PBDE flame retardants

Abstract

Polybrominated diphenyl ethers (PBDEs) are a class of brominated flame retardant chemicals that have been used in large quantities and are now detected worldwide in humans and wildlife. To complement reviews of effects on human health, this review discusses and synthesizes current evidence of PBDE toxicokinetics and toxicity mechanisms leading to perturbations of thyroid hormone homeostasis in fish. PBDE disruptions to thyroid signaling in fish appear to proceed through multiple pathways involving declines in circulating thyroid hormones, disrupted deiodination activity, hindered hormone transport, and altered transcriptional regulation of genes involved in thyroid hormone production, transport, and genomic signaling. PBDE exposures have also been linked to impacts on reproductive health with reductions in fecundity, spawning, hatching success, and offspring survival observed in some species, as well as impaired fertility. These studies on PBDE mediated hormone disruption in fish can help inform future studies seeking to understand potential developmental effects in humans.

Keywords: :

• endocrine disruption
• feminization
• flame retardant
• neurodevelopment
• neurotoxicity
• polybrominated diphenyl ether
• reproduction
• thyroid hormone
• thyroxine
• thyroid receptor

Introduction

Polybrominated diphenyl ethers (PBDEs) are a class of brominated flame retardant (BFR) chemicals that have been added to many consumer and commercial products including textiles, carpeting, construction materials, and electronics in an effort to reduce their combustibility. Although these compounds have been both banned or phased-out from production in a number of countries, human and environmental exposures continue as products that contain these chemicals are still used and recycled, and present legacy contamination problems upon disposal. As a consequence, PBDEs are widespread and persistent contaminants in both living and non-living parts of the global environment.^1-9

PBDEs can have from 1–10 bromine atoms substituted on diphenyl ether (Fig. 1). There are 209 PBDE congeners (BDE-1 to BDE-209) theoretically possible depending on the number and substitution patterns of bromine. In practice, however, the number of congeners formed is limited based on the chemical properties and composition of the PBDE commercial mixtures. Three PBDE commercial mixtures have been produced: PentaBDE, OctaBDE, and DecaBDE. The PentaBDE commercial mixture is a heterogeneous mix of tetra-, penta-, and hexaBDEs and was added mostly to polyurethane foams and textiles, and to a lesser extent in epoxy and phenolic resins and polyesters. The majority (~95%) of PentaBDE was used in the US where it was added to polyurethane foams in furniture cushioning and could constitute up to 30% by weight of these products.¹⁰ The OctaBDE commercial mixture is made up of higher molecular weight (MW) constituents, hepta, octa, and nona-BDEs and was added predominantly to acrylonitrile butadiene styrene (ABS) used in the plastic housing of office equipment and electronics. The production and use of PentaBDE and OctaBDE were phased out in the US and banned in the European Union in 2004 due to concerns about their persistence, bioaccumulation, and toxicity. In 2009, these products were also listed as persistent organic pollutants (POPs) under the United Nations Stockholm Convention.¹¹ The DecaBDE mixture contains the fully brominated congener decabromodiphenyl ether (BDE-209; ~97%) with trace amounts of nonaBDEs. It is used as an additive in high impact polystyrene, polyolefin, and polypropylene plastics used in electronics, automobiles, airplanes, and construction and building materials. To achieve US fire safety standards for plastics, high impact polystyrene plastics typically contain ~10–15% of DecaBDE.¹² DecaBDE is one of two dominant BFRs used worldwide with 2007 global consumption estimated at ~73 000 t (~161 million pounds).^13,14 The production of DecaBDE was discontinued in the US at the end of 2013 under a voluntary phase-out and has been restricted from use in electronics in the European Union since 2008. It is generally unrestricted from use in Asia.¹⁵

Figure 1. Structural comparison of PBDEs, PBBs, PCBs, and thyroxine hormone

PBDEs structurally resemble polybrominated biphenyls (PBBs), polychlorinated biphenyls (PCBs) and some biomolecules, most notably thyroid hormones (THs; Figure 1). Because PBDEs are not chemically bound but are rather added to plastics, they can enter the environment during production and may be released into the surrounding environment and biota with the breakdown and volatilization of the parent polymer. Like other persistent, hydrophobic chemicals, the most important route of uptake in aquatic animals appears to be by trophic transfer and consumption of foods contaminated with PBDEs.^16,17 This dietary exposure pathway in aquatic animals is distinguished from human uptake that appears to depend on both dietary exposures and the incidental ingestion of PBDE-containing dust.^18,19

Despite the discontinued use of PentaBDE, its constituents, including BDE-47 (2,2’4,4’-tetraBDE), BDE-99 (2,2’,4,4’,5-pentaBDE), BDE-100 (2,2’,4,4’,6-pentaBDE), BDE-153 (2,2’,4,4’,5,5′-hexaBDE), and BDE-154 (2,2’,4,4’,5,6’-hexaBDE), continue to be dominant PBDEs frequently detected in humans and wildlife worldwide despite the generally more limited use of PentaBDE outside the US.^20,21 Potential sources of these congeners are likely related to the ongoing use and recycling of products that contain PentaBDE as well as their high environmental persistence and long-range global transport potential.²² Another source of these lower MW PBDEs may be attributable to the breakdown of higher PBDEs, such as BDE-209, which can undergo photolytic degradation,²³ microbial breakdown,²⁴ and metabolic biotransformation²⁵ to lower MW congeners.

Recent attention has focused on the potential for BDE-209 and other highly brominated PBDEs to bioaccumulate. BDE-209 is now the dominant PBDE measured in abiotic compartments, typically at ppb to ppm levels (ng/g dry weight [dw] to low mg/g dw) in dust,^18,26 soils and sediments,^27-29 and biosolids.^30,31 Studies show that BDE-209 is bioaccumulating in a large number and variety of biota residing all over the world, including, for example, in birds and bird eggs,^32,33 terrestrial and aquatic mammals,^34-36 and plankton, fish and shellfish.^20,37 Human body burdens of BDE-209 also appear to be rising including among E-waste workers and people residing near PBDE production facilities^38,39 as well as among the general population, particularly young children in the US.^18,40 People residing in some regions of China with heavy E-waste recycling operations report some of the highest PBDE body burdens in the world, with BDE-209 concentrations in serum at greater than 3000 ng/g lipid. DecaBDE is now more commonly detected among the general population, particularly young children in the US population. For example, recent work in our laboratory,¹⁸ in collaboration with the Centers for Disease Control (CDC) and Boston University, measured BDE-209 levels in the serum of a North Carolina cohort of toddlers (3 y old) ranging from <6–68 ng/g lipid. In addition, other highly brominated PBDEs are being detected more frequently in biota, including the hexa- to nonaBDEs, which may reflect the increased use of BDE-209 and its environmental breakdown and biological metabolism.^17,41

This review summarizes the biological disposition and toxicity of PBDEs in teleost fishes with particular focus on thyroid disruption mechanisms and interactions as this endocrine system is an important target of the PBDEs. Indeed, most laboratory studies published to date in fish (and other vertebrates) have focused on the potential for PBDEs to perturb thyroid signaling, as well as impair neurological development and reproduction. While these are often the primary endpoints of focus, other toxicity outcomes have been observed in fish as well, including immunotoxicity^42,43 and oxidative stress.^44-46 The PentaBDE commercial mixture and its component congeners have been the subject of most study to date, with less known about the toxicity of BDE-209 and the other higher MW PBDE congeners. However, BDE-209 is the only PBDE that has been evaluated for carcinogenicity with “suggestive evidence of carcinogenic potential” based on increased thyroid cell hyperplasia and thyroid adenomas/carcinomas in male mice and liver tumors in male rats.⁴⁷

PBDE data generated in fish not only informs our understanding of potential effects in wild fish species and populations but also provides important information on mechanisms of toxicity and disease in humans due to the high degree of gene and functional conservation shared across vertebrates. Indeed, PBDE toxicity measured in fish shares mutual features and effects to those observed among in vivo and in vitro mammalian models, supporting potentially common biological mechanisms of toxicity that also lend biological plausibility to the human epidemiology data on PBDEs. Therefore, this review includes brief discussion of effects observed in rodent and human epidemiology studies in so far as to frame a fuller picture of the toxicity of these chemicals. PBDE human health effects have been examined in several informative reviews and readers are referred to these papers for more detailed analyses.^48-50

Toxicokinetics of PBDEs

Patterns of PBDE toxicokinetics (i.e., absorption, distribution, metabolism, and excretion; ADME) in fishes have been shown to vary depending on the PBDE congener, species, life-stage, and route of exposure. Table 1 summarizes PBDE toxicokinetic studies in fishes to date.

