized manner. In the rat, high serum
levels of TBG are found in the fetus and the early postnatal pup,150,151adult levels of TBG are low, but low
across taxa appear to have the greatest carrying capacity for T4in serum during development compared to
their respective adult forms.152This may be a mechanism by which T4can be adequately maintained at a
developmental time when it is uniformly important.
THs (T4and T3) exert a negative feedback effect on the release of pituitary TSH and on the activity of
hypothalamic TRH neurons.18,19,124Although it is clear that TH regulates the expression of TSH153−155and
TRH18,123,124,156in a negative feedback manner, it is also clear that the functional characteristics of negative
feedback must include more than simply the regulation of the gene encoding the secreted protein/peptide. In
This fasting-induced suppression of TRH neurons results in the reduction of circulating levels of thyroid
hormone. In humans and perhaps in rodents, circulating levels of T4and of T3fluctuate considerably within
an individual; therefore, TSH measurements are considered to be diagnostic of thyroid dysfunction.15,159,160
However, individual T4levels in humans vary within far narrower limits than the population limits (i.e., the
population reference range).10,161In addition, variance in serum T4in pairs of monozygotic twins is far more
correlated than that in pairs of dizygotic twins or the general population.162Thus, the set-point around which
negative feedback appears to function has a very strong genetic component in humans and perhaps in other
T4and T3are actively transported into target tissues.163–170T4can be converted to T3by the action of
outer-ring deiodinases (ORD, type I and type II).22Peripheral conversion of T4to T3by these ORDs accounts
for nearly 80% of the T3found in the circulation.160Thyroid hormones are cleared from the blood in the liver
following sulfation or sulfonation by sulfotransferases, or following glucuronidation by UDP-glucuronosyl
transferase.171,172These modified THs are then eliminated through the bile.
T4and/or T3are actively concentrated in target cells about 10-fold over that of the circulation, although
transcription.173–177There are two genes that encode the TRs, c-erbA-alpha (TRa) and c-erbA-beta (TRb).
Each of these genes is differentially spliced, forming 3 separate TRs, TRa1, TRb1, and TRb2. The effects of
TH are quite tissue-, cell-, and developmental stage-specific and it is believed that the relative abundance of
the different TRs in a specific cell may contribute to this selective action.
172 ANNALS NEW YORK ACADEMY OF SCIENCES
colloid, the fluid filling the central core of the thyroid follicle. At the time
of hormone release, iodinated Tg is taken up into the cell from the colloid,
digested by lysosomal enzymes, liberating T3and T4into the blood.12T4is
the predominant iodothyronine released by the thyroid gland; circulating T3
is formed largely from peripheral deiodination of T4.13The pituitary glyco-
protein hormone, thyrotropin (TSH),14regulates the synthesis and secretion of
THs by activating adeylate cyclase in thyroid follicular cells. However, there
are a number of important extrathyroidal processes that combine to maintain
circulating THs within a relatively narrow concentration range. Although T4
is the predominant form of TH in the serum, T3is the active hormone at the
paper to “TH” to include both T4and T3, recognizing the differences between
thyrotropin-releasing hormone (TRH).18,19Although it is clear that TRH is a
to TSH regulation, including somatostatin, dopamine, and norepinephrine.16
Moreover, some investigators suggest that the primary role of TRH in the reg-
ulation of TSH secretion is to modulate the set-point around which TH act
on the pituitary.20,21Thus, circulating levels of TH, and the balance between
different forms of these hormones, are controlled by a number of processes.
