EPA/635/R-07/006F
www.epa.gov/iris
vvEPA
TOXICOLOGICAL REVIEW
OF
2,2',4,4',5-PENTABROMODIPHENYL
ETHER (BDE-99)
(CAS No. 60348-60-9)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
June 2008
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
11
-------
CONTENTS—TOXICOLOGICAL REVIEW OF
2,2',4,4',5-PENTABROMODIPHENYL ETHER (CAS No. 60348-60-9)
LIST OF TABLES v
LIST OF FIGURES v
LIST OF ACRONYMS vii
FOREWORD ix
AUTHORS, CONTRIBUTORS, AND REVIEWERS x
1. INTRODUCTION 1
2. CHEMICAL AND PHYSICAL INFORMATION 3
3. TOXICOKINETICS 5
3.1. ABSORPTION 5
3.2. DISTRIBUTION 7
3.2.1. Human Data 7
3.2.1.1. Adipose Tissue 7
3.2.1.2. Liver 9
3.2.1.3. Human Milk 9
3.2.1.4. Blood 11
3.2.1.5. Placental Transport 12
3.2.2. Animal Data 14
3.3. METABOLISM 18
3.4. ELIMINATION 21
3.4.1. Half-life Determinations 23
3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS 23
4. HAZARD IDENTIFICATION 24
4.1. STUDIES IN HUMANS—EPIDEMIOLOGY, CASE REPORTS, CLINICAL
CONTROLS 24
4.2. SHORT-TERM, SUBCHRONIC, AND CHRONIC STUDIES AND CANCER
BIO AS SAYS IN ANIMALS—ORAL AND INHALATION 24
4.2.1. Short-term and Subchronic Studies 24
4.2.1.1. Mice 24
4.2.1.2. Rats 26
4.2.2. Chronic Studies and Cancer Bioassays 26
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES 26
4.3.1. Mice 26
4.3.1.1. Eriksson et al. (2001) 26
4.3.1.2. Eriksson et al. (2002) 27
4.3.1.3. Viberg et al. (2002) 28
4.3.1.4. Viberg et al. (2004a) 29
4.3.1.5. Viberg et al. (2004b) 30
4.3.1.6. Ankarberg (2003) 31
4.3.1.7. Branch! et al. (2002) 31
4.3.1.8. Branch! et al. (2005) 33
iii
-------
4.3.2. Rats 34
4.3.2.1. Kuriyama et al. (2005) 34
4.3.2.2. Viberg et al. (2005) 36
4.3.2.3. Talsness et al. (2005) 37
4.4. OTHER DURATION-OR ENDPOINT-SPECIFIC STUDIES 38
4.4.1. Subcutaneous Exposures 38
4.4.1.1. Lilienthal et al. (2005) 38
4.4.1.2. Lilienthal et al. (2006) 38
4.4.1.3. Ceccatelli et al. (2006) 39
4.4.2. Receptor Site Interactions 41
4.4.2.1. Aryl Hydrocarbon Receptors 42
4.4.2.2. Other CYP-450 Inducing Receptors 44
4.4.2.3. Estrogen Receptors 45
4.4.2.4. Androgen Receptors 47
4.4.2.5. Acetylcholine Receptors 47
4.4.3. Thyroid Effects 49
4.4.4. Neurotoxicity 50
4.4.5. Immunotoxicity 51
4.4.6. Cytotoxicity 52
4.4.7. Genotoxicity 52
4.5. SYNTHESIS OF MAJOR NONCANCER EFFECTS 52
4.5.1. Oral 52
4.5.2. Inhalation 53
4.5.3. Mode-of-Action Information 53
4.6. EVALUATION OF CARCINOGENICITY 55
4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 55
4.7.1. Possible Childhood Susceptibility 55
4.7.2. Possible Gender Differences 56
5. DOSE-RESPONSE ASSESSMENTS 57
5.1. ORAL REFERENCE DOSE (RfD) 57
5.1.1. Choice of Principal Study and Critical Effect—with Rationale and
Justification 57
5.1.2. Methods of Analysis 60
5.1.3. RfD Derivation 63
5.1.4. Previous RfD Assessment 64
5.2. INHALATION REFERENCE CONCENTRATION (RfC) 65
5.3. CANCER ASSESSMENT 65
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD
AND DOSE RESPONSE 66
6.1. HUMAN HAZARD POTENTIAL 66
6.2. DOSE RESPONSE 66
7. REFERENCES 68
APPENDIX A: SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION A-l
iv
-------
APPENDIX B: BENCHMARK DOSE MODELING FOR BDE-99 B-l
LIST OF TABLES
Table 2-1. IUPAC number and bromine substitution pattern of some pentaBDE congeners 3
Table 2-2. Physical and chemical properties of BDE-99 4
Table 3-1. Median PBDE congener concentrations in human biological media in the U.S 8
Table 3-2. Median PBDE congener concentrations in maternal and fetal sera in the U.S.
and Sweden 13
Table 4-1. Receptor interaction studies of pentaBDE congeners 41
Table 5-2. Summary of BMD modeling output results with good data fit in mice 61
Table B-l. Rearing habituation in 2-month-old male mice B-4
Table B-2. Rearing habituation in 8-month-old male mice B-9
Table B-3. Rearing habituation in 2-month-old female mice B-9
Table B-4. Rearing habituation in 8-month-old female mice B-13
Table B-5. Summary of BMD and BMDL results B-17
Table B-6. BMD modeling results for the endpoints duration per day and LBI counts
per phase on PND 36 for male rats exposed to BDE-99 in utero on GD 6 B-20
Table B-7. Summary of the quantal dose-response modeling results based on the
percent of animals with less than two ejaculations in male rats exposed
to BDE-99 in utero on GD 6 B-24
LIST OF FIGURES
Figure 2-1. Chemical structure of pentaBDE 3
Figure 3-1. Proposed metabolic pathway for BDE-99 in male rats 19
Figure B-l. Unrestricted fourth-order polynomial model fit to rearing habituation in
2-month-old male mice exposed to BDE-99 B-2
v
-------
Figure B-2. Low-dose behavior of unrestricted fourth-order polynomial model fit
to rearing habituation in 2-month-old male mice exposed to BDE-99 B-3
Figure B-3. Dose-response relationship based on rearing habituation in 2-month-old
male mice B-8
Figure B-4. Dose-response relationship based on rearing habituation in 2-month-old
female mice B-12
Figure B-5. Dose-response relationship based on rearing habituation in 8-month-old
female mice B-16
Figure B-6. Dose-response relationship based on activity duration per day on PND 36
in male rats exposed to BDE-99 in utero on GD 6 B-20
Figure B-7. Log-logistic model fit based on percent of animals with less than two
ejaculations in male rats exposed to BDE-99 in utero on GD 6 B-25
VI
-------
LIST OF ACRONYMS
Ah
AIC
BDE-99
bDNA
BMD
BMDL
BMDS
BMR
CALUX
CAR
CASRN
CBS
CDD
CDF
cDNA
CYP-450
decaBDE
DHT
DNA
EC50
ER
EROD
FOB
GD
GSH
hexaBDE
i.v.
IC50
IGF
IgG
IRIS
IUPAC
LBI
LDH
LH
LOAEL
Iw
mRNA
MTT
MUP
NOAEL
octaBDE
PBDE
PCB
aryl hydrocarbon
Akaike Information Criterion
2,2',4,4',5-pentabromodiphenyl ether
branched DNA
benchmark dose
95% lower bound on the BMD
benchmark dose software
benchmark response
Chemical-Activated LUciferase expression
constitutive androstane receptor
Chemical Abstracts Service Registry Number
Coxsackie virus B3
chlorinated dibenzo-p-dioxin
chlorinated dibenzofuran
complementary DNA
cytochrome P-450
decabromodiphenyl ether
dihydrotestosterone
deoxyribonucleic acid
median effective concentration
estrogen receptor
ethoxyresorufin O-dealkylase
functional observational battery
gestational day
glutathione
hexabromodiphenyl ether
intravenous
median inhibitory concentration
insulin-like growth factor
immunoglobulin G
Integrated Risk Information System
International Union of Pure and Applied Chemistry
light beam interruption
lactate dehydrogenase
luteinizing hormone
lowest-observed-adverse-effect level
lipid weight
messenger RNA
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide
major urinary protein
no-observed-adverse-effect level
octabromodiphenyl ether
polybrominated diphenyl ether
polychlorinated biphenyl
vn
-------
PCR
pentaBDE
PKC
PND
PR
PROD
PTU
PXR
QNB
RfC
RfD
RNA
s.c.