Table 1. PBDE toxicokinetics measured in teleost fish species

Species	Treatment	Route	Dose/Duration	Effects Observed	Ref
Common carp (Juv.; C. carpio)	BDE-99	GI, liver microsomes	12–29 pmol/mg protein; 60 min incub	Reductive debromination: BDE-99 to BDE-47           Metabolism in liver > intestine No debromination by GI microflora	^74
Atlantic salmon (Juv; S. salar)	PentaBDE; OctaBDE	Diet	10, 50 mg/kg bw; 7 d	No significant hepatic CYP1A induction or protein expression	^75
Chinook salmon (Adu.; O. tshawytscha)	BDE-99	Liver microsomes, cytosol	0.03 - 1.8 µM; 16 h incub	Reductive debromination: BDE-99 to BDE-49 negative GST/CDNB assay	^76
Northern Pike (E. Lucius)	PCBs, PCNs, BDE-47, -99, -153	Diet	90 ng/µl lipid (10 µl injected into rainbow trout); 9 d	Uptake efficiencies: BDE-47 = ~90%; BDE-99 = ~60%; BDE-153 = ~40%	^51
Northern Pike	^14C-BDE-47	Diet	16.2 µg/µl; 9, 18, 36, 65 d	^14C-BDE-47 uptake > 90% Disposition: Highest in liver, adipose tissue, spinal cord-surrounding tissue, eyes, gall bladder; Lowest in muscle, spleen, gills	^52
Crucian carp (C. auratus)	BDE-15	Aqueous	0, 10, 100 µg/l; 50 d	BDE-15 accumulation in gill, liver 2 mono-brominated, 3 hydroxy metabolites	^77
Zebrafish (Larv.)	BDE-209	Spiked sediment	12.5 mg/kg; 4 – 192 hpf	BDE-209 bioaccumulation	^6
Rainbow trout (Juv.; O. mykiss)	DecaBDE (Dow FR- 300BA)	Diet	7.5 - 10 mg/kg bw/day; 16, 49, 120 d; 71 d depuration	Reductive debromination: hexa- to nonaBDE formation (liver, muscle); BDE-154 dominant           BDE-209 uptake: 0.02–0.13% BDE-209 accumulation: 870 ± 219 ng/g ww (liver); 38 ± 14 ng/g ww (muscle)	^59
Zebrafish (Juv.; D. rerio)	PentaBDE (DE-71)	Aqueous	0, 0.1, 1 mg/l; 4 wk	AhR-mediated effects linked to PBDD/F impurities; Weak induction CYP1A; no DR-CALUX response (purified DE-71)	^78
Lake whitefish (Juv.; C. clupeaformis)	BDE-209	Diet	0, 0.1, 1, 2 µg/g; 30 d	Accumulation: BDE-209 + nonaBDEs (BDE-206, -207, -208) in liver	^79
Zebrafish (Adu.)	BDE-28, -183, -209 (mix)	Diet	1 and 100 nmol/g ww food at 2% bw/day; 42 d with 14 d dep	Reductive debromination (high dose); 12 nmol of BDE-154/g ww fish; 3 nmol of BDE-149/g ww fish; < 2 nmol of BDE-153 g ww fish; Uptake: BDE-28 (100%) > BDE-183 (10%) > BDE-209 (< 1%)	^57
Common sole (Juv.; S. solea L.)	BDE-28, -47, -99, -100, -153, -209 (mix)	Diet	82 - 184 ng/g ww food at 0.8% bw/day; 84 d, 149 d dep	Uptake efficiency: BDE-28 = 16%; BDE-47 = 15%; BDE-99 = 13%; BDE-100 = 14%; BDE-153 = 10%; BDE-209 = 1.4%. Reductive debromination: BDE-49; BDE-154; BDE-183; BDE-202; unk tetra-, penta-, hexaBDEs	^56
Common sole (Juv.)	BDE-28, -47, -99, -100, -153, -209 (mix)	Diet	82 - 184 ng/g ww food at 0.8% bw/day; 84 d, 149 d dep	Oxidative metabolism: 6-OH-BDE-47; 4’-OH-BDE-49; 4’-OH-BDE-101; 4’-OH-BDE-103 No MeO metabolites detected	^80
Common carp (Adu.)	BDE-99	Liver microsomes, cytosol	354 pmol; 1 - 250 µM; 90 min incub.	Reductive debromination (BDE-99 to BDE-47) more prevalent in liver microsomes than cytosol THs (rT3, T4) and iodoacetate inhibited debromination; role for deiodinase enzymes in reductive metabolism	^81
Fathead minnow (Juv., P. promelas)	BDE-209	Diet	10 µg/g food at 5% bw/day; 28 d	BDE-209 bioaccumulation           BDE-209 uptake efficiency = 5.8%           Reductive debromination to penta-octaBDEs           BDE-154 dominant metabolite BDE-101 lowest MW metabolite	^61
Fathead minnow (Adu.)	BDE-209	Diet	95 ng/g ww food and 10 µg/g ww food at 3% bw/day; 28 d with 14 d dep	BDE-209 bioaccumulation           Reductive debromination to penta – octaBDEs           BDE-154 dominant metabolite BDE-101 lowest MW metabolite	^62
Atlantic cod (Juv., G. morhua)	BDE-47	Aqueous	5 µg/l; 21 d	Liver: ↓ mRNA transcripts encoding CYP1A, CYP2C33-like, CYP3C1-like, UDPGT No effects on antioxidant genes (GSH-Px, GR)	^82
Rainbow trout           Common carp Chinook salmon (O. tschwatcha) (Juv.)	BDE-28, -47, -49, -99, -100, -153, -154, -183, -203, -208, -209	Liver microsomes	1 µM; 24 h (hepta to BDE-209); 1 h (tri- to hexaBDEs)	Reductive debromination of BDE-99, -153, -183, -203, -208, -209           Carp: meta-position debrom dominated           Salmonids: meta- and para-position debrom No metabolism of PBDEs lacking meta-substituted Br (BDE-28, -47, -100)	^83
Rainbow trout Common carp (Juv.)	BDE-209	Diet, in vitro	940 ng/g ww food, 1% bw/day; 5 mo; 15 pmol/mg protein; 1, 24 h (microsomes)	Reductive debromination (trout): Formation BDE-207, -208, -188, -201, -202, unk. octa-heptaBDEs           BDE-209 uptake (trout): 3.2%; liver > serum > intestine > carcass (lipid-normalized) In vitro: Formation of octa – nonaBDEs (trout), hexa - octaBDEs (carp)	^60
Common carp (Juv.)	BDE-209	Diet	940 ng/day-fish; 60 d w/40 d dep	Reductive debromination: Formation of BDE-154, BDE-155, unknown hexa- to octaBDEs No BDE-209 bioaccumulation	^25
Common carp (Juv.)	BDE-99, BDE-183	Diet	400 ng/day-fish (BDE-99); 100 ng/day-fish (BDE-183); 62 d w/37 d dep	Reductive debromination (GI tract):           BDE-99 → BDE-47           BDE-183 → BDE-154, unk hexaBDE Uptake: BDE-99 = 9.5%; BDE-183 = 17%	^58
Common carp (Juv.)	BDE-28, -47, -99, -153 (mix)	Diet	470 ng/day-fish; 60 d w/40 d dep	BDE-47 accumulation, high assimilation No BDE-99 bioaccumulation; No hydroxy metabolites detected	^84
Japanese medaka (Adu.; O. latipes)	6-OH-BDE-47, 6-MeO-BDE-47, BDE-47	Maternal	21, 8, 0.9 µg/g dw food at 2% bw/day; 14 d	No OH-, MeO-BDEs in BDE-47 treated fish;           In vivo and in vitro conversion of 6-OH-BDE-47 to 6-MeO-BDE-47 (and vice-versa) Maternal transfer to eggs	^85
Common carp (Juv.)	Penta and DecaBDE mixtures	Diet	100, 120, 150 µg/day/fish; 20 d	Reductive debromination facilitated by at least one meta- or para- doubly flanked Br 11 OH-BDEs measured in serum of pentaBDE exposed fish; No OH-BDEs in decaBDE exposed fish	^86
Chinese sturgeon (Adu; A. sinensis)	BDE-209	Field collected	Liver microsomes; PBPK modeling	Reductive debromination; Formation of BDE-126, -154, -188, -202, -204, -197; Low partition coefficients from blood to tissues lead to higher bioaccumulation of hepta to BDE-209 in absorbing tissues	^63

dep = depuration; EROD = ethoxyresorufin-O-deethylase; dpf = days post fertilization; dph = days post hatch; dio = deiodinase; DR-CALUX = chemical-activated luciferase gene expression mediated by Ah-receptor activation; GI = gastrointestinal; HDT = highest dose tested; hpf = hours post fertilization; PBPK = physiologically based pharmacokinetic; UDPGT = uridine diphosphate glucuronosyl phosphate.

PBDE Uptake and Tissue Distribution

Our understanding of PBDE absorption in fishes is somewhat limited by the species and PBDE congeners studied. Two early studies in Northern Pike (E. lucius) found that BDE-47 was readily absorbed with measured uptake efficiencies of ¹⁴C-BDE-47 and unlabeled BDE-47 at 90–100%.^51,52 This research group also measured uptake efficiencies of BDE-99 and BDE-153 in pike at ~60% and ~40%, respectively. Studies in rodents have likewise measured absorption of BDE-47, -99, -100, -153, and -154 in the range of ~70–90%.^53-55 However, other studies that have exposed fish to unlabeled PBDEs through the diet have measured lower assimilation efficiencies of some congeners, including BDE-99, BDE-153, BDE-183, and BDE-209 suggesting species-specific differences in assimilation efficiencies of PBDEs possibly due to differences in metabolic enzyme systems.^56-58 For instance, BDE-209 absorption in some teleost fishes has been shown to occur at a slow rate, which may allow for greater metabolism and elimination than seen in terrestrial species. A dietary study in juvenile rainbow trout receiving 7.5–10 mg/kg bw-day of BDE-209 measured bioavailability at <1%⁵⁹ with higher bioavailability of 3.2% measured in juveniles of the same species receiving a chronic dietary exposure.⁶⁰ In fathead minnow juveniles and adults exposed orally to ~10 µg/g ww food, BDE-209 uptake efficiencies were also low at 5.8%⁶¹ and 1.3%,⁶² respectively. Nonetheless, despite its low bioavailability, BDE-209 appears to be bioaccumulating in fish as shown in field measures and laboratory-based studies with BDE-209 spiked sediments.⁶ While the apparent dichotomy between low bioavailability of some PBDEs, notably BDE-209, and observed bioaccumulation are not fully described, rainbow trout exposed to BDE-209 thru the diet have been shown to bioaccumulate BDE-209 at 1.3 times the concentration of levels in their food, suggesting bioaccumulation that is strongly influenced by tissue composition.⁶⁰ In addition, recent findings in Chinese sturgeon (A. sinensis) also support that tissue distribution patterns of the higher PBDEs, rather than lipid binding, are important factors influencing their bioaccumulation.⁶³ This study showed low partitioning of the higher PBDEs (heptaBDEs to BDE-209) from blood to tissues that could in turn lead to their slower delivery to metabolically active tissues and thus higher bioaccumulation.

The dominant PBDEs measured in biota (i.e., BDE-47, -99, -100, -153, and -154) are deposited to lipophilic tissue compartments, and these congeners continue to be detected in human serum, breast tissue, and milk^18,64,65 and in adipose tissues of a variety of wildlife, including free ranging fish species.^17,66,67 PBDEs have been shown to cross the blood–placenta and blood-brain barriers to accumulate in the brains of perinatally exposed rats exposed to the PentaBDE commercial mixture⁶⁸ and some birds of prey.⁶⁹ In addition to accumulation in lipid-rich tissues, the liver is an important target of PBDE disposition and toxicity. The US EPA National Toxicology Program has published findings showing hepatotoxicity, including elevated liver enzyme activity accompanied by hepatic hypertrophy and vacuolizations in mice exposed orally to the PentaBDE mixture for 13 wk.⁷⁰ The liver also appears to be an important site of PBDE accumulation in fish. In pike fish, ¹⁴C-BDE-47 accumulated in the liver and in lipid rich tissues. In rainbow trout (O. mykiss) exposed orally to BDE-209, the highest concentration of BDE-209 was measured in the liver on both a lipid normalized and body weight basis, followed by accumulation in the serum, with less accumulation in the carcass.⁶⁰ BDE-209 deposition to adipose tissues might be predicted given its low water solubility (<0.1 µg/l) and high octanol-water partitioning coefficient (log K_ow; 6–12).⁷¹ However, studies have shown that this pattern of preferential deposition to lipid depots does not occur substantially for BDE-209. Rather, BDE-209 preferentially distributes to highly perfused, blood-rich tissues, particularly the liver, kidney, heart, and intestinal wall.^25,72,73 The underlying reasons for this distribution pattern appear to be attributable to the large size of BDE-209 and its ability to bind with plasma proteins.⁷²

PBDE Reductive Metabolism in Fish

To understand PBDE metabolic pathways in fish, it is informative to frame the discussion in terms of present knowledge of PBDE metabolism in mammals as this is now fairly well described. The rodent literature supports a PBDE metabolic pathway in mammals that has two major reactions: (1) a cytochrome P450 (CYP450)-mediated epoxidation of PBDE phenyl rings catalyzed predominantly by CYP2B (by constitutive androstane receptor [CAR] inductions) and by CYP3A (by pregnane X receptor [PXR] inductions); and (2) debromination or Phase II conjugation of an OH-intermediate with glucuronides catalyzed by uridine diphosphate glucuronosyl transferases (UDPGTs) and with sulfates by sulfotransferases (SULTs).^87,88 Reductive debromination reactions appear to be minor pathways of PBDE metabolism in mammals (e.g., BDE-209 debromination to octa- and nonaBDEs).