on the receptor. This conversion from T4to T3is mediated by a deiodinase
enzyme, the type 1 or type 2 deiodinase (D1 or D2).22Moreover, a type 3
deiodinase (D3) can inactivate T3. Finally, there are cellular binding proteins
that may play a role in maintaining intracellular TH levels, or delivering the
hormone to the nucleus.23,24However, in the brain, this may be cell-type spe-
cific (FIG. 2).25,26That is, T4may be converted to T3in glial cells before T3
can act in neurons. This would require a T4transporter on glial cells and a
T3transporter on neurons. This appears to be the case inasmuch as the or-
ganic anion transporter 14 (OATP-14, a T4-selective transporter) is expressed
in areas of glia and the monocarboxylate transporter 8 (MCT8, a T3selective
transporter) is expressed in areas of neurons.27
Thus, the thyroid signaling system appears on the surface to be optimized
for maintaining a continuous supply of hormone to cells. Specifically, one can
imagine that as T4levels begin to decline, a number of compensatory mecha-
nisms respond to maintain tissue levels of T3. Specifically, TSH levels would
increase as a consequence of the negative feedback system. In addition, tissue
porters may respond in a manner that would cause a compensatory increase in
the ability of cells to sequester T4and/or T3. In addition, the thyroid glandit-
self appears to be very capable of responding to low iodide with mechanisms
ZOELLER: RISK ASSESSMENT OF THYROID TOXICANTS 173
FIGURE 2. Transport and cellular uptake of THs. The current model is that glial
transporter 8 (MCT8).27T3can act in the neuron to regulate gene expression, and it is
degraded by the action of the inner ring deiodinase, D3. This two-step process may also
occur in the pituitary gland.179
174 ANNALS NEW YORK ACADEMY OF SCIENCES
that increase iodide uptake. For example, the normal percentage uptake of
radioactive iodide in the euthyroid population is about 15–20% (e.g., see
Ref. 28). In contrast, patients with Graves’ disease exhibit a percentage uptake
of radioactive iodide of over 40%.29Thus, even the thyroid gland appears to
be capable of mounting compensatory responses to situations that would tend
to cause a reduction in circulating levels of TH.
thing not often seen in rodent studies), one must capture events downstream
of TH action in tissues before concluding that compensation has occurred.
Moreover, additional measures of compensatory mechanisms cannot make up
for the lack of information about TH action. For example, T4suppresses D2
expression in the brain30and while this observation is certainly important and
is consistent with the hypothesis that compensation has occurred, it does not
directly test the hypothesis that these apparently compensatory changes pro-
tect the tissue from damage due to low TH. Likewise, measurements of tissue
mechanisms have ameliorated effects of low T4on TH action for the same rea-
sons. Therefore, conclusions about whether compensation has occurred must
(e.g., TSH, sodium/iodide symporter [NIS], etc.). This has rarely if ever been
CONSEQUENCES FOR RISK
ASSESSMENT—THE CASE OF PERCHLORATE
Ammonium perchlorate is the principal oxidant for solid propellants in the
defense industry.31,32Perchlorate contamination of ground water across the
United States has recently become apparent33and, therefore, it is important to
determine a “safe” level of perchlorate to which humans can be exposed. Per-
chlorate inhibits iodide uptake into the thyroid gland by the NIS.34This action
of perchlorate can lead to an inhibition of TH synthesis, and thus a reduction
treatment of hyperthyroidism, and its potential toxicity as an environmental
contaminant. To establish the dose–response in humans for perchlorate inhibi-
et al.35gave perchlorate in drinking water at 0.007, 0.02, 0.1, or 0.5 mg/kg
per day to 37 male and female volunteers for 14 days. In 24 subjects, 8- and
exposure, on exposure days 2 and 14 (E2 and E14), and 15 days postexposure
ZOELLER: RISK ASSESSMENT OF THYROID TOXICANTS 175
In general, this study allowed the estimation of a “no-effect level” of per-
chlorate of 5.2 or 6.4 ?g/kg per day.35This value was calculated based on the
dose–response of perchlorate consumption on the ability of the thyroid gland
be a threshold below which perchlorate would have no effect. The concentra-
tion of perchlorate in drinking water required to deliver this dose to a healthy
adult is approximately 200 ppb.
tions of about 200 ppb (and possibly higher) should be of no health concern in
an iodine-sufficient population.35Although this conclusion is marginally de-
fensible for normal, euthyroid adults, several key aspects of the normal adult
thyroid system are significantly different from that of infants, which makes
population—fetuses, neonates, and infants.
Specifically, Greer et al.35postulated that 0.5 mg/kg per day of perchlorate
failed to influence circulating levels of TH in healthy adults in 14 days of
exposure because the normal adult thyroid gland contains a very large storage
capacity of unreleased TH. In fact, these authors estimate that there should be
sufficient hormone stored in the adult thyroid gland to last for several months.