SD
SXR
T3
T4
TCDD
tetraBDE
triBDE
TTR
UDPGT
UF
polymerase chain reaction
pentabromodiphenyl ether
protein kinase C
postnatal day
progesterone receptor
pentoxyresorufm O-dealkylase
6-n-propyl-2-thiouracil
pregnane X receptor
quinuclidinyl benzilate
reference concentration
reference dose
ribonucleic acid
subcutaneous
standard deviation
steroid X receptor
triiodothyronine
thyroxine
2,3,7,8-tetrachlorodibenzo-p-dioxin
tetrabromodiphenyl ether
tribromodiphenyl ether
transthyretin
uridine diphosphoglucuronosyl transferase
uncertainty factor
Vlll
-------
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and rationale
for the hazard and dose-response assessment in IRIS pertaining to chronic exposure to 2,2',4,4',5-
pentabromodiphenyl ether. It is not intended to be a comprehensive treatise on the chemical or
toxicological nature of 2,2',4,4',5-pentabromodiphenyl ether (BDE-99).
The majority of the available toxicological information on the pentabromodiphenyl ether
homolog group (CAS No. 32534-81-9) relates to the pentabromodiphenyl congener BDE-99
(CASRN 60348-60-9). Toxicological information related to other congeners in the
pentabromodiphenyl ether homolog group is also discussed. However, this health assessment
does not deal with commercial mixtures of brominated diphenyl ether homologs containing
pentabromodiphenyl ether as one of the constituents of commercial formulations. In addition to
BDE-99, IRIS health assessments have also been prepared for three other polybrominated
diphenyl ether congeners: tetraBDE-47, hexaBDE-153, and decaBDE-209. These four
congeners are those for which toxicological studies suitable for dose-response assessments were
available and are the ones most commonly found in the environment and human biological
media.
The intent of Section 6, Major Conclusions in the Characterization of Hazard and Dose
Response, is to present the major conclusions reached in the derivation of the reference dose,
reference concentration and cancer assessment, where applicable, and to characterize the overall
confidence in the quantitative and qualitative aspects of hazard and dose response by addressing
the quality of data and related uncertainties. The discussion is intended to convey the limitations
of the assessment and to aid and guide the risk assessor in the ensuing steps of the risk
assessment process.
For other general information about this assessment or other questions relating to IRIS,
the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
hotline.iris@epa.gov (email address).
IX
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGER
Joyce Morrissey Donohue, Ph.D.
Office of Water, Office of Science and Technology
Health and Ecological Criteria Division
U.S. Environmental Protection Agency
Washington, DC
AUTHORS
Joyce Morrissey Donohue, Ph.D.
Office of Water, Office of Science and Technology
Health and Ecological Criteria Division
U.S. Environmental Protection Agency
Washington, DC
Hend Galal-Gorchev, Ph.D.
Office of Water, Office of Science and Technology
Health and Ecological Criteria Division
U.S. Environmental Protection Agency
Washington, DC
Mary Manibusan
Office of Research and Development
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
OFFICE OF RESEARCH AND DEVELOPMENT CO-LEAD
Samantha J. Jones, Ph.D.
Office of Research and Development
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
REVIEWERS
This document has been reviewed by EPA scientists, interagency reviewers from other
federal agencies, and the public, and peer reviewed by independent scientists external to EPA. A
summary and EPA's disposition of the comments received from the independent external peer
reviewers and from the public is included in Appendix A of the Toxicological Review of
2,2',4,4',5-Pentabromodiphenyl ether (BDE-99).
-------
INTERNAL EPA REVIEWERS
TedBerner, M.S.
National Center for Environmental Assessment
Office of Research and Development
Linda Birnbaum, Ph.D.
Experimental Toxicology Division
National Health and Environmental Effects Research Laboratory
Office of Research and Development
Kevin Crofton, Ph.D.
Neurotoxicology Division
National Health and Environmental Effects Research Laboratory
Office of Research and Development
Karen Hogan, M.S.
National Center for Environmental Assessment
Office of Research and Development
Tammy Stoker, Ph.D.
Reproductive Toxicology Division
National Health and Environmental Effects Research Laboratory
Office of Research and Development
EXTERNAL PEER REVIEWERS
Ahmed E. Ahmed, Ph.D.
University of Texas Medical Branch
Richard J. Bull, Ph.D.
MoBull Consulting
Lucio G. Costa, Ph.D.
University of Washington
Ralph Kodell, Ph.D.
University of Arkansas for Medical Sciences
Deborah Rice, Ph.D. (chair)
State of Maine
XI
-------
1. INTRODUCTION
This document presents background information and justification for the Integrated Risk
Information System (IRIS) Summary of the hazard and dose-response assessment of 2,2',4,4',5-
pentabromodiphenyl ether (BDE-99). IRIS Summaries may include oral reference dose (RfD)
and inhalation reference concentration (RfC) values for chronic and other exposure durations,
and a carcinogenicity assessment.
The RfD and RfC, if derived, provide quantitative information for use in risk assessments
for health effects known or assumed to be produced through a nonlinear (presumed threshold)
mode of action. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with
uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
population (including sensitive subgroups) that is likely to be without an appreciable risk of
deleterious effects during a lifetime. The inhalation RfC (expressed in units of mg/m3) is
analogous to the oral RfD, but provides a continuous inhalation exposure estimate. The
inhalation RfC considers toxic effects for both the respiratory system (portal of entry) and for
effects peripheral to the respiratory system (extrarespiratory or systemic effects). Reference
values are generally derived for chronic exposures (up to a lifetime), but may also be derived for
acute (<24 hours), short-term (>24 hours up to 30 days), and subchronic (>30 days up to 10% of
lifetime) exposure durations, all of which are derived based on an assumption of continuous
exposure throughout the duration specified. Unless specified otherwise, the RfD and RfC are
derived for chronic exposure duration.
The carcinogenicity assessment provides information on the carcinogenic hazard
potential of the substance in question and quantitative estimates of risk from oral and inhalation
exposure may be derived. The information includes a weight-of-evidence judgment of the
likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic
effects may be expressed. Quantitative risk estimates may be derived from the application of a
low-dose extrapolation procedure. If derived, the oral slope factor is a plausible upper bound on
the estimate of risk per mg/kg-day of oral exposure. Similarly, an inhalation unit risk is a
plausible upper bound on the estimate of risk per ug/m3 air breathed.
Development of these hazard identification and dose-response assessments for BDE-99
has followed the general guidelines for risk assessment as set forth by the National Research
Council (1983). EPA Guidelines and Risk Assessment Forum Technical Panel Reports that may
have been used in the development of this assessment include the following: Guidelines for the
Health Risk Assessment of'ChemicalMixtures (U.S. EPA, 1986a), Guidelines for Mutagenicity
Risk Assessment (U.S. EPA, 1986b), Recommendations for and Documentation of Biological
Values for Use in Risk Assessment (U.S. EPA, 1988), Guidelines for Developmental Toxicity
1
-------
Risk Assessment (U.S. EPA, 1991), Interim Policy for Particle Size and Limit Concentration
Issues in Inhalation Toxicity (U.S. EPA, 1994a), Methods for Derivation of Inhalation Reference
Concentrations and Application of Inhalation Dosimetry (U.S. EPA, 1994b), Use of the
Benchmark Dose Approach in Health Risk Assessment (U.S. EPA, 1995), Guidelines for
Reproductive Toxicity Risk Assessment (U.S. EPA, 1996), Guidelines for Neurotoxicity Risk
Assessment (U.S. EPA, 1998), Science Policy Council Handbook: Risk Characterization (U.S.
EPA, 2000a), Benchmark Dose Technical Guidance Document (U.S. EPA, 2000b),
Supplementary Guidance for Conducting Health Risk Assessment of Chemical Mixtures (U.S.
EPA, 2000c), A Review of the Reference Dose and Reference Concentration Processes (U.S.
EPA, 2002), Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), Supplemental
Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens (U.S. EPA,
2005b), Science Policy Council Handbook: Peer Review (U.S. EPA, 2006a), and A Framework
for Assessing Health Risks of Environmental Exposures to Children (U.S. EPA, 2006b).
The literature search strategy employed for this compound was based on the Chemical
Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
scientific information submitted by the public to the IRIS Submission Desk was also considered
in the development of this document. The relevant literature was reviewed through November
2007.
-------
2. CHEMICAL AND PHYSICAL INFORMATION
Pentabromodiphenyl ether (CASRN 32534-81-9) is one of the possible 10 homologs of
polybrominated diphenyl ethers (PBDEs). Figure 2-1 shows the chemical structure of
pentabromodiphenyl ether (pentaBDE). The number of possible congeners of pentaBDE is 46,
with International Union of Pure and Applied Chemistry (IUPAC) numbers 82 to 127. The
IUPAC number and bromine substitution pattern of some pentaBDE congeners that have been
investigated in various studies are given in Table 2-1.
m + n = 5
Figure 2-1. Chemical structure of pentaBDE.