PBDE metabolism in teleost fish appears to be different from that in mammals. As outlined in Table 1, a large number of studies have shown reductive debromination of PBDEs to be a major route of metabolism, including in common carp (C. carpio),²⁵ fathead minnow (P. promelas)^61,62; rainbow trout,^59,60 lake trout (S. namaycush),⁸⁹ Chinook salmon (O. tshawytscha),^76,83 and zebrafish (D. rerio).⁷⁸ However, while PBDE reductive debromination appears to be a major metabolic pathway in fish, the role of specific enzyme systems in catalyzing this biotransformation remains unclear. One pathway that has been hypothesized to mediate PBDE reductive metabolism in fish is by the activity of iodothyronine deiodinase (Dio) enzymes.^25,89 Dios are membrane-bound enzymes that are expressed on plasma membranes and in the endoplasmic reticulum, and regulate TH levels in vertebrates.⁹⁰ There are three known Dio isoforms in fish, Types 1, 2, and 3 (Dio 1, Dio 2, and Dio 3, respectively) that share functional homology with mammalian Dio isoforms. As illustrated in Figure 2, the conversion of the TH thyroxine (T4) to the genomically active 3,3′,5-triiodothyronine (T3) hormone is catalyzed by the cleavage of iodine from the meta-position of the outer phenyl ring of T4. The reductive debromination of PBDEs in fishes is also dominated by meta-cleavages of bromine, suggesting a possible role for these enzymes in catalyzing PBDE debromination.^25,89 More recent studies have shown that the reductive debromination of BDE-99 to BDE-47 can be substantially inhibited by co-incubating liver microsomes from common carp with THs, suggesting that BDE-99 may be a substrate that competes with THs for Dio enzyme activity.^74,81

Figure 2. Outer-ring deiodination of thyroxine (A) catalyzed by Dio 1 and 2 enzymes resulting in meta-cleavage of iodine from T4. Reductive metabolites and meta-position debromination measured in common carp (C. carpio) exposed orally to (B) BDE-99, (C) BDE-183, and (D) BDE-209.^25,58,60

Although BDE-glutathione metabolites have been measured in rodents^53,91 and birds,⁹² glutathione-S-transferases (GSTs) have not been found to be involved in the reductive debromination of BDE-99 to BDE-49 in Chinook salmon⁷⁶ or of BDE-99 to BDE-47 in common carp,^74,81 suggesting that they may not play an important role in PBDE debromination in fish. This finding has been subsequently confirmed in negative results from in vitro testing with liver microsomes from Chinook salmon, rainbow trout, and common carp incubated with several PBDE congeners⁸³ and in juvenile fathead minnows exposed in vivo to BDE-209.⁶¹ Taken together, despite these lines of indirect evidence, additional work is needed to better understand the underlying enzymes catalyzing the reductive debromination of PBDEs in fishes, and the potential role of Dio enzymes and/or other possible reductases that have yet to be described.

Mixed Evidence of PBDE Oxidative Metabolism in Fish

Laboratory studies in fish have presented mixed results concerning the formation of hydroxylated PBDE (OH-BDE) metabolites. For instance, no OH-BDEs were detected in the serum of common carp receiving dietary exposures to a mixture of BDE-28, -47, -99, and -153.⁸⁴ Likewise, OH-BDEs were not detected in liver microsomes from common carp, rainbow trout, or Chinook salmon incubated with various PBDEs, including BDE-209.^76,83 In contrast, 11 OH-BDEs were reported in the serum of juvenile common carp exposed to the PentaBDE mixture with no OH-BDEs in DecaBDE exposed fish.⁸⁶ Another BDE-209 study, however, reported substantial OH-BDE and methoxy PBDE (MeO-BDE) formation in trout.⁹³ OH-BDE metabolites were also reported in juvenile common sole (S. solea L.) exposed to a mixture of congeners (BDE-28, -47, -99, -100, -153, -209).⁸⁰ In studies showing positive OH-BDE results, the gas chromatography-mass spectrometry (GC/MS) injection techniques used in the PBDE analyses were either not specified⁸⁶ or employed split/splitless injection.^80,93 GC/MS splitless injection techniques for PBDE analysis can lead to thermal degradation of parent PBDEs and the formation of byproducts in the GC/MS inlet that may confound identification of MeO-BDEs (GC/MS derivatives of OH-BDE metabolites).⁹⁴ Although potential analytical confounders were not addressed, if oxidative metabolism occurs in fish, it appears to be a minor metabolic pathway compared with reductive debromination.

In vivo and in vitro studies in fish have shown both weak induction^78,95 and inhibition^82,96 of ethoxyresorufin-O-deethylase (EROD) activity (biomarker of CYP1A and aryl hydrocarbon (AhR) induction) upon exposure to individual PBDE congeners. However, a larger number of PBDE studies in teleost fish have shown no AhR/CYP1A activation.^75,89,97,98 Thus, it appears that PBDEs operate predominantly through non-dioxin, AhR independent toxicity mechanisms and are not metabolized by CYP1A. Conversely, the PentaBDE commercial mixture contains small amounts of polybrominated dibenzo-p-dioxins/dibenzofurans (PBDDs/PBDFs) that are trace byproducts formed by thermal stress during production and that activate the AhR. The reasons for the disparate results in the literature are not clear but may be attributable to variations in the purity of formulations tested.

While mammalian studies support PBDE oxidative metabolism by CYP2B and CYP3A catalyzed by CAR/PXR inductions, similar pathways have not been observed in fishes. The expression of CYP2B in fish and the regulatory mechanisms involved in its induction are still unclear. In mammals, phenobarbital (PB) and ortho-substituted halogenated aromatics (e.g., ortho-chlorine-substituted PCBs) are strong inducers of CYP2B through activation of CAR.⁹⁹ In teleost fish, however, the induction of CYP2B in the presence of PB-type inducers has not been observed.¹⁰⁰ Thus, there appear to be important functional differences between piscovorous and mammalian CAR/CYP2B that may play a role in its lack of induction in PBDE-exposed fish, although this has not been examined. With regard to other AhR-independent mechanisms, the function and tissue distribution of enzymes in the CYP3 gene family are not well characterized in fish, but the CYP3A isoforms appear highly versatile with broad substrate affinities.¹⁰¹

PBDE Conjugation in Fish

There has been little research to date to examine the role of UDPGTs and SULTS in the metabolism of PBDEs in fishes, although these enzymes are important catalytic drivers of Phase II metabolism in fishes.¹⁰⁰ The UDPGTs catalyze the glucuronidation of an array of endogenous and exogenous substrates to more polar, water-soluble compounds for elimination. The SULTs catalyze the transfer of the sulfonate from 3′-phosphoadenosine-5′-phosphosulfate (PAPS) to hydroxylated and amine substituents on numerous exogenous and endogenous substrates to facilitate elimination. In zebrafish, several UDPGT¹⁰² and SULT genes^103,104 have been characterized with prototypical substrates such as bilirubin, TH, estradiol (E2), testosterone (T), and phenolic contaminants. As many as 10 different UDPGT isoforms have been identified in zebrafish, with nucleotide similarities to some mammalian UGT1 and UGT2 gene families.¹⁰⁵ Two studies have shown a decrease in the relative mRNA abundances of genes encoding UDPGT1ab in zebrafish larvae exposed to BDE-209¹⁰⁶ and UGT1 in juvenile Atlantic cod (G. morhua) exposed to BDE-47.⁸² This decline may be a response to reduced TH levels as UDPGTs are involved in the metabolism of THs.

PBDE Elimination

Table 2 provides whole body elimination half-lives (t_1/2) reported in fish for PBDEs frequently detected in the environment and biota, with inclusion of data in humans and rodents for comparison. In rodents, the major route of PBDE elimination is by the fecal route with low levels of excretion in the urine and bile depending on the PBDE congener.^88,107 In fish, routes of PBDE elimination have not been targeted specifically but the early studies in pike suggest that biliary and fecal excretion also occurs.⁵² Some reports estimate apparent half-lives in humans for the tetra - hexaBDEs that are substantially longer than those reported in rodents and some fish. Conversely, some data in rodents suggest relatively short half-lives (e.g., BDE-99) that are incongruous with the human and fish data. Thus, there continues to be uncertainty about PBDE elimination half-lives in fish and other biota with substantial species variability apparent that appears related to differential metabolism.

Table 2. Whole-body elimination half-lives (t_1/2) of environmentally relevant PBDEs measured and estimated in humans, laboratory rodents, and teleost fish species

PBDE Congener	Teleost Fish half-life (days)	Humans half-life	Rodents half-life (days)
BDE-47	18–42 (Common carp)^84 173–519 (Lake trout)^89 12 (Japanese medaka)^108	1.5–2.5 y^109	25c (mouse)^55 19–30e (rat, tetraBDE)^110
BDE-99	173–519 (Lake trout)^89	1.8–3.4 y^109	6d (rat)^91
BDE-100	46–80 (Lake trout)^89	1.3–1.8 y^109	42–52e (rat, pentaBDE)^110
BDE-153	4–23 (Common carp)^84 154–308 (Lake trout)^89	3.6–12.4 y^109	50–105e (rat, hexaBDE)^110
BDE-154	17–53 (Common carp)^84 70–208 (Lake trout)^89	2.3–4.3 y^109	Unknown
BDE-183	173–519 (Lake trout)^89 10–15 (Zebrafish)^57	94 d^111	Unknown
BDE-209	21–34 (Lake trout)^89 6.5 (Zebrafish)^57	15 d^111	2.5, 8.6 (rat)^112^,^113

Fish Thyroid System

To discuss PBDE-related TH disruption, it is informative to broadly highlight current knowledge of the structure and function of the fish thyroid with some comparison to the mammalian thyroid. The vertebrate thyroid is well-conserved across taxa, and chemical effects in lower level vertebrates like fish can reveal mechanisms of thyroid dysregulation in higher level species. THs are key regulators of vertebrate development, endothermic basal metabolism, and organ system physiology. The importance of TH in brain and somatic development is well established, and small changes in maternal or fetal TH can cause severe motor skill deficiencies and irreversible cognitive impairments.¹¹⁴ Recent attention has focused on the permissive role of THs in regulating physiological processes in adults, including neurological plasticity, mood, cognition, and reproduction.^90,115 In fish, THs are important mediators of many physiological, developmental, and behavioral processes, including growth and metamorphic transitioning,^116,117 osmoregulation,¹¹⁸ olfactory imprinting,¹¹⁹ interrenal regulation,¹²⁰ otolith formation,¹²¹ and reproduction,^90,122 often acting in concert with other hormones.^123,124