Thus, no concentration of perchlorate should cause a reduction in circulating
levels of TH of a normal euthyroid adult within 14 days. The case is quite dif-
Ref. 37) estimated that the neonatal thyroid gland contains TH equivalent to
only a single day’s secretion. This estimate was revised by van den Hove et
al.37and Savin et al.38who empirically measured intrathyroidal stores of TH
in human fetuses and neonates and found that the amount of hormone stored
in the colloid is less than that required for a single day. Thus, inhibiting TH
synthesis in a fetus/neonate would immediately be manifested as a decrease in
circulating levels of TH.
ered performing risk analysis of a thyroid toxicant such as perchlorate. First,
neonates have sole source of nutrition—milk—and perchlorate may be con-
centrated in milk. Perchlorate acts on the NIS,34a protein that is induced in
lactating breast tissue by prolactin.39–42Thus, it is possible that perchlorate
is concentrated in milk.43,44In support of this, Kirk et al. have found some
perchlorate is found in a number of human milk samples and that the level
of perchlorate in milk was not predicted by the level of perchlorate in their
drinking water. Thus, it is not possible to conclude from this study that milk
widespread than is predicted by contaminated drinking water.
Second, a short period of TH insufficiency may produce permanent neuro-
176 ANNALS NEW YORK ACADEMY OF SCIENCES
insufficiency of as little as 14 days may be long enough to produce permanent
neurological deficits in neonates.46Finally, infants drink six times as much
fluid per unit body weight as adults. Therefore, in the example of perchlorate,
a drinking water concentration of 200 ppb would deliver about 129 ?g per-
chlorate to a 4 kg infant being fed a formula prepared with tap water. This
would be a dose of about 32.2 ?g/kg, which would undoubtedly inhibit iodide
uptake in this infant.
Considering these differences between the thyroid system of adults and in-
fants, it is predictable that any chemical, such as perchlorate, that reduces TH
extent iodide uptake must be inhibited by perchlorate to cause a reduction in
inhibition would be required for deleterious effects to be produced by perchlo-
rate.47This estimate, as articulated in the report of the National Academy
Report is without citation. To our knowledge, there are no quantitative studies
defining the degree to which iodide uptake inhibition must occur to inhibit TH
synthesis. Thus, for human infants, it remains to be determined the degree to
which they can tolerate iodide uptake inhibition.
A significant issue is also that infants cannot tolerate TH insufficiency
for many days. For example, long-term studies of children with CH that
have been treated with T4replacement indicate that very subtle differences
in circulating levels of T4are associated with significant differences in in-
tellectual performance later in life.48–51Thus, the consequences of perchlo-
rate exposure to infants could be the production of a lifetime of cognitive
deficit, the severity of which would be proportional to the severity of TH
that recent reviews of perchlorate risk assessment fail to mention the infant.
the 10-fold default uncertainty factor is needed for intraspecies (i.e., within
human) variability to protect such hypothetical susceptible sub-populations”
(p. 52). Although their focus was on vegetarians who may (or may not) be
iodide insufficient, their presumptive concern about the 10-fold uncertainty
factor to protect vegetarians failed to mention infants as a subpopulation of
humans with known vulnerabilities to TH insufficiency. Perhaps even more
remarkably, Crump and Gibbs53performed benchmark dose analysis using
changes in serum free T4and TSH (which are not changed in the three studies
used in their analysis). Thus for the Greer study, which could not have iden-
tified reduced T4or TSH in a 14-day study for reasons described above, the
benchmark dose levels ranged from about 50 to 60 mg per day (around 850
?g/kg per day). Again for this analysis, there is no mention of the differences
in the thyroid system between adults and infants or the differences in conse-
quences of TH insufficiency between adults and infants. Moreover, 850 ?g/kg
ZOELLER: RISK ASSESSMENT OF THYROID TOXICANTS 177
In contrast, Ginsberg and Rice54come to a conclusion that is the oppo-
site of that proposed by Fields et al.52Specifically, they argue that additional
sumption proposed by different authors appears to be dependent upon whether
infants are considered in the analysis. However, because most infants are nor-
mal and all normal infants have very well-characterized differences in their
fail to include infants as a vulnerable subpopulation worthy of protection.