Table 2-1. IUPAC number and bromine substitution pattern of some
pentaBDE congeners
IUPAC number
BDE-85
BDE- 99
BDE-100
BDE-105
BDE-119
BDE-126
Bromine substitution pattern
2,2',3,4,4'-PentaBDE
2,2',4,4',5-PentaBDE
2,2',4,4',6-PentaBDE
2,3,3',4,4'-PentaBDE
2,3',4,4',6-PentaBDE
3,3',4,4',5-PentaBDE
PentaBDE is found in commercial pentaBDE, which is usually composed of a mixture of
pentaBDE (50-62%), tetraBDE (24-38%), and hexaBDE (4-12%) (U.S. EPA, 2005c). The
relative proportions by weight of various PBDE congeners in the commercial pentaBDE
DE-71™ are approximately 43% (pentaBDE-99), 28% (tetraBDE-47), 8% (pentaBDE-100), 6%
(hexaBDE-153), and 4% (hexaBDE-154). Tribromodiphenyl ether (triBDE)-28 and -33 and
tetrabromodiphenyl ether (tetraBDE)-49 and -66 are present at about 1% or less in the
-------
formulation (Great Lakes Chemical Corporation, 2003). Manufacture of the commercial
pentaBDE mixture was phased out in the U.S. at the end of 2004.
The predominant pentaBDE congener in environmental media, biota, and human tissues
is usually BDE-99 (CASRN 60348-60-9), followed by BDE-100 (CASRN 189084-64-8).
Physical and chemical properties of BDE-99 are listed in Table 2-2.
Table 2-2. Physical and chemical properties of BDE-99
Parameter
Synonym
CASRN
Chemical formula
Physical form
Molecular weight
Vapor pressure (Pa) at 25°C
Melting point (°C)
Solubility in water (ug/L)
Henry's law constant (Pa x m3 x mor1)
at 25°C
Log octano I/water partition coefficient
(KoW) at 25°C
Log octano I/air partition coefficient
(Koa) at 25°C
Relative density (at 25°C)
Value
Benzene, l,2,4,-tribromo-5-(2,4-
dibromophenoxy)-;
2,2',4,4',5-pentabro mo diphenyl ether;
BDE-99
60348-60-9
C12H5Br50
Amber solid
564.7
5 x 10~5
93
2.4
0.60
6.5-8.4
11.3
2.28
Reference
U.S. EPA (2004)
U.S. EPA (2004)
U.S. EPA (2004)
Flemming et al. (2000)
U.S. EPA (2004)
Wong etal. (2001)
Palm etal. (2002)
Stenzel and Markley
(1997)
Cetin and Odabasi
(2005)
Braekevelt et al. (2003);
ATSDR3 (2004)
Chen etal. (2003)
Flemming et al. (2000)
3ATSDR = Agency for Toxic Substances and Disease Registry.
-------
3. TOXICOKINETICS
Data on the toxicokinetics of the pentaBDEs in humans are limited to findings on levels
in adipose tissues, blood, liver, and maternal milk, which demonstrate that they are absorbed
from the environment and distributed to tissues. Studies of BDE-99 in several strains of rats
suggest that absorption varies between 60 and 90%. Quantitative data on absorption in mice are
more limited; however, there is uptake from the gastrointestinal track based on radiolabel found
in tissues and excreta. The highest levels of radiolabel following oral exposure are found in
adipose tissues, muscle, skin, and liver. Much of the metabolism data have been derived from
studies in rats rather than mice. Both species form a variety of hydroxylated metabolites via the
activity of cytochrome P-450 (CYP-450) isozymes. In rats there appears to be some hydrolysis
of the ether bond resulting in production of brominated phenols. There is evidence that
conjugation with glutathione (GSH), sulfate, and/or glucuronic acid can occur. There are no data
on whether the radiolabel in tissues is present as metabolites or parent, although it is presumed
that the material in adipose tissue is parent compound based on its lipophilicity. The metabolites
are excreted with bile and feces and to a lesser extent in urine. Some of the urinary metabolites
in rats are excreted bound to albumin, while in mice metabolites are found bound to one
(BDE-99) or two (BDE-100) major urinary proteins.
3.1. ABSORPTION
There are no direct studies of BDE-99 absorption in humans. The data that demonstrate
human absorption come from measurements of BDE-99 in human biological media after
anthropogenic exposures but do not permit estimation of route-specific uptake parameters.
Data on absorption of BDE-99 in several strains of rats and mice are available. There are
also some data for BDE-85 and BDE-100. In a study by Hakk et al. (2002a), 14C-BDE-99 (>98%
purity) in corn oil was given as a single oral dose of 8 mg/kg (1.0 uQ/rat) to groups of
conventional (three/group) and bile-duct-cannulated (five/group) male Sprague-Dawley rats.
Urine, feces, and bile were collected at daily intervals over 3 days. In the first 24 hours after
administration, 22 ± 16% of the dose was present in the feces in the conventional rats. This
suggests that absorption was very variable among the rats, and estimates of absorption range
from about 60-90%. The fecal excretion on day 1 was much higher in the bile-duct-cannulated
rats (53 ± 27%), indicating that gastrointestinal emulsion by bile was a major factor governing
absorption. The estimated amounts absorbed by the bile-duct-cannulated rats were also quite
variable and ranged from 21-74%. There was excretion of BDE-99 in bile on day 1 after
exposure (0.7 ± 0.8% of the dose). However, since it was low, it had little effect on the
estimates of absorption.
-------
Chen et al. (2006) estimated the absorption of 14C-BDE-99 (95.6% purity; 34-37 jiCi/kg)
to be 85% in male F344 rats administered a 0.6 mg/kg (1 |imol/kg) dose in corn oil by gavage.
This estimate was derived from a comparison of the excretion data from male rats receiving the
same dose intravenously compared with those exposed orally. The percent of the dose in the
feces 24 hours after oral exposure was 43.1 ± 4.7% compared to 27.3 ± 5.1% for the intravenous
(i.v.) exposure. The dose used by Chen et al. (2006) was lower than that used by Hakk et al.
(2002a) and, thus, is consistent with the higher average estimate of absorption.
The toxicokinetics of BDE-100 (2,2',4,4',6-pentaBDE) were studied in male Sprague-
Dawley rats (Hakk et al., 2006). Doses of 4.1 mg/rat of 14C-BDE-100 (purity >95%; 0.9 uCi/rat,
equivalent to about 17 mg/kg) were administered to groups of nine conventional or bile-duct-
cannulated rats. Urine, bile, and feces were collected at 24-hour intervals for 72 hours. About
11.5 ± 10.1% of the radiolabel was found in the feces of conventional rats 24 hours after
administration, suggesting that absorption of BDE-100 varies among animals and is greater than
80%. As was the case for the Hakk et al. (2002a) study discussed above, the level in the
collected 24-hour fecal matter from the bile-duct-cannulated rats was higher than in the
conventional rats (16.8 ± 25.7%).
Chen et al. (2006) found that the absorption of 14C-BDE-99 (95.6% purity; 34-37
|iCi/kg) in male B6C3F1 mice administered 0.6 mg/kg (1 |imol/kg) was comparable to that in
F344 rats (85%), based on the comparison of fecal loss after oral (27.1 ± 5.3%) versus i.v.
(12.9 ± 2.2%) exposures.
Darnerud and Risberg (2006) examined the tissue distribution of uniformly radiolabeled
BDE-85 or BDE-99 in adult female C57BL mice after i.v. or gavage exposure to 20 |imol/kg of
the congener in dimethyl sulfoxide (25 |iCi/g body weight; about 11 mg/kg), using a qualitative
whole-body autoradiography technique. The presence of radiolabel in tissues 24 hours and
96 hours after oral dosing demonstrated absorption via the gastrointestinal tract. The levels of
radiolabel in the gastrointestinal tract were higher after oral exposures than after i .v. exposures,
demonstrating some excretion of unabsorbed compound with the fecal matter.
Eriksson et al. (2002) demonstrated that radiolabeled BDE-99 can be taken up and
retained in the neonatal mouse brain. Neonatal NMRI male mice (five/group) were administered
0 or 8 mg/kg of 14C-BDE-99 (purity >98%; 40.5uCi/kg) in a fat emulsion on postnatal day
(PND) 3, 10, or 19, and the animals were sacrificed 24 hours or 7 days after administration. The
amount of radioactivity in the brain was between 0.4 and 0.5% of the administered dose in the
three different age categories 24 hours after administration. Seven days after administration,
14C-BDE-99 or its metabolites could still be detected in the brain. These data cannot be used to
quantify absorption but do demonstrate uptake from the gastrointestinal tract and transport
across the blood-brain barrier in mice.