Structure and Function of Fish Thyroid

The general architecture of the thyroid appears to be similar across vertebrates whereby circulating levels of TH are tightly controlled by both a centrally operating hypothalamic-pituitary-thyroid (HPT) axis and in peripheral tissues through the activity of Dio enzymes, among other dynamically operating regulatory processes (Fig. 3). The functional unit of the central HPT is the thyroid follicle where the THs T4 and 3,5,3′-triiodothyronine (T3) are synthesized and secreted into circulation. However, there continue to be gaps in our understanding of the fish thyroid in comparison to the more thoroughly studied mammalian and amphibian thyroid systems. Typically in teleost fish, thyroid follicles are found dispersed predominantly in the ventral pharyngeal region, rather than being organized in a compact lobular gland as seen in higher vertebrates. One important differentiating feature of the fish thyroid may relate to how THs are produced and regulated. In fish, T4 may be the primary, possibly only TH produced in the thyroid where it is under negative feedback control by the HPT axis. The production of T3 in fish, in contrast, is thought to be under exclusive control of peripheral tissues, although this has not been studied recently or beyond Salmonid fishes.¹²⁵ In some contrast, the mammalian thyroid gland produces both T4 and to a lesser extent T3 under negative feedback control by the central HPT. Thus, while extrathyroidal regulation of T3 is important in all vertebrates, including for instance in the brains of developing mammals, in fish the formation and regulation of T3 may be dominated by local control in response to the needs of individual tissues rather than by T4 availability and through the central HPT axis. THs circulate in plasma bound to TH binding proteins, including thyroid binding globulin (TBG), transthyretin (TTR), and albumin. In humans, the primary transporter of TH is TBG while in rodents it is TTR.¹²⁶ Less is known about the dominant transporters in fish although it has recently been shown that TTR may bind TH in some species.^127,128 Most TH in fish circulation is bound to protein with only a small amount (< 1%) thought to be free and available for uptake into cells. As demonstrated in rodents, the cellular transport of THs in fish is mediated largely by membrane bound transporters, including the high affinity monocarboxylate transporter 8 (MCT8) and organic anion transporter polypeptides (e.g., OATP1c1), among others.^129-132

Figure 3. Overview of TH regulation and signaling in teleost fishes. TRH = thyrotropin releasing hormone; TSH = thyroid stimulating hormone; T4 = thyroxine; T3 = 3,3′,5-triiodothyronine; rT3 = 3,3′,5′-reverse T3; T2 = 3,3′-diiodothyronine; UGT = uridine diphosphate glucuronosyl transferase; SULT = sulfotransferase; TH-G = glucuronidated thyroid hormone; TH-S = sulfated thyroid hormone; Mrp = multidrug resistance associated protein; Mdr1 = multidrug resistance protein 1 or P-glycoproteins; MCT = monocarboxylate transporter; OATP = organic anion transport polypeptide; TR = thyroid hormone receptor; RXR = retinoic x receptor.

Peripheral TH Regulation and Signaling in Fish

Once in the cell, T4 can be deiodinated to the active T3 hormone or inactivated to 3,3′,5′-triiodothyronine (rT3) or 3,3′-diiodothyronine (T2) (Fig. 4). Dio 1 and Dio 2 enzymes in vertebrates may catalyze T4-outer ring deiodination (ORD) to produce the active T3 hormone, while Dio 1 and Dio 3 catalyze T4-inner ring deiodination (IRD) to inactive rT3. Thus, Dio 1 can be involved in both ORD and IRD. In addition, T3-IRD and rT3-ORD can metabolize hormone to T2. THs in vertebrates are further conjugated in the liver to glucuronides or sulfates catalyzed by UDPGTs and SULTs, respectively, and excreted through the bile and urine. The transcriptional activity of T3 is mediated by complexing with nuclear thyroid receptors (TRs) that bind to TH response elements (TREs) to induce the expression of TH responsive genes.¹³³ Increased attention is also focusing on nongenomic pathways of THs signaling by integrin mediated signaling and kinase activation.^134,135

Figure 4. Pathways of outer and inner ring deiodination of thyroid hormones in extrathyroidal peripheral tissues of vertebrates.

While Dio enzymes in fish and mammals are believed to share many functional features, relative tissue distributions and activity vary, which may have implications for their role in TH regulation.^136,137 For instance, Dio 1 has been localized to the kidneys and liver of fish and mammals (as well as in the mammalian thyroid gland).¹³⁸ However, hepatic T4-ORD activity is thought to be catalyzed mostly by Dio 1 in mammals and Dio 2 in fish.^136,139 One notable exception to this peripheral TH production may occur in the teleost fish brain with its relatively low T4-ORD activity. Early studies that measured deiodination in rainbow trout attributed reduced brain T4-ORD in fish to absent or negligible Dio 2 expression.¹⁴⁰ However, more recent studies using quantitative PCR have localized mRNA expression of dio2 genes in the fish brain, although at comparatively lower levels than in other tissues.^62,141,142 Limited evidence also suggests that the transcriptional response of Dio 2 in the fish brain may be more sensitive to systemic TH changes than Dio 2 in the liver.¹⁴¹ Likewise, in mammals, Dio 2 has demonstrated substantial physiological plasticity in the brain with a short half-life of ∼40 min suggesting that it might also be an important regulator of intracellular T3 in these tissues.¹⁴³

PBDE Thyroid Disruption in Fish

Because PBDEs, particularly OH-BDE metabolites, are similar in structure to THs, concerns have been raised about their effects on thyroid system functioning in both mammalian and non-mammalian vertebrates.⁵⁰ In fish, PBDEs have been shown to perturb the thyroid system at several points along the central HPT and in extrathyroidal tissues with most study to date focused on the PentaBDE mixture, BDE-47, and BDE-209. Many of these laboratory studies in fish have shown declines in circulating levels of THs and altered TH signaling in some species exposed to these PBDE congeners and mixtures. Table 3 provides a summary of studies in fish conducted to date. Similar to results in fish, in vivo rodent studies and cell-based assays have reported PBDE-induced disruption of thyroid homeostasis, including declines in circulating THs^144-146; altered expression and activity of TH metabolizing enzymes,^147,148 and competitive binding with plasma transporters.¹⁴⁹ Human epidemiology studies have shown associations between altered plasma concentrations of THs in adults and PBDE levels in serum and dust.^150-152

Table 3. PBDE effects on thyroid endocrine systems of teleost fish

Species	Treatment	Route	Dose/Duration	Effects Observed	Ref
Zebrafish (Larv.)	BDE-209	Aqueous	0, 0.08, 0.38, 1.92 mg/l; 14 dpf	↑ mRNA transcripts for CRH, TSHβ, Pax8, Nkx2.1, NIS, Tg, dio1, dio2, trα, trβ ↓ TTR mRNA transcripts; ↑ T3 (0.38, 1.92 mg/l), ↓T4 (1.92 mg/l)	^106
Zebrafish (Juv.)	BDE-47	Diet	100 ng/g; 0.5–1 mg food/fish-d; 20–60 dpf	No change in TTR, dio1, TSHβ mRNA transcripts	^95
Zebrafish (Larv.)	6-OH-BDE-47	Aqueous/WISH	1, 10, 100 nM; 4 – 22 hpf	↑ dio1 mRNA in brain periventricular zone; ↑ dio 3 in pronephric duct	^153
Rainbow trout (O. mykiss Juv.)	BDE-209	IP Injection	50 – 1000 ng/g bw/day; 21 d	↑ TT4 (1000 ng/g bw); ↓ FT3 (all doses); ↓ FT4 (100, 200, 500 ng/g bw)	^93
European flounder (P. flesus Adu.); Zebrafish (Juv.)	PentaBDE (DE-71 purified of PBDD/Fs)	Spiked sediments, Diet	0+0.014 -700+14000 µg/g TOC + µg/g lipid; 101 d (flounder) 0–500 µg/l; 30 d (zebrafish)	↓ plasma TT4 (flounder)           ↑ plasma TH, ↓ larval survival; (zebrafish)	^96
Fathead minnow (Adu. breeding pairs)	BDE-47	Diet	2.4 µg/pair/ day; 12.3 µg/pair/day; 21 d	↓ TT4 (no change in TT3); ↑ mRNA transcripts for TSHβ (low dose), trα (female brains); ↓ trβ mRNA transcripts (both sexes, brains)	^154
Chinese rare minnow (G. rarus; Adu., Larv.)	BDE-209	Aqueous	0 – 10 µg/l; 21 d	↑ dio2, NIS mRNA transcripts (larvae) ↓trα, dio2, NIS mRNA transcripts (adults)	^155
Gilthead sea bream (S. aurata; Adu.)	PBDEs, 6-OH-BDE47	In vitro	0–10 µM; 2 h	BDE-28, 49, -47, -99 potent inhibitors of ^125I-T3 binding to TTR; IC50s < < T3, T4 6-OH-BDE-47 moderate inhibitor; IC50s > T3,T4	^156
Fathead minnow (Juv.)	BDE-209	Diet	10 µg/g food at 5% bw/day; 28 d with 14 d dep	↓ Dio activity (T4-ORD and T4-IRD) by 74%           Thyroid histology: Over-stimulation, injury (thickened follicular epithelium, irregular follicle outlines, colloid depletions, ↑ inflammatory cells) Liver alterations: vacuolated hepatocytes	^61
Fathead minnow (Adu. males)	BDE-209	Diet	95 ng/g ww food and 10 µg/g ww food at 3% bw/day; 28 d with 14 d dep	↓ TT4 (53%) and TT3 (46%) at low dose           ↓ TT4 (59%) and TT3 (62%) at high dose           ↓ brain Dio activity ~65% ↑ mRNA transcripts for dio1, dio2, trα, trβ (↓ in liver), mct8, oatp1c1, oatp2a1(↓ in liver) in brain and liver	^62
Lake trout (S. namaycush; Juv.)	13 PBDE congener mix	Diet	0, 2.5, 25 ng/g dw food at 1.5% bw/day; 56 d; 112 d depuration	↓ FT4 (low, high dose); ↓FT3 (low dose only)	^89
Zebrafish (Larv.)	PentaBDE (DE-71)	Aqueous	1, 3, 10 µg/l; 14 d	↓ whole fish T4 (T3 not measured) ↑ mRNA transcripts for CRH, TSHβ, NIS, Tg, Pax8, Nkx2.1, dio1, dio2; ↓TTR mRNA transcripts	^157
Zebrafish (Adu., offspring)	PentaBDE (DE-71)	Aqueous	1, 3, 10 µg/l (parents and offspring) + (parent alone); 5 min to sexual maturation (parents), 5 and 10 dph (offspring)	↑ plasma TT4 (no ∆ TT3) (parents, ELISA);           ↓ mRNA transcripts for CRH, TSHβ (parent brain) ↑ whole fish T4, T3; altered HPT axis mRNA transcripts (offspring w/w/o DE-71 exposure)	^158

BTEB = basic transcription element-binding protein; CRBP = cellular retinal binding protein; CRH = corticotrophin releasing hormone; CSF = cerebral spinal fluid; dep = depuration; dpf = days post fertilization; dph = days post hatch; dio = iodothyronine deiodinase; FT3 = free triiodothyronine; FT4 = free thyroxine; HDT = highest dose tested; hpf = hours post fertilization; HPT = hypothalamic-pituitary-thyroid; MBP = myelin basic protein; PBDD/f = polybrominated dibenzo-p-dioxins/dibenzofurans; NIS = sodium/iodide symporter; Tg = thyroglobulin; TH = thyroid hormone; TOC = total organic carbon; TR = thyroid receptor; TSH = thyroid stimulating hormone; TT4 = Total thyroxine; TTR = transthyretin; WISH = whole mount in situ hybridization.