ENVIRONMENTAL CHEMICALS AS THYROID HORMONE
MIMICS: A CHALLENGE FOR RISK ASSESSMENT
Despite early speculations that environmental chemicals may act as imper-
several recent reports show that a broad range of chemicals to which humans
are routinely, and inadvertently, exposed can bind to TRs and may produce
rinated biphenyls (PCBs)—industrial chemicals consisting of paired phenyl
was banned in the mid 1970s, these contaminants are routinely detected in the
environment58and in human tissues59at high concentrations. PCB body bur-
den is associated with lower full-scale IQ, reduced visual recognition memory,
attention, and motor deficits,60–65and structural differences in areas of white
PCBs reduce circulating levels of T4 in animals,68–72and some authors
propose that PCBs exert neurotoxic effects on the developing brain by caus-
ing a state of relative hypothyroidism.73–75This concept is supported by the
observations that the ototoxic effect of PCB exposure can be partially amelio-
rated by T4replacement,76and that the cerebellum, a tissue highly sensitive
to TH insufficiency,77–79is targeted by PCB exposure. PCB exposure alters
motor behavior associated with cerebellar function,80,81as well as cerebel-
lar anatomy.81Interestingly, PCB exposure is associated with an increase in
expression of glial fibrillary acidic protein,81which is also increased by TH
insufficiency.82Finally, in young children, the association between PCB body
burden and behavioral measures of response-inhibition is stronger in those
children that have a smaller corpus callosum,66an area of the brain affected by
TH.83–85Thus, it is possible that PCBs exert at least some neurotoxic effects
on the developing cerebellum by causing a state of relative hypothyroidism.
However, PCB exposure does not produce effects on animals that are fully
consistent with effects of TH insufficiency, such as body weight gain during
178 ANNALS NEW YORK ACADEMY OF SCIENCES
development68,86,87or the timing of eye opening.69In addition, despite the
reduction in serum T4, PCB exposure increases the expression of several TH-
responsive genes in the fetal86,87and neonatal68brain, indicating that at least
vivo. Recently, Kitamura et al.88reported that nine separate hydroxylated PCB
congeners can bind to the rat TR with an IC50as low as 5 ?M. In addition,
using a human neuroprogenitor cell line, Fritsche et al.89found that a specific
PCB congener could mimic the ability of T3in increasing oligodendrocyte
differentiation, and that this effect was blocked by the selective TR antagonist
NH3. Finally, Arulmozhiraja and Morita90have identified several PCB con-
geners that exhibit weak TH activity in a yeast two-hybrid assay optimized to
identify such activity.
However, not all recent reports indicate that PCBs act as agonists on the TR.
Kimura-Kuroda et al.91found that two separate hydroxylated PCBs interfere
(Aroclor 1254) exhibited specific binding to the rat TR? at approximately
10 ?M. This concentration inhibited TR action on the malic enzyme (ME)
promotor in a CAT assay and this effect required an intact TR element (TRE).
However, the PCB mixture did not alter the ability of TR to bind to the ME
TRE in a gel shift assay. In contrast, Iwasaki et al.93found that a specific
hydroxylated PCB congener inhibits TR-mediated transcriptional activation
in a luciferase assay at concentrations as low as 10−10M. This effect was
observed in several cell lines, but was not observed using a glucocorticoid
response element. Miyazaki et al.94followed this report by showing that PCBs
can dissociate TR:RXR heterodimers from a TRE.