-------
3.2. DISTRIBUTION
The high Kow of BDE-99 suggests a strong potential for accumulation in lipid-rich
tissues. This property of BDE-99 is quite evident from the data on distribution in humans and
animals.
3.2.1. Human Data
The human data described below come from monitoring of PBDEs in human populations
rather than from measured dosing studies. The data demonstrate that humans are exposed to
PBDEs and that absorption and distribution to some tissues occur. The data do not provide
information on the quantitative aspects of absorption or the kinetics of tissue distribution and
retention. The PBDE congener profiles in human biological media differ from the congener
profiles of the commercial PBDE mixtures. The reasons for the differences in congener
distributions are not known with any certainty. Monitoring data are available for human adipose
tissue, liver, milk, and blood samples and indicate a tendency for PBDEs to distribute to these
media. However, distribution studies have not been conducted in humans, and therefore it is not
known whether BDE-99 distributes to other tissues as well. The number of samples examined in
various studies and countries is small, and therefore the data should not be construed as
representative at the national level.
Biomonitoring data with emphasis on levels of PBDE congeners found in the U.S. are
summarized in Tables 3-1 and 3-2.
3.2.1.1. Adipose Tissue
Breast adipose samples were collected between 1996 and 1998 from 23 San Francisco
Bay area women as part of a case-control study on organochlorine compounds and breast cancer
(She et al., 2002). Women ranged from 28-62 years of age and were predominantly Caucasian
and born in the U.S. Pathology reports indicated 12 women had malignancies, 8 had benign
tumors, and 3 had ductal carcinomas in situ, a condition considered by some as transitional to
malignancy. Breast adipose samples were collected during biopsy or breast surgery and were
analyzed for tetraBDE (BDE-47), pentaBDEs (BDE-99 and -100), and hexabromodiphenyl
ethers (hexaBDEs) (BDE-153 and -154). Mean and median concentrations of the sum of these
PBDEs were 86 and 41 ng/g lipid weight (Iw), respectively. Median concentrations of individual
PBDE congeners are given in Table 3-1. The highest concentrations found were for tetraBDE,
followed by hexaBDEs and pentaBDEs, a distribution that does not follow that of the
commercial pentaBDE used in the U.S. There was an inverse relationship between the sum of
the concentrations of these PBDEs in breast adipose tissue and age, with women younger than
the median age of 48 years having significantly higher concentrations of PBDEs in adipose tissue
than women older than 48. This may imply that different activities may expose different age
7
-------
groups more than others or that some PBDE congeners may accumulate differently with age.
However, only 23 women were tested and firm conclusions cannot be made. Five paired
samples of breast and abdominal adipose tissues were also analyzed for tetra- to hexaBDEs.
Abdominal and breast concentrations of PBDEs were highly correlated and of comparable
magnitude.
Table 3-1. Median PBDE congener concentrations in human biological
media in the U.S.
Adipose
tissue
Adipose
tissue
Breast
milk
Breast
milk
Maternal
serum
Fetal
serum
Serum
pools
Serum,
pregnant
women
Year3
1996-
1998
2003-
2004
2002
2003
2001
2001
2000-
2002
1999-
2001
Nb
23
52
47
40
17
12
T
24
BDE-
47
BDE-
99
BDE-
100
BDE-
85
BDE-
153
BDE-
154
BDE-
183
£
PBDEs
ng/glw
18
29
18
28
28
25
34
11
7
10
6
0.6
6
7
11
0.3
3
12
3
5
4
4
6
2.9
-
<1
0.4
5
0.7
1.8
4
<1
2
5
3
4
7
1.5
6
<1
0.2
0.4
0.3
0.7
1
0.3
-
—
0.1
0.2
0
0
-
0.1
41
75
34
50
37
39
61
21
Reference
She et al.
(2002)
Johnson-
Restrepo et
al. (2005)
Schecter et
al. (2003)
She et al.
(2007)
Mazdai et al.
(2003)
Mazdai et al.
(2003)
Sjodin et al.
(2004)
Bradman et
al. (2007)
3Year = year of sampling.
'TST = number of donors.
°Seven serum pools with number of donors in each serum pool ranging from 40-200.
In a study in New York City, adipose fat tissue samples (n = 52) were collected in 2003-
2004 from patients undergoing liposuction procedures (Johnson-Restrepo et al., 2005). BDE-47
was the major congener detected, followed by BDE-100 and -99. Median concentrations of
individual PBDE congeners and the sum of these PBDEs are given in Table 3-1 and are the
highest human levels reported so far. No significant difference was found in the concentrations
of PBDEs between genders. Concentrations of PBDEs were, on average, similar to those for
polychlorinated biphenyls (PCBs). PBDE concentrations did not increase with increasing age of
the subjects, whereas concentrations of PCBs increased with increasing age in males but not in
females. These results suggest differences between PBDEs and PCBs in their sources and/or
time course of exposure and disposition.
-------
In a study in Japan, 10 human adipose samples taken from the general Tokyo population
in 1970 and in 2000 were analyzed for BDE-28, -47, -99, -100, -153, -154, and -183. Median
concentrations of the sum of these PBDEs were 0.03 and 1.3 ng/g Iw in 1970 and 2000,
respectively. In 2000, median concentrations in ng/g Iw of PBDE congeners were, in decreasing
order, tetraBDE-47 (0.5), hexaBDE-153 (0.4), pentaBDE-100 (0.3), pentaBDE-99 (0.1),
hexaBDE-154 (0.06), and heptaBDE-183 (0.05) (Choi et al., 2003).
Tetra-, penta-, and hexaBDEs were analyzed in the adipose tissues from 3 women and
10 men between the ages of 28 and 83 and living in Spain for at least 10 years. The average
concentrations of tetraBDE-47, pentaBDE-99, and hexaBDE-153 were 1.4, 0.4, and 1.8 ng/g Iw,
respectively. An unidentified pentaBDE congener was found at an average concentration of
0.5 ng/g Iw. The predominant congener in both men and women in this study was hexaBDE-153
(Meneses et al., 1999).
3.2.1.2. Liver
In a Swedish study, paired samples of human liver and adipose tissue obtained at autopsy
from one woman (age 47) and four men (ages 66-83) were analyzed for nine tri- to hexaBDE
congeners. PBDEs were found in all samples. BDE-47, -99, and -153 were the predominant
PBDE congeners in both liver and adipose tissue. Generally, BDE-47 occurred at similar levels
in adipose tissue and liver (mean approximately 2.7 ng/g Iw). For the pentaBDEs, BDE-99 was
the predominant congener in both liver and adipose tissue, followed by BDE-100 and -85. The
mean total concentrations of these three pentaBDEs were higher in liver than in adipose tissue
and amounted to 4.3 and 1.7 ng/g Iw, respectively (Guvenius et al., 2001).
3.2.1.3. Human Milk
In a study conducted in 2002 of levels of PBDEs in human milk in the U.S., 47 samples
from Caucasian, African-American, and Hispanic nursing mothers 20-41 years of age and living
in Texas were analyzed for 13 PBDE congeners (Schecter et al., 2003). Mean and median total
concentrations of tri- through decaBDEs were 74 and 34 ng/g Iw, respectively. The maximum
and mean concentrations of BDE-99 were 111 and 14 ng/g Iw, respectively. Median
concentrations of individual PBDE congeners are given in Table 3-1. There was no correlation
between age and level of PBDEs in human milk.
Milk samples were collected in 2003 from 40 first-time mothers with 2- to 8-week-old
infants and residing in urban areas in the Pacific Northwest of the U.S. (Montana, Oregon, and
Washington State) and Canada (British Columbia) (She et al., 2007). Mean and median total
concentrations of 12 triBDE through decabromodiphenyl ether (decaBDE) congeners were
96 and 50 ng/g Iw, respectively. These values are substantially higher than the values reported in
the study of Schecter et al. (2003) and could be due to the fact that the mothers in the later study
-------
had been nursing for longer periods of time. BDE-47 was found at the highest level with median
concentration of 28 ng/g Iw, followed by BDE-99 (5.4 ng/g Iw), BDE-100 (5.3 ng/g lw), and
hexaBDE-153 (4.8 ng/g lw). Except for triBDE-28 with a median concentration of about 2 ng/g
lw, all other concentrations of PBDE congeners were <1 ng/g lw. In 7% of the samples,
hexaBDE-153 was the dominant congener. DecaBDE-209 with median concentration of 0.4
ng/g lw was a minor congener in breast milk, contributing 1.2% to the total PBDE concentration.