BDE-47, PentaBDE, and OH-BDE Thyroid Dysfunction

In one of the earlier studies in fish, depressed levels of circulating free T4 and T3 were measured in the plasma of lake trout (S. namaycush) dosed through the diet with a mixture of 13 PBDE congeners (e.g., BDE-28, -47, -99, -100, -153, -154, -209) at 2.5 ng/g and 25 ng/g per congener for 56 d.⁸⁹ In fathead minnows, dietary exposures to BDE-47 at 2.4 and 12.3 µg/breeding pair-day for 21 d elicited depressed total T4 (TT4) but not total T3 (TT3) that was accompanied by elevated mRNA transcripts encoding TSHβ in the pituitary of low dose fish.¹⁵⁴ In addition, elevated and reduced levels of mRNA transcripts encoding TRα and TRβ, respectively, were detected in the brain but not the liver, suggesting that the adult fish brain may be a sensitive target of BDE-47.¹⁵⁴ These results in minnows and trout are consistent with observations in European flounder (P. flesus) whereby declines in circulating TT4 with no change in TT3 were detected after a 101 d dietary exposure to the PentaBDE commercial mixture purified to remove polybrominated dibenzo-p-dioxins/dibenzofurans (PBDDs/Fs).⁹⁶

Reductions in whole fish T4 have also been measured in zebrafish larvae subjected to waterborne exposures of the PentaBDE commercial mixture.¹⁵⁷ In some divergence, however, this same research group measured elevated whole fish T4 and T3 in zebrafish offspring exposed to PentaBDE at the same dose as their previous study (Yu et al., 2010) but by a different exposure route (maternal) and longer duration (5 mo). The opposing effects of PentaBDE on TH levels that were measured in the Yu et al. (2010–11) zebrafish work demonstrate potentially important differences in PBDE impacts on vertebrate thyroid signaling that may be influenced by the pathway and duration of exposure (e.g., aqueous, short-term vs. maternal, chronic). BDE-209 and the PentaBDE mixture also have been found to enhance the relative mRNA expression of genes encoding dio1 and dio2 in zebrafish larvaeas well as thyroidal genes encoding the sodium iodide symporter (NIS), thyroglobulin (TG) and other transcription factors regulating NIS and TG expression (Nkx2.1a, Pax8).^106,157

There has been only limited thyroid-related study in fish to date focused on the OH-BDEs. Recent work in our laboratory using whole mount in situ hybridization (WISH) measured significantly upregulated mRNA expression of dio1, but not dio2 or dio3, in the periventricular brains of 22 hpf embryos exposed to the hydroxylated metabolite 6-OH-BDE-47. An increase in dio3 mRNA expression was also detected in the pronephric duct, which is the earliest form of the kidney in vertebrates and constitutes the central component of the excretory system. Thus, this study demonstrated that effects of 6-OH-BDE-47 on the developing zebrafish thyroid may elicit localized and age-specific transcriptional responses that then potentially contribute to downstream effects on neurological, renal, and reproductive development. To better understand the implications of these findings, additional data are needed to clarify the cellular and tissue distribution of Dios during ontogeny of the fish brain. In the mammalian brain, Dio 2 is expressed in glial cells (astrocytes, tanycytes), which could play a role in transporting and maintaining T3 supplies to neurons.¹⁵⁹

BDE-209 Thyroid Dysfunction

BDE-209 reduced circulating levels of free T4 and T3 in early life-stages of rainbow trout⁹³ with declines in whole fish T4 also reported in zebrafish larvae exposed aqueously to BDE-209.¹⁰⁶ In addition, an increase in dio2 mRNA transcripts have been measured in the larvae of Chinese rare minnows exposed to BDE-209 with a decrease in dio2 transcripts measured in the brains of adults minnows.¹⁵⁵ Recently published data of ours has shown TH regulation and signaling in juvenile and adult fathead minnows to be disrupted by low dose exposures to BDE-209. Specifically, in juvenile fathead minnows, compared with vehicle controls, the activity of Dio enzymes (T4-ORD and T4-IRD) declined by ~74% in fish dosed with 9.8 µg/g ww food at 3% bw/day for 28 d.⁶¹ This extrathyroidal perturbation was accompanied by evidence of thyroid follicle hypertrophy indicative of over-stimulation and injury. In adult male fathead minnows, BDE-209 caused a > 53% and > 46% decline in circulating total T4 and T3, respectively, upon a 28-d exposure to low doses of BDE-209 at ~3 and 300 ng/g bw-day.⁶² Depressed levels of circulating THs were accompanied by a 65% decline in Dio activity (T4-ORD) in the brains of treated fish at both BDE-209 doses tested. This hypothyroid response in BDE-209 exposed minnows was accompanied by possible localized compensatory signaling, including increased T4-ORD activity in the liver and transient, tissue-specific upregulation of genes encoding several important thyroidal proteins (Table 4). However, similar to results in minnows exposed to BDE-47,¹⁵⁴ this study suggested that the fish brain may be particularly sensitive to BDE-209 based on severe reductions in brain Dio activity (T4-ORD) and potentially muted adaptive responses of the brain to reduced TH levels. Consistent with observations in the brains of adult fish, data collected in developing rodents suggest weak adaptive responses of the brains of younger mammals to TH insufficiency caused by low level chemical exposures.^160-162 Additional work is needed to better understand how tissues, especially the brains, of developing and adult animals adapt or not to contaminant-induced TH insufficiency and whether this ameliorates downstream apical endpoints.

Table 4. Relative expression of genes encoding deiodinase (dio) enzymes, nuclear thyroid receptors (tr), monocarboxylate transporters (mct), and organic anion transporter polypeptides (oatp) in brains and livers of adult male fathead minnows exposed orally to BDE-209 and the positive control 6-propyl-2-thiouracil (PTU) for 28 d with a 14-d depuration.⁶²*

Gene target	Treatment	Day 14		Day 28		Day 42
		Brain	Liver	Brain	Liver	Brain	Liver
dio1	BDE-209 Low Dose	—	—	↑	—	—	↑
	BDE-209 High Dose	—	—	↑	—	—	—
	PTU Pos Ctrl	—	—	—	—	—	—
dio2	BDE-209 Low Dose	↑	↑↑	—	—	—	↑
	BDE-209 High Dose	↑↑	↑	—	—	—	—
	PTU Pos Ctrl	—	—	—	—	—	—
trα	BDE-209 Low Dose	—	—	—	—	↑	—
	BDE-209 High Dose	↑↑↑	—	↑↑	—	—	—
	PTU Pos Ctrl	—	—	—	—	—	—
trβ	BDE-209 Low Dose	—	↑	—	—	—	—
	BDE-209 High Dose	—	—	—	↓	—	—
	PTU Pos Ctrl	—	↑	—	—	—	—
mct8	BDE-209 Low Dose	↑	↑	—	—	↑↑	—
	BDE-209 High Dose	↑↑↑	—	—	—	—	—
	PTU Pos Ctrl	↑	—	—	—	—	—
oatp1c1	BDE-209 Low Dose	—	↑	—	—	—	—
	BDE-209 High Dose	—	↑	—	—	—	—
	PTU Pos Ctrl	—	↑	—	—	—	—
oatp2a1	BDE-209 Low Dose	—	—	—	—	—	—
	BDE-209 High Dose	↑↑	↓	—	—	—	—
	PTU Pos Ctrl	—	—	—	—	—	—

* Relative mRNA transcript abundances of genes not affected by BDE-209 or PTU: dio3, mct10, oatp1f1, oatp1f2, oatp2b1,oatp3a1, oatp4a, and oatp5a1. Statistical significance evaluated within sampling day with a one-way ANOVA and Tukey’s test; one arrow = P < 0.05; two arrows = P < 0.01; three arrows = P < 0.005.

Mechanisms of PBDE Thyroid Disruption

The vertebrate thyroid system maintains normal physiological functioning by responding to endogenous and exogenous perturbations with changes in TH production from the thyroid and through changes in the capacity and sensitivity of peripheral tissues. Such integrated compensatory responses at the central HPT and in peripheral target tissues make it challenging to evaluate mechanisms of action for thyroid disruptors like PBDEs. Nonetheless, several mechanisms appear to play a role in the thyroid perturbations measured in fish (and other vertebrates) exposed to PBDEs including: interference with Dio enzyme activity/expression; enhanced metabolism and elimination of THs; altered expression and activity of plasma transporters and membrane bound transporters; and altered genomic signaling.

Interference with Dio Enzymes

One mechanism that might be contributing to thyroid perturbations in fish and other vertebrates could involve PBDEs interfering with the expression and activity of Dio enzymes. PBDEs may be acting as competitive substrates for Dio enzymes or otherwise altering the expression and activity of these enzymes. As discussed, altered mRNA expression and enzymatic activity of some Dio isoforms has been observed in fish exposed to BDE-209^61,62,155 and 6-OH-BDE-47,¹⁵³ as well as in rodents exposed to PentaBDE¹⁴⁸ and human microsomes incubated with 5′-OH-BDE-99 and 2,4,6-TBP.¹⁴⁷ However, it remains unclear whether PBDEs (or OH-BDEs) can bind directly to Dio enzymes or whether they may elicit other allosteric effects that affect the capacity of Dios to mediate TH regulation.

Induction of TH Metabolizing Enzymes

Another hypothesis for PBDE-induced thyroid disruption is that PBDE detoxification responses may induce the expression and/or activity of TH catabolizing enzymes. In some support of this hypothesis, studies have measured increased mRNA expression of TH-conjugating UDPGT and SULT enzymes in rodents exposed to BDE-47¹⁴⁵ and the PentaBDE commercial mixture.^148,163 In these studies, declines in circulating levels of T4 from PBDEs were linked to enhanced glucuronidation associated with UDPGT inductions.^148,163,164 Other studies, however, have shown little to no change in UDPGT levels in rodents following exposure to PBDEs despite decreased T4 levels in circulation,^145,165 although increased mRNA expression of UDPGT transcripts has been observed in some of this work.¹⁴⁵ In contrast to the rodent data, limited evidence in fish has shown declines in mRNA transcript abundances of some UDPGT isoforms suggesting that PBDEs may be acting as TH mimics that then downregulate the expression of these TH metabolizing enzymes in PBDE exposed fish.^82,106

Altered Expression/Activity of Plasma and Cellular Transporters

Few studies have explored the role of plasma and membrane bound transporters in PBDE metabolic detoxification pathways and in contributing to or ameliorating PBDE effects on TH signaling.¹⁶⁶ OH-BDE metabolites produced in rat liver microsomes enriched with CYP2b (i.e., PB-induced) have been found to compete with THs for binding to the plasma transport protein TTR, potentially leading to greater elimination of TH and hypothyroidism.^149,167 Likewise in a recombinant sea bream TTR assay, several parent PBDEs (BDE-28, -49, -47, -99) were shown to be potent inhibitors of T3 binding to TTR, suggesting competitive interferences, while 6-OH-BDE-47 had less affinity for sea bream TTR than T3 or T4.¹⁵⁶ Other studies have also used biosensor screening methods to show that the OH-BDEs may bind to TTR and TBG with high potency.¹⁶⁶ More recently, newly designed fluorescent probes and competitive binding assays have shown that the binding of OH-BDEs with TBG and TTR increases with bromine number and OH position (i.e., 3-meta OH).¹⁶⁸