It is clear that PCBs are neurotoxic in humans and animals, and that they
on TR action appear to be quite complex. This complexity includes acting as
an agonist or antagonist and may include TR isoform selectivity inasmuch as
most studies have been performed using the TR?, leaving the TR? relatively
unstudied in this context. In addition, considering that there are 209 different
chlorine substitution patterns on the biphenyl backbone and that these can be
metabolized (hydroxyl- and methylsulfonyl-metabolites95), it is possible that
different chemical species exert different effects. Finally, PCBs may exert dif-
structure, or different co-factors. This complexity will be important to pursue
because the effects of PCB exposure in humans is far better studied than for
structurally related compounds such as polybrominated biphenyls (PBBs) and
polybrominated diphenyl ethers (PBDEs). Thus, mechanistic studies on PCBs
ample of an environmental chemical to which humans are exposed that may
produce complex effects on TH action. BPA is produced at a rate of over
ZOELLER: RISK ASSESSMENT OF THYROID TOXICANTS 179
800 million kg annually in the United States alone,96and is used primarily
in the manufacture of plastics including polycarbonate plastics, epoxy resins
that coat food cans, and in dental sealants.97,98Howe et al.97estimated hu-
man consumption of BPA from expoxy-lined food cans alone to be about 6.6
?g/person per day. BPA has been reported in concentrations of 1–10 ng/mL
in serum of pregnant women, in the amniotic fluid of their fetus, and in cord
serum taken at birth.99,100Moreover, BPA concentrations of up to 100 ng/g
were reported in placenta.99BPA is also halogenated (brominated or chlori-
nated) to produce flame-retardants. Tetrabromobisphenol-A (TBBPA) is the
most commonly used with over 60,000 tons produced annually.101,102Thom-
sen et al.103recently reported that brominated flame retardants, including
TBBPA, have increased in human serum from 1977 to 1999 with concen-
trations in adults ranging from 0.4 to 3.3 ng/g serum lipids. However, infants
(0–4 years) exhibited serum concentrations that ranged from 1.6 to 3.5 times
Considering this pattern of human exposure, it is potentially important that
BPA has been shown to bind to the TR.104Best characterized as a weak estro-
gen,105binding to the estrogen receptor (ER) with a Kiof approximately 10−5
M,106,107BPA binds to and antagonizes T3activation of the TR108,109with a
Kiof approximately 10−4M, but as little as 10−6M BPA significantly inhibits
TR-mediated gene activation.109Moreover, Moriyama et al. found that BPA
reduced T3-mediated gene expression in culture by enhancing the interaction
with the co-repressor N-CoR.104Interestingly, we have found that develop-
mental exposure to BPA in rats produces an endocrine profile similar to that
observed in thyroid resistance syndrome.110Specifically, T4levels were ele-
vated during development in the pups of BPA-treated animals, but TSH levels
were not different from controls.111This profile is consistent with BPA inhi-
bition of TR?-mediated negative feedback. However, the TH-response gene,
RC3, was elevated in the dentate gyrus of these BPA-treated animals.111Be-
cause the TR? isoform is expressed in the dentate gyrus, we concluded that
BPA may be a selective TR? antagonist in vivo.
If BPA acts as a TR antagonist in vivo, it is predictable that specific devel-
opmental events and behaviors would be affected by developmental exposure
to BPA. In this regard, Seiwa et al.112have shown that BPA blocks T3-induced
oligodendrocyte development from precursor cells (OPCs). In addition, it is
This is important because the apparent endocrine profile induced by BPA, in
which serum T4is elevated despite normal TSH levels, and in which the TR?
receptor appears to be overstimulated, is similar to the endocrine profile in
humans caused by a mutation in the TR? receptor. The subsequent syndrome,
thyroid resistance syndrome,114is associated with a number of symptoms in-
cluding attention deficit disorders.110,115Thus, it is possible that the endocrine
180 ANNALS NEW YORK ACADEMY OF SCIENCES
Despite the antagonistic effects of BPA on the TR?, halogenated BPAs
appear to act as TR agonists.108Both TBBPA and tetrachlorobisphenol A
TR and this could be important during early brain development. For example,
TH of maternal origin can regulate gene expression in the fetal brain,116–118
one of these genes codes for Hes1.87Considering the role of HES proteins
in fate specification in the early cortex,119–121the observation that industrial
chemicals can activate the TR and increase HES expression87may indicate
that these chemicals can exert subtle effects on early differentiative events. In
addition, TH exerts equal and opposite effects on the numbers of oligodendro-
cytes and astrocytes found in the corpus callosum and anterior commissure,122
indicating that TH controls the balance of production of these two cell types
in the postnatal brain. Similarly to events in the early fetal brain controlled by
Hes1, inappropriate TH stimulation may have long-term consequences on the
balance of glial cell types that make up major bridging structures in the brain.