Breast milk was collected from 12 primiparous 24- to 33-year-old nursing women in
Japan, at one month after delivery. The most abundant PBDE congeners in human milk were
BDE-47 and the hexaBDE congener BDE-153, followed by pentaBDEs (BDE-99 and -100) and
triBDEs. The sum of the concentrations of tri- to hexaBDEs ranged from 0.7 to 2.8 ng/g lw.
There was a strong positive relationship between total PBDE levels in human milk and the
frequency offish consumption. The average total PBDE concentration (1.7 ng/g lw) in five
women representing the highest frequency offish consumption (every day) was double that
found in three women (0.8 ng/g lw) representing relatively low fish consumption (1-2
days/week) (Ohta et al., 2002).
In another study in Japan, PBDEs were not detected in eight pooled human milk samples
collected in 1973 (Akutsu et al., 2003). In 2000, BDE-47 was the predominant congener
(0.5 ng/g lw), followed by hexaBDEs (0.4 ng/g lw), pentaBDEs (0.3 ng/g lw), triBDEs
(0.1 ng/g lw), and heptaBDEs (0.04 ng/g lw). Of the pentaBDEs, BDE-100 was the most
abundant (0.17 ng/g lw), followed by BDE-99 (0.15 ng/g lw) and BDE-85 (0.01 ng/g lw) (Akutsu
et al., 2003). The relatively large concentration of hexaBDEs in mothers' milk seen in Japan
was explained by the authors to be due to past use in Japan of a hexaBDE commercial product
consisting mostly of BDE-153.
The breast milk concentrations of BDE-47, two pentaBDEs (BDE-99 and -100), and two
hexaBDEs (BDE-153 and -154) were determined in samples from 93 primiparous women,
collected from 1996-1999 in Uppsala County, Sweden. The women ranged in age from 20-
35 years. BDE-47 was the major congener (mean value 2.4 ng/g lw) and constituted 60% of the
mean concentration of PBDEs of 4.0 ng/g lw, followed by BDE-99 and -153 (0.6 ng/g lw each),
BDE-100 (0.4 ng/g lw), and BDE-154 (0.07 ng/g lw). No significant relationship was found
among breast milk concentrations of PBDEs and dietary intakes of PBDEs (through fish,
meat/poultry, dairy products, and egg consumption), age, body mass index, alcohol consumption,
or computer usage. After adjustments for these factors, a weak but significant association
between PBDE concentrations and smoking was observed. Time-trend analysis for samples
collected between 1996 and 2001 suggested a peak in BDE-47 and total PBDE concentrations
around 1998, followed by decreasing levels (Lind et al., 2003).
Pooled samples of breast milk collected at eight time periods between 1972 and 1997
from primiparous Swedish women were analyzed for tri- to hexaBDEs. In 1997, BDE-47 was
10
-------
the most abundant congener (2.3 ng/g Iw), followed by BDE-99 and -100 at 0.5 and 0.4 ng/g Iw,
respectively. The sum of the concentrations of PBDE congeners in human milk increased from
0.1 to 4.0 ng/g Iw during the 25-year period studied (Meironyte et al., 1999).
Levels of PBDEs found in breast milk in Japan and Sweden in comparable sampling
years were substantially lower than those found in the U.S. or Canada.
3.2.1.4. Blood
Levels of PBDEs in the blood are representative of either recent exposures or the slow
release of PBDEs from tissue stores. Seven tetra- to decaBDEs were analyzed in serum samples
collected in the U.S. in 1988 from male blood donors. The median serum concentration of the
sum of tetra- to decaBDEs was 1.6 ng/g Iw (Sjodin et al., 2001). In 2000-2002, the sum of the
median concentrations of six tetra- to hexaBDEs in serum pools collected in the U.S. was 61
ng/g Iw. PBDEs included in this study were tetraBDE-47; pentaBDE-85, -99, and -100; and
hexaBDE-153 and -154. Median concentrations of these individual PBDE congeners are given
in Table 3-1 (Sjodin et al., 2004).
Serum samples from 24 pregnant Mexican immigrant women living in an agricultural
community in California were collected during 1999 and 2001 (Bradman et al., 2007). Tetra-,
penta-, hexa-, and hepta-BDE congeners were measured in the serum samples. The median
concentration of the sum of tetra-, penta-, hexa-, and heptaBDE congeners was 21 ng/g Iw with a
median concentration for BDE-47, -99, -100, and -153 of 11, 2.9, 1.8, and 1.5 ng/g Iw,
respectively. There was no clear association between blood levels of PBDEs and demographic
characteristics, including age, lactation, and parity. There was a slight correlation between
number of years living in the U.S. and PBDE blood levels.
Concentrations of tetra-, penta-, hexa-, and decaBDE congeners were measured in serum
samples collected during 2004 from a family residing in Berkeley, California (35- and 36-year-
old father and mother, respectively, 5-year-old daughter, and 18-month-old son) (Fischer et al.,
2006). The 18-month-old was exclusively breast-fed for 6 months and was breast-feeding during
the study period. PBDE levels for the sum of the five lower brominated congeners BDE-47, -99,
-100, -153, and -154 were much higher in the infant (418 ng/g Iw) and child (247 ng/g Iw) than
in their parents (mother 106 ng/g Iw; father 64 ng/g Iw). BDE-47 was the predominant congener
for all ages, followed by hexaBDE-153, -100, and -99. Levels of BDE-209 in the infant
(233 ng/g Iw) and child (143 ng/g Iw) were unusually high compared with those in the parents
(mother 14 ng/g Iw; father 23 ng/g Iw). The authors suspected house dust and breast milk to
contribute appreciably to the child and infant exposures. However, no firm conclusions can be
drawn from this study given the small number of subjects investigated.
In Norway, pooled serum samples collected in 1998 from eight population groups of
different ages (0 to >60) and genders were analyzed for tri- to hexaBDEs. Total concentration of
11
-------
these PBDEs in men older than 60 years was 5.3 ng/g Iw, with tetraBDE-47 being the most
abundant congener (3.4 ng/g Iw), followed by hexaBDE-153 (0.6 ng/g Iw); pentaBDE-100,
pentaBDE-99, and hexaBDE-154 (all at approximately 0.4 ng/g Iw each); and triBDE-28
(0.1 ng/g Iw). The sum of the concentration of these PBDEs in serum was highest for the 0- to
4-year-old children (12 ng/g Iw) but was about one-third lower and relatively constant for the
different age groups above 4 years. Except for the 0- to 4-year-olds who seemed to experience
elevated exposure, there was a lack of an age-related trend of PBDE body burdens. This may be
explained by the fact that PBDEs are relatively new contaminants in the environment; the time
period for human exposure is therefore relatively short, and different age groups (except the 0- to
4-year-old group) may thus have experienced a similar exposure period (Thomsen et al., 2002).
The high level of PBDEs in the serum of the 0- to 4-year-olds could be due to higher exposure
from human milk and/or certain behavioral activity, such as crawling or sucking on flame-
retardant materials.
Concentrations of BDE-47, hexaBDEs (BDE-153 and -154), heptaBDE (BDE-183), and
decaBDE (BDE-209) were determined in blood serum from groups of 19-20 Swedish male and
female subjects in the following occupational groups: hospital workers (control), clerks working
full-time at computer screens, and personnel at an electronic-dismantling plant (Sjodin et al.,
1999). Commercial PBDEs used as flame retardants in the electronic industry are usually
decaBDE and to a lesser extent octabromodiphenyl ether (octaBDE). The median concentration
of BDE-47 in serum was about the same in the controls and computer clerks (-1.5 ng/g Iw) but
almost double that level in the electronic-dismantling personnel. Serum concentrations of all
PBDE congeners decreased in the electronic-dismantling workers after vacation. The median
decreases, standardized to 30 days of leave, were 14% for BDE-47, -153, and -154; 30% for
BDE-183; and 66% for BDE-209. These results indicate shorter half-lives of the more highly
brominated diphenyl ethers.
3.2.1.5. Placental Transport
Twelve paired samples of maternal and cord blood collected in 2001 from women in
Indiana were analyzed for tetra- to heptaBDE congeners (Mazdai et al., 2003). None of the
mothers had work-related potential for exposure to PBDEs and none smoked. Median
concentrations of the various PBDEs found in maternal and fetal sera are given in Tables 3-1 and
3-2 for comparison with a Swedish study (Guvenius et al., 2003) described below. TetraBDE-47
was the most abundant congener, followed by BDE-99 and BDE-100. PBDE concentrations
were highly correlated between mother and fetal sera, indicating that PBDEs cross the placenta
into the fetal circulation. In addition, the results indicate that all tetra- through hepta-substituted
congeners have approximately the same potential to cross the placenta. There was a decreasing
12
-------
trend in the concentration of PBDE congeners in maternal and fetal sera with increasing degree
of bromination.