In addition to plasma transport, a few studies have explored PBDE effects on cell membrane bound transporters of TH. Our recently published data in fathead minnows measured upregulated mRNA transcripts encoding mct8 and oatp1c1 in the brains and livers of fish exposed to BDE-209 (Table 4).⁶² As observed with dio transcription, upregulated mRNA expression patterns of these transporters in minnows exposed to BDE-209 may be indicative of additional compensatory responses to hypothyroidism as mct8 and oatp1c1 are specific and active transporter of THs in fish^129,169 and mammals.^170,171 Only a limited number of the OATP transporters have been characterized in vertebrates with recent work in zebrafish¹³¹ and fathead minnows¹⁶⁹ to clarify their tissue distribution and function. Some OATPs have been found to be potentially important PBDE transporters. BDE-47, -99, and -153 have been shown to be effective substrates for human OATP1B1, OATP1B3, and OATP2B1 expressed in Chinese hamster ovary (CHO) cells.¹⁷² OATP1B1 and OATP1B3 were found to transport BDE-47 with the highest affinity, while OATP2B1 was found to transport all three tested congeners with similar affinities. Using human embryonic kidney cells transiently expressing mouse hepatic OATPs, this same research group also measured that OATP1a4, OATP1b2, and OATP2b1 were able to bind and transport BDE-47, -99, and -153.¹⁷³ Consistent with these results, upregulated mRNA expression of genes encoding OATP1a4, which transports THs, bile acids, and xenobiotics, was also detected in young rats exposed to PentaBDE.¹⁴⁸ PBDEs have also been shown to affect the Phase III hepatic efflux transporters, P-glycoproteins (i.e., Pgp; Mdr1) and multidrug resistance-associated proteins (Mrps) in rodents. Pgp and Mrp transporters are members of the ATP-binding cassette (ABC) superfamily, are regulated by AhR, CAR, and PXR, and play important roles in the efflux of xenobiotics and THs into the bile for elimination.

Binding to Thyroid Receptors

Limited evidence in fish and other vertebrates supports that PBDEs may, in part, be mediating effects on the thyroid by altering TR expression and signaling. The transcriptional activity of nuclear TRs is thought to be mediated by both the presence and absence of T3 due to its ability to bind to TRE regions of regulated genes in both the presence and absence of ligand. As in other vertebrates, two genetically distinct receptors TRα and TRβ have been identified in teleost fishes, including zebrafish,^174,175 flounder (P. olivaceus)¹⁷⁶; goldfish,¹⁷⁷ fathead minnows,^178,179 Nile tilapia (O. niloticus),¹⁸⁰ and Atlantic salmon.¹⁸⁰ Additional receptor subtypes with the capacity to bind TH have also been identified attributable to gene duplication and alternative mRNA splicing. Two trα genes, thraa (original) and thrab (duplication) have been described in zebrafish with the thraa gene shown to encode two protein variants, TRαA1 and TRαA1–2.^181,182 Two TRβ isoforms have also been identified in teleosts.^175,180 While TR variants arising from TRα and TRβ have also been described in other vertebrates, including humans, the general structure and function of TRs appears well conserved across vertebrates, which has been the topic of several reviews.^183-185

Questions remain, however, concerning whether and how parent PBDEs and their OH-BDE metabolites interact with and bind to TRs. For instance, BDE-47 was reported to not interact as either an agonist or antagonist with TRβ1 in an in vitro binding assay and did not interfere with TRβ-responsive gene expression.¹⁸⁶ However, altered expression in thyroid-responsive genes was observed in the brain and livers of rodent pups exposed perinatally to BDE-47, suggesting that BDE-47 operated through alternative mechanisms to TRβ signaling. Studies in rat pituitary GH3 cell proliferation assays have shown BDE-127 and BDE-185 to be TR agonists while BDE-206 was a TR antagonist.^167,187,188 Cell based assays have shown that some OH-BDEs, including 3-OH-BDE-47, can inhibit T3 binding to TRs by antagonizing the receptor, whereas several other parent PBDEs and OH-BDEs have shown no TR affinity.^189,190 Another study showed TR antagonistic activity for BDE-209, -153, -154, -100, and PentaBDE.¹⁹¹ Limited evidence has shown some OH-BDEs to behave as weak TR agonists.^192-194 It has been suggested that hydroxy moieties in the 3 or 4 position of the phenyl ring along with two adjacent bromine substituents are necessary for OH-BDE to bind to TRs.¹⁸⁹ In addition, recent molecular docking assays have shown that OH-BDEs may have varied interactions with TR binding pockets depending on the degree of bromination with lower OH-BDEs (e.g., 6-OH-BDE-47, 5-OH-BDE-47) showing weak TR agonism while higher OH-BDEs (e.g., 3-OH-BDE-100, 3′-OH-BDE-154) antagonized TRs.¹⁹⁴ Taken together, data on TR binding of PBDEs and OH-BDEs continue to be inconsistent and have demonstrated both antagonistic and weak agonistic activities toward TRs, as well as no interactions, that may be attributable to hydroxylation and bromination patterns that influence binding geometries with the receptor.

PBDEs also may alter patterns of TR-responsive gene transcription, although this remains understudied. For instance, reductions in relative mRNA transcript abundances of brain transcription element binding protein (BTEB) were measured in adult fathead minnows exposed to BDE-47.¹⁵⁴ The downregulated expression of BTEB, which is a thyroid-responsive transcription factor involved in neurogenesis, was accompanied by declines in circulating T4 and reductions in trβ gene transcripts in the brains of treated minnows, suggesting that hypothyroidism in BDE-47 treated fish may elicit downstream effects on neurogenic capacity in adults. Similarly, in primary rat cerebellar granule cell cultures, BDE-99 was found to disrupt trα1 and trα2 mRNA transcript abundances, alter TR-responsive gene transcription (e.g., brain-derived neurotrophic factor), and increase the production of reactive oxygen species.¹⁹⁵ Finally, a study with CV-1 cell cultures measured suppressed TR binding with TREs through the DNA binding domain (vs. between THs and TRs) upon exposure to several PBDEs and OH-BDEs, with BDE-209 showing the greatest suppression at the lowest dose.¹⁹¹ The suppressed TR-TRE binding was then shown to inhibit TH-dependent dendrite arborization of cerebellar Purkinje cells, suggesting TR-TRE mediated impacts on PBDE neurotoxicity.

Recently, a second potential binding site for T3 and T4 was suggested in the ligand binding domain (LBD) of TRα ¹⁹⁶. This second binding site was identified on the surface of the TRα LBD in the same region where the F-domain (i.e., additional C-terminal amino acids) is located in some species. While human TRs do not appear to have F-domains, it has been identified within the TRαA1 isoform of zebrafish where it has been shown to constrain transcriptional activity by altering TR coactivator recruitment.¹⁸² Thus, one hypothesis put forth is that this second binding site within TRα may serve to suppress TR activation when elevated concentrations of TH are present.¹⁹⁶ While more work is needed, this second binding site could also play a role in mediating how environmental contaminants like PBDEs interact with TRs and alter the functioning of these nuclear receptors.

Neurodevelopmental Toxicity

Limited evidence of PBDE effects on neurodevelopment of fishes has been observed in early life stages of fish (zebrafish typically) for a subset of PBDE congeners. Specifically, neurodevelopmental abnormalities, including impaired normal motor behavior and inhibited neuron growth and differentiation, as well as morphological deformities have been measured in zebrafish larvae exposed to: PentaBDE^157,197,198; a mixture of BDE-47, -99, -100, -153, and -183¹⁹⁹; BDE-47,^200-203 BDE-49²⁰⁴; and BDE-209²⁰⁵ (Table 5).

Table 5. PBDE neurodevelopmental/developmental malformations measured in teleost fish

Species	Treatment	Route	Dose/Duration	Effects Observed	Ref
Zebrafish (Larv.)	BDE-47	Aqueous	0, 1.25, 5, 20 µM; 6–96 hpf	Impaired motor behavior           ↓ touch-response, swimming speed Inhibited axon growth	^200
Zebrafish (Larv.)	PentaBDE (DE-71)	Aqueous	0, 31, 68.7, 227.6 µg/l; 2–120 hpf	Altered behavior (light-dark stimulation) ↑AChE activity, ↑Ach; ↓ mRNA transcripts for MBP, a1-tubulin, Shh	^197
Zebrafish (Larv.)	PentaBDE (DE-71)	Parental	0.16, 0.8, 4.0 µg/l; 150 d	No effect on F1 hatching success, survival, or malformations; decreased locomotor activity (light-dark stimulation) ↓AChE activity, ↓ mRNA transcripts for MBP, GAP-43, GFAP, α1-tubulin, SYN2a; ↓ protein levels of α1-tubulin, SYN2a	^198
Zebrafish (Adu. females)	PentaBDE (DE-71)	Aqueous	0.45 µg/l, 9.6 µg/l; 60 d	↓ retinyl ester protein; ↓ CRBP mRNA transcripts (GI); ↑ retinoids (eyes, ovaries, eggs); ↑ CRBP mRNA transcripts (liver, eyes); ↓ retinal dehydrogenase, ↑ CYP26A (eyes)	^220
Zebrafish (Larv.)	BDE-209	Spiked sediment	12.5 mg/kg; 4 – 192 hpf	Impaired motor behavior (light/dark stimulation) In silico profiling: BDE-209 binding to neurologically active cocaine esterase, AChE, 5HTR2A, 5HTR2C, 5HTR3A, and tubulin α1A	^6
Zebrafish (Adu., offspring)	BDE-209	Aqueous	0.001 – 1 µM; 150 dph (Adu); bred at 120 dph	Parent: ↑ mortality (~44% high dose); Neg ctrl mortality ~38%; PBDE bioaccumulation (congeners not specified) Offspring: Delayed hatching, motor neuron development, loose muscle fibers, slow locomotion; hyperactivity (light-dark test)	^205
Zebrafish (Larv.)	BDE-47	Aqueous	100 – 5000 µg/l; 3–168 hpf	Delayed hatching, reduced growth, dorsal curvature, impaired CSF flow Cardiac toxicity at 96 hpf (tachycardia, arrhythmias)	^202
Zebrafish (Larv)	BDE-49	Aqueous	4 – 32 µM; 5 and 24 hpf	Dorsal curvatures, cardiac toxicity (reduced heart rate); neurobehavioral effects (impaired touch-escape responses)	^204
Mummichog (Larv., Juv.)	PentaBDE (DE-71)	Aqueous	0.001 – 100 µg/l; 0–7 hpf	Delayed hatching; No major deformities but tail curve asymmetry; ↓ activity; impaired fright response (Larv) Impaired learning (Juv.)	^98
Zebrafish (Larv.)	BDE-28, -47, -99, -100, -153, -183	Aqueous	0.635 – 10 mg/l; up to 168 hpf	Swimming rate ↑ (96–120 hpf), ↓ (168 hpf)           ↓ swimming rates (BDE-47; 168 hpf)           Developmental deformities (dorsal curvature at 120 hpf) w/mortality (BDE-28, -47, -99, -100) No effects w/BDE-153, -183	^199