The human population is exposed to a large number of chemicals that can
influence thyroid function and, perhaps, TH action. Current screens and tests
on thyroid function as indicated by changes in hormone levels. Mechanistic
studies will be required to fully characterize thyroid toxicants because it is
becoming clear that some compounds can exert direct effects on TH signaling
toxicants at different life stages because both the sensitive to various toxicants
and the ultimate consequences of exposures will differ among individuals of
1. JACOBSEN, B.B. & N.J. BRANDT. 1981. Congenital hypothyroidism in Denmark:
incidence, types of thyroid disorders and age at onset of therapy in children:
1970-1975. Arch Dis Child. 56: 134–136.
2. ALM, J. et al. 1984. Incidence of congenital hypothyroidism: retrospective study
to diagnosis. Br. Med. J. 289: 1171–1175.
3. KLEIN, R. 1980. History of congenital hypothyroidism. In Neonatal Thyroid
Screening. G.N. Burrow & J.H. Dussault, Eds.: 51–59. Raven Press. New York.
The Thyroid: a Fundamental and Clinical Text. L.E. Braverman & R.D. Utiger,
Eds.: 984–988. Lipponcott-Raven. Philadelphia.
ZOELLER: RISK ASSESSMENT OF THYROID TOXICANTS 181
5. CHAN, S. & M.D. KILBY. 2000. Thyroid hormone and central nervous system
development. J. Endocrinol. 165: 1–8.
6. KILBY, M.D. et al. 2000. Expression of thyroid receptor isoforms in the human
fetal central nervous system and the effects of intrauterine growth restriction.
Clin Endocrinol. (Oxf.) 53: 469–477.
7. CHAN, S. et al. 2002. Early expression of thyroid hormone deiodinases and re-
ceptors in human fetal cerebral cortex. Brain Res. Dev. Brain Res. 138: 109–
8. KILBY, M.D. 2003. Thyroid hormones and fetal brain development. Clin. En-
docrinol (Oxf). 59: 280–281.
9. BRADLEY,D.J.,H.C.TOWLE&W.S.Y OUNG.1992.Spatialandtemporalexpression
of alpha- and beta-thyroid hormone receptor mRNAs, including the beta-2
subtype, in the developing mammalian nervous system. J. Neurosci. 12: 2288–
10. ANDERSEN, S. et al. 2002. Narrow individual variations in serum T(4) and T(3)
in normal subjects: a clue to the understanding of subclinical thyroid disease.
J. Clin. Endocrinol. Metab. 87: 1068–1072.
11. SPENCER, C.A. 2000. Thyroglobulin. In Werner and Ingbar’s The Thyroid: a Fun-
damental and Clinical Text. L.E. Braverman & R.D. Utiger, Eds.: 402–413.
Lippincott Williams and Wilkins. Philadelphia, PA.
12. TAUROG, A. 2004. Hormone synthesis: thyroid iodine metabolism. In The Thy-
roid: a Fundamental and Clinical Text. L.E. Braverman & R.D. Utiger, Eds.:
61–85. Lippincott-Raven. Philadelphia.
13. LEONARD, J.L. & J. KOEHRLE. 1996. Intracellular pathways of iodothyronine
metabolism. In The Thyroid: a Fundamental and Clinical Text. L.E. Braver-
man & R.D. Utiger, Eds.: 125–161. Lippincott-Raven. Philadelphia.
14. SPAULDING, S.W. 2000. Biological actions of thyrotropin. In The Thyroid: a Fun-
damental and Clinical Text. L.E. Braverman & R.D. Utiger, Eds.: 227–233.
Lippincott Williams and Wilkins. Philadelphia.
15. STOCKIGT, J.R. 2000. Serum thyrotropin and thyroid hormone measurements and
assessment of thyroid hormone transport. In The Thyroid: a Fundamental and
16. MORLEY, J.E. 1981. Neuroendocrine control of thyrotropin secretion. Endocrine
Rev. 2: 396–436.