Table 3-2. Median PBDE congener concentrations in maternal and fetal sera
in the U.S. and Sweden
PBDE congener
BDE-47
BDE-99
BDE-100
BDE-153
BDE-154
BDE-183
E PBDEs
Maternal serum (ng/g Iw)
Mazdai et al.
(2003)a
28
5.7
4.2
2.9
0.3
0
37
Guvenius et al.
(2003)b
0.8
0.2
0.2
0.6
0.04
0.06
2.1
Fetal serum (ng/g Iw)
Mazdai et al.
(2003)
25
7.1
4.1
4.4
0.7
0
39
Guvenius et al.
(2003)
1.0
0.07
0.07
0.17
0.01
0.01
1.7
"U.S.: year of sampling, 2001; number of donors = 12.
bSweden: year of sampling, 2000-2001; number of donors =15.
Samples of maternal and cord blood plasma were collected during 2000-2001 from
15 Swedish mothers (Guvenius et al., 2003). BDE-47 was the most abundant of all the
congeners, and comparable median concentrations were found in maternal and cord blood
plasma (Table 3-2). The levels of the higher brominated congeners, BDE-99 to -183, were
higher in maternal blood than in cord blood, indicating that the higher brominated PBDEs do not
pass through the placenta to the same extent as the lower brominated congener BDE-47. This
trend was not apparent in the Mazdai et al. (2003) study, where comparable levels were found in
maternal and fetal sera for all PBDE congeners studied. The concentrations of PBDEs found in
maternal and fetal blood samples in Indiana women (Mazdai et al., 2003) were substantially
higher than those found in Swedish women (Guvenius et al., 2003).
In summary, median concentrations of PBDE congeners in the U.S. are available for
human adipose tissue (Johnson-Restrepo et al., 2005; She et al., 2002), human milk (She et al.,
2007; Schecter et al., 2003), and serum (Bradman et al., 2007; Sjodin et al., 2004; Mazdai et al.,
2003). The concentration profiles in the U.S. of PBDEs in adipose tissue, serum, and human
milk are similar, although these studies were conducted in different regions of the U.S. (Table
3-1). The predominant congener found in adipose tissue, human milk, and blood samples in the
U.S. is tetraBDE-47, followed by pentaBDE-99 and -100, with current median concentrations in
human biological samples of approximately 25, 7, and 4 ng/g Iw, respectively. Few
measurements have been made of other PBDE congeners, such as triBDE and heptaBDE to
decaBDE. Median concentrations of the sum of PBDEs measured in human biological media
are about 40 ng/g Iw. These levels are substantially higher than the levels found in human
populations in Europe or Japan.
13
-------
3.2.2. Animal Data
The animal data on BDE-99 distribution are limited but more quantitative than the
human data because they represent the distribution after deliberate dosing studies.
In a study by Hakk et al. (2002a), tissue distribution of BDE-99 was assessed in young
adult male Sprague-Dawley rats. 14C-BDE-99 (>98% purity) in corn oil was given as a single
oral dose of 8 mg/kg (1.0 uCi/rat) to groups of conventional (three/group) and bile-duct-
cannulated (five/group) rats. Adipose tissue, adrenals, blood, carcass, gastrointestinal tract,
heart, kidney, liver, lung, testes, and thymus were analyzed for radioactivity on day 3 after
exposure. In the conventional rat, BDE-99 was preferentially found in lipophilic tissues, with
39% of the administered dose being found in the carcass, 6% in the gastrointestinal tract, and 4%
in adipose tissue. No other tissues in the conventional or bile-duct-cannulated rats contained
greater than 1% of the 14C on day 3. The tissue data support the hypothesis that bile salts are
necessary for the intestinal absorption of BDE-99, since tissue levels of 14C for the bile-duct-
cannulated rats were much lower when compared with those of conventional rats (2% in the
carcass, 1.5% in the gastrointestinal tract, and 0.8% in adipose tissue).
The remaining carcasses from conventional rats were fractionated into skin, bone, brain,
eyes, and muscle. An estimated 21% of the BDE-99 dose was found in the skin. When the
tissue distribution data were expressed on a concentration basis, the most lipophilic tissues
(adipose, adrenal, gastrointestinal tract, and skin) contained the highest concentrations of 14C.
The lipid content of selected tissues was determined, but the observed distribution pattern for
BDE-99 did not consistently correlate with tissue lipid content. Adipose tissue had the highest
lipid content and the highest BDE-99 concentration. However, kidney and lung with higher lipid
content than liver had lower concentrations of BDE-99 than liver, an indication that selectivity in
hepatic retention was not occurring.
In the study by Chen et al. (2006), the tissues that contained the largest portion of the
radiolabel from a 0.6 mg/kg-day oral dose in both F344 rats and B6C3F1 mice were the adipose
deposits, muscle, skin, and liver. Adipose tissue contained the highest percentage of the dose in
rats and mice. In rats, the next-to-the-highest percentage was in the skin, while in mice it was
the muscle tissue. All other tissues evaluated contained less than 1% of the dose. Following i .v.
injection of 1 mg/kg uniformly 14C-labeled BDE-99 (1.5 Ci) to C57BL/6 mice, Staskal et al.
(2006) found that the percentage in muscle was greater than that in skin 5 days after exposure.
This is in agreement with the Chen et al. (2006) finding for mice and may suggest a species
difference in distribution.
After oral dosing, the radiolabel in adipose tissues as well as in kidney and lung (Hakk et
al., 2002a) appeared to be totally the parent compound. The apparent requirement for bile in
absorption and the extrahepatic tissue data indicating the presence of parent compound rather
14
-------
than hydroxylated metabolites in peripheral tissues could be interpreted as suggesting initial
postabsorption distribution of a substantial portion of the parent compound by way of the
chylomicrons. This absorption route would explain the presence of unmetabolized parent
compound in the adipose and other high-affinity tissues. Hakk et al. (2002a) found there was
some binding of nonextractable metabolites to proteins in the liver. The lower extractability of
the label from the liver was presumably the result of binding to cellular biomolecules.
Chen et al. (2006) compared tissue levels of 14C-labeled BDE-99 (95.6% purity;
35.6 mCi/mmol) in male F344 rats at 24 hours after a single exposure to 0.6 or 6 mg/kg with the
tissue levels at 24 hours after 10 days of daily exposure to 0.6 mg/kg (a total of 6 mg/kg). For all
tissues except adipose tissue, the single 6 mg/kg dose resulted in a higher concentration
(nmol-eq/g) than the 10 single 0.6 mg/kg-day doses. In adipose tissues, the level of label
24 hours after the last of the 10 0.6 mg/kg-day single doses was greater than that after the single
6 mg/kg dose, illustrating the potential for accumulation of BDE-99 in body lipids.
Hakk et al. (2006) conducted a study comparable to their 2002 study of BDE-99 in
Sprague-Dawley rats, using BDE-100. About 73% of the radiolabel (0.9 uCi/rat) remained in the
body of conventional rats after 72 hours and was found in the adipose tissue, gastrointestinal
tract, skin, liver, and lungs. The other tissues evaluated contained less than 0.1% of the label
after 72 hours. The results with BDE-100 were similar to those from the study of BDE-99.
Eriksson et al. (2002) demonstrated that radiolabeled BDE-99 can be taken up and
retained in the neonatal mouse brain. Male NMRI mice (five/group) were administered 8 mg/kg
of 14C-BDE-99 (purity >98%; 40.5 uCi/kg) in a fat emulsion on PND 3, 10, or 19 and were
sacrificed 24 hours or 7 days after administration. The amount of radioactivity in the brain was
between 0.4 and 0.5% of the administered dose, 24 hours after administration. Seven days after
the administration, 14C-BDE-99 (or its metabolites) could still be detected in the brain,
decreasing to between 0.1 and 0.3% of the administered dose in mice exposed on PND 3, 10, or
19. The amount of radioactivity in the brain was similar in mice exposed on PND 3 or 10
compared with mice exposed on PND 19 and therefore does not appear to explain the different
behavioral effects seen in adult mice exposed to BDE-99 on PND 3, 10, or 19 (see section
4.3.1.5).