α1-tubulin = neuron microtubulin protein; ACh = acetylcholine; AChE = acetylcholinesterase; BTEB = basic transcription element-binding protein; CRBP = cellular retinal binding protein; CSF = cerebral spinal fluid; dep = depuration; dpf = days post fertilization; dph = days post hatch; dio = iodothyronine deiodinase; GAP-43 = growth associated protein 43; GFAP = glial fibrillary acidic protein; HDT = highest dose tested; hpf = hours post fertilization; MBP = myelin basic protein; Shh = Sonic hedgehog; SYN2a = synapsin IIa

PentaBDE Neurotoxicity

For instance, in 96-hpf zebrafish parentally exposed to the PentaBDE mixture, several genes involved in central nervous system development were downregulated, including α1-tubulin, synapsin IIa, and myelin basic protein.¹⁹⁸ This decline in mRNA expression was accompanied by reduced protein expression of α1-tubulin and synapsin IIa as well as reduced locomotor activity among treated larvae. In another study by this same research group, the PentaBDE commercial mixture also caused a downregulation in mRNA transcript abundances of sonic hedgehog (Shh).¹⁹⁷ This suggests a possible contributory role for thyroid dysregulation in some PBDE related neurodevelopmental toxicity as the Shh pathway, as well as its coreceptors patched (Ptc) and smoothened (Smo), have been shown to be regulated by THs in embryonic forebrain signaling and development in mammals.²⁰⁶ Only one study has examined neurotoxicity endpoints in a fish species beyond zebrafish. This study, conducted in mummichogs (F. heteroclitus) detected hindered behavior and learning ability, as well as dorsal curvatures, in fish exposed to PentaBDE.⁹⁸

BDE-47, BDE-49, and BDE-209 Neurotoxicity

In BDE-47 exposed zebrafish, delayed hatching, neural defects, and cardiac arrhythmias were measured at 168 hpf in larvae exposed to 5 mg/l of BDE-47.²⁰² The cardiac toxicity and dorsal curvature deformities observed in this study were coupled with reduced flow rates of cerebrospinal fluid in neural tubes and brain ventricles of the hindbrain. The reduced flow rates of cerebrospinal fluid were postulated to be related to the dorsal curvatures that then possibly contributed to the measured neurotoxicity. Dorsal curvatures, attenuated heart rates, and impaired touch-escape responses were also measured in zebrafish larvae exposed to another tetra PBDE congener BDE-49.²⁰⁴ Moreover, neurodevelopmental impairments, including delayed hatching and hindered motor neuron development, loose muscle fiber deformities and slow locomotion, and hyperactivity under a light/dark stimulation test, have also been observed in zebrafish larvae exposed maternally to BDE-209.²⁰⁵ These behavioral findings with BDE-209 are consistent with another recent study in which zebrafish larvae exposed to sediment spiked with 12.5 mg/kg of BDE-209 from 4–192 hpf experienced hyperactive responses to light stimulation that may be linked to impaired neurodevelopment.⁶ Additional in silico binding assays in this study also predicted BDE-209 binding with several human proteins involved in neurological functioning, including: tubulin α1A involved in microtubule formation; acetylcholinesterase (AChE) involved in the breakdown of acetylcholine; 5HTR2A, 5HTR2C, and 5HTR3A, which are serotonin receptor system genes; and cocaine esterases.

Toxicity Mechanisms and Thyroid Interactions

In mammals, like in fish, neurodevelopmental toxicity of PBDEs is also an important toxicological endpoint of concern. For instance, PBDEs (BDE-47, -99, and -100) measured in umbilical cord blood of women have been found to be correlated with reduced performance of gestationally-exposed children (aged 0–6) on mental performance tests.²⁰⁷ Moreover, maternal prenatal and childhood PBDE exposures have been associated with reduced attention, fine motor coordination, and cognition (declines in IQ scores) among a California cohort of Mexican-American children.⁶⁴ A substantial number of studies in rodents, spanning different laboratories, have demonstrated also that PBDEs can elicit adverse neurobehavioral outcomes in early development.^208-210 Recent in vitro work has even shown that the prominent OH-BDE metabolite in humans, 6-OH-BDE-47, can disrupt adult neurogenesis by inhibiting neuronal differentiation and oligodendrocyte differentiation, proliferation, and survival of primary cultured adult neural stem/progenitor cells isolated from the brains of adult mice.²¹³ Other in vitro study has shown that the hydroxylated metabolites of BDE-47 may disturb intracellular calcium release.²¹⁴

Some neurological deficits and alterations measured in fish¹⁵⁴ and rodents^191,211,212 have been accompanied by reductions in circulating T4 and altered TH signaling, suggesting that one of the contributing neurotoxicity mechanisms may proceed through PBDE interference with TH regulation and signaling. However, PBDE neurotoxicity has been observed absent impacts on TH regulation, suggesting other mechanistic pathways. Indeed, a growing body of evidence suggests that PBDE mechanisms of neurotoxicity may operate by several pathways that include disrupted TH signaling, altered cholinergic neurotransmissions^215,216; impaired neuronal proliferation and plasticity,^191,214,217 and oxidative stress.^218,219 While the underlying mechanisms of PBDE neurotoxicity are unclear, continued testing in fish would be informative to better understand these underlying mechanisms of PBDEs effects on the development and functioning of the central and peripheral nervous systems of vertebrates.

Reproductive Toxicity

PBDE impacts on fish reproduction and reproductive development have been evaluated in a limited number of studies and congeners (Table 6). Studies that have examined PBDE effects on fish reproduction have presented mixed evidence of reduced fecundity, spawning, hatching success, and offspring survival as well as impaired fertility, particularly among male fish, with PBDE-induced alterations in spermatogenesis, declines in sperm counts, and feminization possibly playing important roles.

Table 6. PBDE reproductive alterations measured in teleost fish

Species	Treatment	Route	Dose/Duration	Effects Observed	Ref
Atlantic salmon (S. salar; Juv.)	PentaBDE, OctaBDE mixtures	Diet	10 mg/kg bw (d 1), 50 mg/ kg bw (d 4); 7 d, 14 d	No effects on protein expression/activity of Vtg, zona radiata, CYP1A	^75
Zebrafish (Juv.)	BDE-47	Diet	100 ng/g; 0.5–1 mg food/fish-d; 20–60 dph	No significant change in Vtg mRNA	^95
Zebrafish (Adu.)	PentaBDE (DE-71)	Aqueous	2, 300, 500 ng/l (measured); 120 dpf	Gene Expression: Males: ↑ GnRH, ERβ (brain); ↑ FSHβ, LHβ (pituitary); ↑ FSH-R, LH-R, CYP19a and ↓ CYP11a, 3β-HSD (testis); ↓ ERα, AR, Vtg (liver) / Females: ↑ GnRH and ↓ ERβ (brain); ↑ FSHβ, LHβ and ↓ GnRH-R (pituitary); ↑ 3β-HSD (ovary); ↑ ERα, AR (liver) / Sex hormones: ↓ E2 (males (high dose only) and females); ↑ T, 11-KT; ↑ T/E2 and 11-KT/E2 (males)	^223
Zebrafish (Adu.)	PentaBDE mixture (DE-71)	Aqueous	2, 300, 500 ng/l (measured); 120 dpf	Reduced spawning, fertilization, hatching, larval survival; ↑ GSI Increased malformations and percentages of male offspring	^224
Zebrafish (Adu., offspring)	BDE-209	Aqueous	0.001 – 1 µM; 150 dph (Adu); bred at 120 dph	↓ male/female GSI; ↓sperm count, motility Offspring: Delayed hatching, motor neuron development, loose muscle fibers, slow locomotion; hyperactivity (light-dark test)	^205
European flounder (P. flesus; Adu.); Zebrafish (Juv.)	PentaBDE (DE-71 purified of PBDD/Fs)	Spiked sediments, Diet	0+0.014 -700+14000 µg/g TOC + µg/g lipid; 101 d (flounder) 0–500 µg/l; 30 d (zebrafish)	↓ ovarian aromatase (CYP19) activity (flounder); ↓ fecundity (zebrafish)	^96
Fathead minnow (Adu. Breeding pairs)	BDE-47	Diet	2.4 µg/pair/ day; 12.3 µg/pair/day; 21 d	↓ mature spermatozoa	^154
Chinese rare minnow (G. rarus; Adu., Larv.)	BDE-209	Aqueous	0 – 10 µg/l; 21 d	↓ spermatogenesis	^155
Fathead minnow (Adu. Breeding pairs)	BDE-47	Diet	28.7 ± 1.6 µg/pair (bioencapsulated artemia); 25 d	Spawning ceased by 2-wks of exposure           Reduced fecundity           > 50% reduction in sperm counts No change in GSI, LSI; reduced condition, males	^108
Fathead minnow (Adu. males)	BDE-209	Diet	95 ng/g ww food and 10 µg/g ww food at 3% bw/day; 28 d with 14 d dep	↓ GSI at both low and high dose that extended into 14-d dep PTU positive control did not affect GSI	^62
Zebrafish (embryos)	BDE-28, BDE-183, BDE-209	Diet, Maternal	10 and 100 nmol/g food at 2% bw/day; 42 d	All three PBDEs transferred to eggs (BDE-28 > BDE-183 > BDE-209) Egg/maternal fish concentration ratios significant > 1.0 for BDE-183, BDE-209	^221
Rainbow trout (Adu. Males)	BDE-47	Diet	55 µg/kg-day; 17 d	No effects on embryonic survival at first cleavage (0.5 dph) or during eye development (19 dph)	^225
Atlantic salmon (Juv. males)	BDE-47, -153, -154 (alone and mixture)	Hepatocytes	0.01 – 100 µM; 48 h	Disturbed glucose homeostasis (PBDE mix, BDE-153); Altered cell proliferation processes (PBDE mix) Estrogenic responses (↑ Vtg; ZP3 mRNA) in males (BDE-47, PBDE mix)	^226
Marine medaka (O. melastigma; embryos)	BDE-47	Diet, Maternal	Breeding pairs: 1.3 ± 0.2 µg/day; 18 d	Maternal transfer of BDE-47           ↑ BDE-47 to 25 ng/egg (day 18) Maternal concentrations BDE-47 < males	^222
Zebrafish (Adu., offspring)	PentaBDE (DE-71)	Parental, Aqueous	1, 3, 10 µg/l (parents and offspring) + (parent alone); 5 min to sexual maturation (parents), 5 and 10 dph (offspring)	↓ hatching rate           ↑ malformation rates to offspring exposed parentally; DE-71 exposures by parental transfer plus directly to offspring worsened malformations	^158

11-KT = 11-keto-testosterone; 3β-HSD = 3-β-hydroxy steroid dehydrogenase; CRH = corticotrophin releasing hormone; CSF = cerebral spinal fluid; dep = depuration; dpf = days post fertilization; dph = days post hatch; E2 = estradiol; ER = estrogen receptor; FSH = follicle stimulating hormone; FSH-R = follicle stimulating hormone receptor; GSI = gonado-somatic index; GnRH = gonadotropin releasing hormone; HDT = highest dose tested; hpf = hours post fertilization; LH = luteinizing hormone; LH-R = luteinizing hormone receptor; PTU = 6-propyl-2-thiouricil; t = testosterone; TOC = total organic carbon; Vtg = vitellogenin; ZP3 = zona pellucida 3

PentaBDE Reproductive Effects

Reduced fecundity and larval survival have been measured in zebrafish exposed orally to the PentaBDE mixture purified of PBDDs/PBDFs.⁹⁶ Reductions in larval survival in this and other studies may be partly attributable to maternal transfer of PBDEs to eggs, which has been shown in zebrafish and marine medaka (O. melastigma) and could hinder normal developmental progression.^221,222 A study in zebrafish adults with the PentaBDE mixture reported sex-specific alterations in the relative expression of genes encoding an array of reproductive hormones and receptors along the HPG axis, as well as disruption in circulating levels of E2, T, and 11-keto-testosterone in males.²²³ In a second related study, this same research group also measured PentaBDE-induced reductions in spawning, fertilization, and hatching success along with reduced larval survival and higher percentages of male offspring.²²⁴ These reproductive impairments were accompanied by altered counts of spermatogonia, spermatocytes, and spermatids in the testis of treated fish. In some contradiction, however, this study reported an increase in the gonado-somatic index (GSI) in treated males.