17. SCANLON, M.F. & A.D. TOFT. 2000. Regulation of Thyrotropin Secretion. In The
Thyroid: a Fundamental and Clinical Text, 8th Edition. L.E. Braverman & R.D.
Utiger, Eds.: 234–253. Lippincott William & Wilkins. Philadelphia.
18. KOLLER, K.J. et al. 1987. Thyroid hormones regulate levels of thyrotropin-
releasing hormone mRNA in the paraventricular nucleus. Proc. Natl. Acad.
Sci. USA. 84: 7329–7333.
19. RONDEEL, J.M.M. et al. 1989. In vivo hypothalamic release of thyrotropin-
releasing hormone after electrical stimulation of the paraventricular area: com-
parison between push-pull perfusion technique and collection of hypophysial
portal blood. Endocrinology 125: 971–975.
hormone negative feedback on the pituitary thyrotroph. Neuroendocrinology.
182 ANNALS NEW YORK ACADEMY OF SCIENCES
21. TAYLOR, T. et al. 1990. The paraventricular nucleus of the hypothalamus has a
major role in thyroid hormone feedback regulation of thyrotropin synthesis and
secretion. Endocrinology 126: 317–324.
22. ST. GERMAIN, D.L. & V.A. GALTON. 1997. The deiodinase family of selenopro-
teins. Thyroid 7: 655–668.
23. NISHII, Y. et al. 1993. Induction of cytosolic triiodo-L-thyronine (T3) binding
protein (CTBP) by T3 in primary cultured rat hepatocytes. Endocr. J. 40: 399–
24. MORI, J. et al. 2002. Nicotinamide adenine dinucleotide phosphate-dependent
cytosolic T(3) binding protein as a regulator for T(3)-mediated transactivation.
Endocrinology 143: 1538–1544.
25. BERNAL, J., A. GUADANO-FERRAZ & B. MORTE. 2003. Perspectives in the study
of thyroid hormone action on brain development and function. Thyroid. 13:
26. BERNAL, J. 2005. The significance of thyroid hormone transporters in the brain.
Endocrinology. 146: 1698–1700.
27. HEUER, H. et al. 2005. The monocarboxylate transporter 8 linked to human psy-
chomotor retardation is highly expressed in thyroid hormone-sensitive neuron
populations. Endocrinology 146: 1701–1706.
28. BRAVERMAN, L.E. et al. 2005. The effect of perchlorate, thiocyanate, and ni-
trate on thyroid function in workers exposed to perchlorate long-term. J. Clin.
Endocrinol. Metab. 90: 700–706.
29. MESTMAN, J.H. 1999. Diagnosis and management of maternal and fetal thyroid
disorders. Curr. Opin. Obstet. Gynecol. 11: 167–175.
30. BURMEISTER, L.A., J. PACHUCKI & D.L. ST. GERMAIN. 1997. Thyroid hor-
mones inhibit type 2 iodothyronine deiodinase in the rat cerebral cortex
by both pre- and posttranslational mechanisms. Endocrinology 138: 5231–
31. EPA, U. S. 1998. Perchlorate environmental contamination: toxicological review
and risk characterization based on emerging information. http://www.epa.gov/
33. URBANSKY, E.T. 1998. Perchlorate chemistry: implications for analysis and re-
mediation. Bioremed. J. 2: 81–95.
34. WOLFF, J. 1998. Perchlorate and the thyroid gland. Pharmacol. Rev. 50: 89–
contamination: the dose response for inhibition of thyroidal radioiodine uptake
in humans. Environ. Health Perspect. 110: 927–937.
36. VULSMA, T. 1991. Etiology and Pathogenesis of Congenital Hypothyroidism.
Evaluation and Examination of Patients Detected by Neonatal Screening in the
Netherlands. Academisch Proefschrift, Amsterdam.
37. VAN DEN HOVE, M.F. et al. 1999. Hormone synthesis and storage in the thyroid of
human preterm and term newborns: effect of thyroxine treatment. Biochimie.
of human neonates. J. Pediatr. Endocrinol. Metab. 16: 521–528.
39. SPITZWEG, C. et al. 1998. Analysis of human sodium iodide symporter gene
expression in extrathyroidal tissues and cloning of its complementary deoxyri-