The overall qualitative distribution of 14C-labeled pentaBDE-85 and -99 was studied in
C57BL mice by using whole-body autoradiography (Darnerud and Risberg, 2006). 14C-BDE-85
or -99 (>95% purity) was administered to male and female C57BL mice by i.v. injection or by
gavage at 20 jimol/kg (25 |iCi/g body weight; ~11 mg/kg). The animals were sacrificed at time
intervals varying from 1 hour to 16 days after administration. The distribution of radioactivity in
mice after i.v. administration was characterized by a high initial uptake of radioactivity in fatty
tissues. In addition, the liver, adrenal cortex, lung, ovaries, and nasal epithelium accumulated
radioactivity. Initially, intermediate radioactivity levels were found also in the brain tissue. No
15
-------
radioactivity was observed in the thyroid gland. At 4 and 16 days after the administration, the
radioactivity concentration was weaker, indicating significant 14C excretion. In the male mouse
after 6 hours, the concentration of radioactivity in the testis was low; in females, labeling in the
ovaries was localized to the follicular structure. At 16 days postinjection, labeling was still
visible in the fat tissues, liver, lung, and adrenal cortex; elimination from the lungs seemed to be
slower than from the liver. Some faint labeling remained in the brain. The distribution pattern
after oral administration of BDE-85 and -99 was similar to what was found after i.v. injection,
which showed that the gastrointestinal uptake is effective. In spite of the lipophilic nature of
these PBDEs, retention in the body fat depot was only moderate, probably because substantial
metabolism and/or excretion occurred in mice.
The qualitative distribution of 14C-BDE-85 and -99 was also studied in pregnant mice
sacrificed 1 day after i.v. administration of 11 mg/kg of 14C-BDE-85 or -99 on gestational days
(GDs) 16-17 or 4 days after i.v. administration on GDs 13-17. In general, the uptake of
radioactivity observed in pregnant mice was comparable to that in nonpregnant mice.
Radiolabel was observed in the membranes surrounding the fetus, and labeling of fetal liver and
intestinal contents was higher than in surrounding tissues. Faint radiolabeling was observed in
the fetal brain.
Darnerud and Risberg (2006) also studied the partition of 14C-BDE-85 and 14C-BDE-99
to maternal milk in lactating C57BL mice. The lactating dams were injected intravenously with
2.0 |imol/kg (1 mg/kg) of each of the pentaBDE congeners on day 11 postpartum. Quantitative
measurements were made of 14C-BDE-85 and 14C-BDE-99 radioactivity in the liver, kidney, fat,
and plasma from lactating dams and their offspring on day 12 postpartum. No significant
differences in 14C levels of the studied tissues were observed between the two congeners. In the
dams, fat contained about 10 times as much 14C as did the liver, and the liver had higher
14C concentrations than both the kidney and the plasma. In the offspring, liver and kidney
radioactivity levels were similar to what was found in corresponding tissues from dams, whereas
a two to four times higher plasma concentration was found in offspring.
Radioactivity was also measured in breast milk of lactating dams 1 and 4 days after i.v.
administration of 1 mg/kg of 14C-BDE-85 or -99 at day 11 postpartum (Darnerud and Risberg,
2006). Breast milk collected on days 12 and 15 postpartum contained a substantial amount of
radioactivity, the levels decreasing with time after administration. Results from this distribution
study indicate that, in spite of the lipophilic nature of these PBDEs, retention in the body fat
depot is only moderate, probably because substantial metabolism and/or excretion occur in mice.
Breast milk transport of these pentaBDE congeners was substantial, and radioactivity was found
in the milk and in tissues of the suckling offspring 4 days after administration of a single dose of
radioactive pentaBDEs to the dam. However, neonatal excretion seems to prevent accumulation
and high levels in offspring compared with maternal levels.
16
-------
Adult male and female offspring of pregnant Long-Evans rats exposed to BDE-99 (purity
>99%) by nine subcutaneous (s.c.) injection (1 or 10 mg/kg-day) on GDs 10-18 were found to
have detectable parent compound in the brain, plasma, and adipose tissues 120 days after birth.
For the 1 mg/kg-day dose, the level in the adipose tissue was about 400 times greater than that in
plasma. For the 10 mg/kg dose, the level in adipose tissue was about 1,500 times greater than
that in plasma. There was considerable variability in the levels found in the adipose tissue
among the individual samples analyzed (Ceccatelli et al., 2006).
Darnerud et al. (2005) examined whether an infection of Coxsackie virus B3 (CBS), a
common human virus, changes tissue distribution of 14C-BDE-99. On day 0, adult female Balb/c
mice were infected with CBS; on day 1 of the infection, they were dosed orally with 0.2 mg/kg
of 14C-BDE-99; and on day 3 of the infection, they were sacrificed for studies of 14C-BDE-99
distribution. Clinical signs of disease started to appear on day 2. In comparison with control
values, there was no change in distribution of 14C-BDE-99 in the brain, heart, spleen, kidney,
blood, or thymus. However, 14C-BDE-99 concentrations were increased in the liver (186%) and
decreased in the lung (47%) and pancreas (51%). This correlated with decreased enzyme
activities of CYP-450-mediated ethoxyresorufin O-dealkylase (EROD) and pentoxyresorufin
O-dealkylase (PROD) in mice (see section 3.3), possibly suggesting decreased metabolism of
BDE-99 and increased retention of unmetabolized compound by the liver during the acute phase
of the viral infection.
In the study of Branch! et al. (2005), BDE-99 at 0 or 18 mg/kg-day was administered to
CD-I Swiss mice (nine/group) from GD 6 to PND 21. Two modes of administration of BDE-99
were investigated: by gavage or by letting the mouse drink BDE-99 dissolved in corn oil from a
syringe (see also section 4.3.1.7). The brains collected from treated male pups on PND 22 (n = 2
per group) showed significantly elevated BDE-99 concentration: 640 ug/kg compared to
<5 ug/kg for controls. No effects of administration routes were seen in the concentration of
BDE-99 in the brain of treated mice.
Kodavanti et al. (2005) carried out a study using cultures of cerebellar granule cells from
7- to 8-day-old Long-Evans rat pups. The cultures were treated with 14C-labeled PBDE-99
(0.05 uCi/mL) combined with different concentrations of unlabeled compound (0-30 uM) for
15 minutes to 1 hour. For each concentration tested there was a linear increase in percent uptake
of the label from the culture medium over the 1-hour exposure period. When time was held
constant and concentration varied, the percent accumulation increased linearly for the 0.67 and
3.67uM concentrations at 15, 30, and 60 minutes and then decreased linearly between the
3.67 and 30.67 uM concentrations, suggesting saturation of uptake.
17
-------
3.3. METABOLISM
BDE-99 and -100 appear to be the most extensively metabolized of the individual PBDE
congeners. Staskal et al. (2006) estimated that about 70% of the BDE-99 in urine and 80% in
fecal matter were present as metabolites in mice, following i.v. injection of a 1 mg/kg dose. For
BDE-100 the percentages were similar. When comparing urinary excretion and tissue loads
across PBDE congeners, there appeared to be an inverse relationship (Staskal et al., 2006): tissue
retention was higher when urinary excretion was low. Urinary excretion in male mice appears to
exceed that in male rats (Chen et al., 2006).
Data on the metabolites present in bile and excreta in rats from the Hakk et al. (2006,
2002a) and Chen et al. (2006) studies and in mice from the Chen et al. (2006) and Staskal et al.
(2006) studies support a metabolic pathway that involves epoxidation as the initial reaction for
BDE-99 and -100. This can be followed by debromination in some cases and possibly by
conjugation of the hydroxylated derivatives with GSH, glucuronate, and/or sulfate. Hydrolysis of
the ether linkage between the two phenyl rings may also occur, producing brominated phenols.
There is consistency among these studies of metabolites with regard to the production of mono-
and dihydroxylated metabolites as well as hydroxylated/debrominated metabolites. The data for
conjugation with GSH, glucuronate, and/or sulfate are more variable.
Chen et al. (2006) have proposed a metabolic pathway for BDE-99 (Figure 3-1) that is
consistent with much of their data as well as those from the Staskal et al. (2006) and Hakk et al.
(2006, 2002a) studies. There are some differences among the studies that relate to the extent to
which metabolism occurs for the different congeners (BDE-99 and -100) and in different
species.
According to the proposed metabolic pathway, BDE-99 can undergo two primary
reactions (Figure 3-1). Both involve CYP-450 epoxidation of the phenyl ring. In one case,
epoxidation targets carbons 5 and 6 of the disubstituted phenyl ring. This reaction favors
BDE-99 to a greater extent than BDE-100 because there is less steric hindrance of unsubstituted
carbon 6 of the disubstituted ring by the bromines on the trisubstituted phenyl ring in BDE-99.
The 5,6-epoxy intermediate can undergo three subsequent reactions as follows (Figure 3-1):
(1) rearrangement of the epoxy by way of a 1,2-hydride shift, (2) protonation of the epoxide
generating a carbocation intermediate, and (3) conjugation with GSH.
The 1,2-hydride shift generates monohydroxylated pentaBDEs with the hydroxyl group
located on the dibrominated phenyl ring. Hydrolysis of the epoxide, although not proposed as
part of the Chen et al. (2006) pathway, could also occur and would produce dihydroxylated
pentabromo metabolites.