BDE-47 Reproductive Effects

With regard to constituents of the PentaBDE commercial mixture, most of the limited research to date that has studied PBDE impacts on reproduction has been conducted with BDE-47. Reduced spawning has been observed in adult fathead minnow male/female pairs exposed orally to ~14 µg/fish of BDE-47 for 25 d, with reproduction completely ceased within 10 d of exposure.¹⁰⁸ The impaired reproduction may have been attributable to selective toxicity in male fathead minnows as decreased mature spermatozoa and reduced condition factors were noted in these fish. In another study, declines in mature spermatozoa also were measured, but with no effect on fecundity, in adult male fathead minnows exposed orally to ~6 µg/fish-day of BDE-47 for 21 d.¹⁵⁴ BDE-47 also elicited no effects on the morphological development of offspring from adult male rainbow trout exposed to 55 µg/kg-day of BDE-47 for 17 d and then bred with untreated females.²²⁵ These data suggest that other components in the PentaBDE mixture in addition to BDE-47 may play an important role in affecting fecundity, embryonic development, and adult male reproduction.

PBDE Effects on Male Fish Reproduction

Some studies have suggested a role for PBDEs in the feminization of male fish. Specifically, dose-dependent increases in relative mRNA transcripts encoding vitellogenin (Vtg; egg yolk precursor protein) and zona radiata protein (eggshell protein) were measured in hepatocytes of juvenile male Atlantic salmon (S. solar) exposed to BDE-47 and a PBDE mixture (BDE-47, 153, -154).²²⁶ However, some of these in vitro findings have not been reproduced with in vivo studies of juvenile Atlantic salmon exposed to the PentaBDE mixture⁷⁵ or in zebrafish larvae exposed to BDE-47,⁹⁵ whereby no change in the expression of Vtg mRNA transcript abundances were detected. Thus, while the in vitro evidence is limited because it may not account for the extensive and coordinated interactions among cells and tissues, conversely the in vivo research conducted to date targeting PBDE effects on male fish reproduction has only been in two species, targeting largely transcriptional changes in immature animals for just a limited number of genes. In addition, very little is known about the potential reproductive effects of BDE-209 in fishes but limited evidence suggests alterations to some reproductive endpoints in male fish. Specifically declines in spermatogenesis were observed in male Chinese rare minnows (G. rarus) exposed aqueously to BDE-209,¹⁵⁵ and reductions in the gonadal somatic index (GSI) have been measured in adult male fathead minnows exposed to BDE-209^62,205 with evidence of declining sperm counts in some of this research.²⁰⁵ Taken together, more research is necessary to understand whether PBDEs play a role in eliciting potential feminization and reproductive impairments in male animals.

Toxicity Mechanisms and Thyroid Interactions

Fish have evolved diverse reproductive strategies that are closely integrated with gonadal differentiation and functioning and demonstrate sensitivity to external environmental cues, such as photoperiod and temperature. While there are important differences between reproduction in mammalian and non-mammalian vertebrates, reproduction in jawed vertebrates is controlled by the HPG axis and the structure of this endocrine system is highly conserved. It is possible that there could be shared pathways of PBDE impacts on mammalian and non-mammalian vertebrate sex steroid synthesis and gonadal development and functioning. It is also possible that there may be important species- and sex-specific differences in reproductive responses to PBDEs as current studies are limited by teleost species and focused largely on altered reproduction in males with less known about reproductive effects in female fish.

For instance, in female fish, declines in ovarian aromatase (CYP19) activity were measured in European flounder (P. flesus) exposed to the PentaBDE mixture purified of PBDDs/PBDFs, suggesting that PBDEs may be altering pathways of steroidogenesis in female fish.⁹⁶ Somewhat consistent with these findings, in vitro testing also has shown potential inhibitory effects of OH-PBDEs on CYP19 and CYP17 in human placental microsomes²²⁷ and human adrenocortical carcinoma cells.^228,229 Several OH-BDEs have also been found to bind with the estrogen receptor (ER) with the general trend that para-OH metabolites displayed the highest affinity for ERs with the lower OH-BDEs (1–4 bromines) tending to act as weak agonists while higher OH-BDEs (4 bromines or more) had antagonistic properties.^189,230,231In addition, rodent studies^232-236 and cell-based assays^187,237 have shown that some PBDEs (PentaBDE, BDE-47, -99, -100, and -209) may be estrogenic and/or induce feminization in male animals by anti-androgenic pathways. In humans, epidemiology studies have measured PBDE associations with: cryptorchidism²³⁸; early onset of menarche²³⁹; decreased testosterone, luteinizing hormone (LH), and follicle stimulating hormone (FSH) in adult men¹⁵¹; increased E2 in 3-mo old boys (BDE-154)²⁴⁰; and decreased sperm counts and testis size in young adults (BDE-153).²⁴¹ Taken together, there appear to be important common mechanistic pathways of PBDE effects on vertebrate reproduction that are not well understood at this time.

While some evidence points to PBDEs affecting the reproductive health of fish and other vertebrates, few studies have examined interactions between thyroid and reproductive functioning. Thus, questions remain as to whether PBDE effects on reproduction are being mediated directly and/or indirectly by altered TH regulation. In mammals, both hypothyroidism and hyperthyroidism have been shown to impair reproductive physiology and lower fertility.²⁴² An early review described important interactions between TH regulation and reproductive physiology in fishes.²⁴³ More recently, studies in goldfish (C. auratus) and zebrafish suggest that THs may have important inhibitory effects on teleost reproductive functioning at different levels of the HPG axis, including by: inhibiting pituitary LH and FSH; and reducing steroidogenesis and gonadal aromatase expression.^244,245 Recent evidence, in some contrast, also supports a stimulatory role for THs in the proliferation of sertolli cells and spermatogonia in zebrafish testes.²⁴⁶

Little is known about the role of PBDE-induced TH disruption in potential reproductive toxicity in fishes or other vertebrates. One hypothesis is that PBDEs could be mimicking THs that is in turn leading to altered steroidogenesis and steroidal hormone regulation among exposed animals, although this has not been examined. However, in data collected in male fathead minnows exposed orally to BDE-209, the model goitrogen 6-propyl-2-thiouracil (PTU), which was used as a positive control, reduced TH levels as predicted but had no effect on the GSI, unlike BDE-209 which caused substantial reductions in the GSI. These results suggest that BDE-209 reproductive effects in male fish could be acting through thyroid-independent pathways.⁶² Still though, it remains unclear from the limited data whether PBDE effects on reproduction are mediated by directly impairing HPG functioning or whether these effects are also mediated in cross-talk with perturbations of the thyroid.

Research Needs and Conclusions

A substantial body of evidence suggests that PBDE metabolism in teleost fishes proceeds through reductive debromination pathways. Studies report both the presence and absence of OH-BDE metabolites forming in fish, and given their bioactivity, it would be informative to investigate further whether these metabolites form in vivo in fish. It might be expected that the OH-BDEs, if they are being produced by fish, would be found at higher concentrations in the blood and highly perfused tissues like the liver where they are formed mostly. Related to this, questions remain as to the identity and kinetics of the enzymatic biotransformation pathways of PBDE metabolism in fish with some indirect evidence implicating Dio enzymes. In mammals, PBDEs appear to operate predominantly through AhR-independent toxicity pathways. However, the extent to which these pathways (e.g., CAR, PXR) are operational in fish exposed to PBDEs is less clear, but potentially relevant to the hypothyroidism observed given their additional role in TH metabolism and elimination.

PBDEs have been shown now to disrupt TH regulation and signaling in several teleost species with mechanisms of action that proceed through multiple pathways depending on the congener, fish life stage, and tissue type, including by: enhancing the metabolism and elimination of THs; binding competitively with plasma and membrane bound transporter proteins; altering interactions of T3 ligand with nuclear TRs; disrupting Dio activity; and altering the transcription of genes involved in TH production, transport, and genomic signaling. Despite expanding knowledge of these thyroid disruption mechanisms, there continues to be a limited understanding of PBDE impacts on thyroid-response gene expression and downstream apical endpoints of concern, including neurodevelopment and reproduction. PBDE effects on neurological, developmental, and reproductive physiology have been indicated across fish and other vertebrates exposed to PBDEs, suggestive of common toxicity pathways that remain unclear. Related to this, additional work is needed to understand whether there exist localized compensatory or adaptive physiological responses of the thyroid system to TH insufficiency caused by PBDE exposures. Continued study of these apical endpoints of concern in fish would be helpful for not only understanding potential toxicities in free ranging fishes, but also would contribute to understanding PBDE mechanisms of toxicity in humans.

The thyroid endocrine systems of fish and mammals are well conserved and similar in structure and function, supporting the relevance of fish as models for understanding PBDE effects more broadly across vertebrate taxa. However, there are differences between mammalian and non-mammalian thyroid signaling that could have implications for using fish as models for evaluating PBDE thyroid toxicity in higher level animals. One important difference is that unlike in mammals, the fish thyroid may not to be centrally directed through the HPT but rather appears to rely strongly on localized peripheral tissues for T3 production and regulation. There is evidence for this dominant peripheral signaling in mammalian models as well, but nonetheless demonstrates the need to continually evaluate the choice of animal model when studying chemically induced thyroid disruption. For some PBDE congeners, evidence in fish demonstrates thyroid dysregulation at low doses (i.e., ~ppb levels) that roughly compares to levels detected in the environment and tends to be lower than doses typically administered in rodent studies. Additional low dose studies are needed to determine whether non-monotonic dose responses are occurring in fish and other vertebrates exposed to PBDEs.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

This review was supported by National Institute of Environmental Health Sciences research grants (R01-ES016099 and T32ES007060) and US EPA STAR graduate fellowship (FP-917145010). Findings and conclusions in this article are those of the authors and do not necessarily represent the views of the NIEHS or EPA. The authors would also like to thank Dr. Linda Birnbaum, Dr. Rich Di Giulio, Dr. David Hinton, Dr. Sean Lema, and Dr. Joel Meyer for their insightful comments as committee members on the Ph.D. dissertation that served as the basis for this review.

Citation: Noyes PD, Stapleton HM. PBDE flame retardants: Toxicokinetics and thyroid hormone endocrine disruption in fish. Endocrine Disruptors 2014; 2:e29430; 10.4161/endo.29430