Protonation of the 5,6-epoxide would lead to an unstable carbocation with the capability
of accepting an electron pair from a nucleophile, providing an additional route for formation of
dihydroxypentabromodiphenyl ethers and for protein binding in the liver. A
18
-------
1,2-dihydroxypentabromodiphenyl ether formed via this route could fragment, generating
3,5-dibromo-l,2-dihydroxybenzene and 2,4,5-tribromophenol (Figure 3-1). Chen et al. (2006)
tentatively identified 2,4,5-tribromophenol in rat feces and its sulfate and glucuronate conjugates
in urine.
Br
OH Br OGluc Br
UDPGA ^JssvJJ-s^Ss*. + «tleiBt two otrwrmono-hydroityl Bled
Br
1mm®
o es "
or or
Br
tetrabramoelphenyl ether
glueuronides excreted in toil©
Br Br
Br'
Br
Br
BDEti
Br Br Br Br Br Br
UDP6*
r Br'
OH
JP4503
Br Br
,Ck
»
shift
F4
Br Br
FS
Br Hr
Br
B7
Br Br
Gluoir
B7
Br
GSH
Br
H*
Br
GSH
SS Br
32
I2O
Br Br
Br y^ y^ Br
SG Br
BS
OH Br
13
|-H20
Br Br
SB
Br
IS0
Ir
Br
GlycO,
Br
ODPGA.
Br
Br
Br
U1.B1
Br
Br
PAP8
Br
U3.F1
Br
Br
ua
Figure 3-1. Proposed metabolic pathway for BDE-99 in male rats.
Note: PAPS = 3'-phosphoadenosine-5'-phosphosulfate; UDPGA = undine diphosphate glucuronic
acid.
Source: Chenetal. (2006).
Conjugates of metabolites with GSH have been identified as potential metabolites in bile
and thereby in fecal matter by Hakk et al. (2002a) and Chen et al. (2006). GSH conjugates could
be modified through the activity of y-glutamyl transpeptidase, carboxypeptidase, and cysteine
P4yase to produce the thiols occasionally identified in urine (Hakk et al., 2002a).
In the alternate primary pathway, the epoxide would form between a brominated and
nonbrominated carbon (Figure 3-1). Chen et al. (2006) proposed that this would lead to
formation of tetrabromo-hydroxylated metabolites. Any brominated carbon with a neighboring
19
-------
nonbrominated carbon is liable to this reaction, permitting generation of several tetrabromo-
hydroxylated derivatives. Tetrabromo-hydroxylated metabolites have been identified in rats for
both BDE-99 and -100 by Hakk et al. (2006, 2002a) and in mice by Staskal et al. (2006).
Additional research is needed to support the metabolic pathway proposed by Chen et al.
(2006). The primary steps involving the formation of arene oxide intermediates, leading to
hydroxylated and hydroxylated/debrominated metabolites, are consistent with the available data
on metabolites in feces and urine of rats and mice. Support for the mechanisms of reaction and
the extent and variability in conjugate formation and protein binding are fertile areas for
additional research.
There have been no whole animal studies of the CYP-450 isozymes that participate in the
epoxidation of BDE-99 or -100, but involvement of the CYP-1 Al/2 and -2B isozymes has been
investigated in vitro with BDE-99 because of their link to the activation of the aryl hydrocarbon
(Ah) receptor (Sanders et al., 2005; Chen and Bunce, 2003; Chen et al., 2001). The activity of
CYP-1 Al/2 and -2B is generally evaluated through analysis of the phase I enzymes EROD for
CYP-1 Al/2 activity and PROD for CYP-2B activity. In the study by Darnerud et al. (2005) of
the effects of a viral infection on the distribution of 14C-BDE-99 (see section 3.2.2), both EROD
and PROD were active in the noninfected control adult female Balb/c mice treated with
14C-BDE-99, EROD to a greater extent than PROD. On day 3, in the infected mice treated with
14C-BDE-99, the enzyme activities of EROD and PROD were about 17 and 31%, respectively, of
those in the noninfected 14C-BDE-99-treated mice.
Sanders et al. (2005) used a different approach for measuring the induction of the
CYP-450 isozymes in the liver. Male F344 rats were treated with 0, 0.57, 5.7, or 57 mg/kg-day
BDE-99 for 3 consecutive days and sacrificed 24 hours after the last dose. Messenger
ribonucleic acid (mRNA) was isolated from a portion of the right medial lobe of the liver and
converted to its complementary deoxyribonucleic acid (cDNA) by using real-time polymerase
chain reaction (PCR). Target gene amplification was evaluated by using specific probes for
CYP-1A1, -2B, and -3A. These analyses indicated that CYP-1A1 expression was significantly
up-regulated (eightfold) only with the 57 mg/kg-day dose of BDE-99. On the other hand,
BDE-99 doses up-regulated expression of CYP-2B in a dose-related fashion at the 5.7 and
57 mg/kg doses in one assay and at the highest dose in another assay. At the highest dose,
CYP-2B mRNA levels were 14-25 times those in the corn oil controls. These data conflict with
the data from the control mice from Darnerud et al. (2005), where EROD appeared to be up-
regulated to a greater extent than PROD. The expression of CYP-3 A was up-regulated (four- to
fivefold) with the highest dose.
C57BL/6 (10 weeks of age) mice were injected with doses of 0, 10, or 100 umol/kg (0,
5.7, or 57 mg/kg) BDE-99 in corn oil for 4 days in a study by Pacyniak et al. (2007). The livers
were removed 24 hours after the last dose and the levels of mRNA measured by Northern blot
20
-------
and branched DNA (bDNA) analyses. The bDNA was considered to be the more accurate of the
two assay systems. Northern blot analysis indicated that the levels of CYP-2B10 were induced 4
and 74 times, respectively, at the two doses tested, while the bDNA results indicated 51- and
38.9-fold inductions. CYP-3A11 did not show a difference with respect to the dose administered
but was induced fivefold by the Northern blot analysis and 1.8-fold by the bDNA analysis.
Kester et al. (2002) evaluated whether or not the human estrogen sulfotransferase and the
human phenol sulfotransferase were able to conjugate sulfate from 3'-phosphoadenosine-5'-
phosphosulfate to several hydroxylated PBDEs. The highest degree of sulfation was observed
with the tetraBDE hydroxy congener with both enzymes; the values for the pentaBDE hydroxy
congener fell between those for the tested tri- and tetraBDE hydroxy congeners. Estrogen
sulfotransferase was approximately 20 times more active toward 4-OH-3,5,2',4',6'-pentaBDE
than was phenol sulfotransferase. Studies of excreted pentaBDE metabolites do not indicate that
sulfate conjugation is a major metabolic process.
3.4. ELIMINATION
Elimination of BDE-99 and -100 occurs by way of the fecal matter and urine. Urinary
excretion appears to be more substantial in mice than rats. Both parent compound and
metabolites are identified in the excreta; the amounts of each vary among the studies.
Radiolabel in fecal matter represents absorbed and unabsorbed material. Bile is an important
contributor to the radiolabel in fecal matter.
In the study by Hakk et al. (2002a), a single oral dose of 8 mg/kg 14C-BDE-99
(1.0 uCi/rat) was given to conventional and bile-duct-cannulated male Sprague-Dawley rats.
Parent BDE-99 compound and metabolites were analyzed in urine, feces, and bile collected at
daily intervals for 3 days. Excretion in urine was low and amounted to 0.9% of the dose in the
conventional rat and 0.4% in the bile-duct-cannulated rat within 3 days; biliary elimination was
only 4% over the same period of time. Excretion in both urine and bile of bile-duct-cannulated
rats peaked at 1-2 days after exposure. Feces were the major route of elimination of BDE-99.
Approximately 43% of the dose in conventional rats and 87% in bile-duct-cannulated rats were
found in the feces within 3 days. In both conventional and bile-duct-cannulated rats, fecal
radioactivity was highest on the first day (22 and 53%, respectively) and then declined steadily
thereafter, suggesting that absorption is increased by bile emulsion and that enterohepatic
circulation of BDE-99, if it occurs, plays a minor role in the male rat.
The Hakk et al. (2006) study of 14C-BDE-100 in male Sprague-Dawley rats found only
0.1% of the dose in urine of the conventional rat at 72 hours. This is lower than the 0.9%
observed with BDE-99. The bile contained about 1.7% of the radiolabel after 72 hours, which is
also lower than that observed with BDE-99 in the study described above (Hakk et al., 2002a).
There was evidence of glucuronidation of some of the biliary material but no evidence of sulfur-
21
-------
containing metabolites. Most of the biliary label was protein bound. Approximately 20-30% of
the extractable fecal radiolabel was unmetabolized in the conventional and bile-duct-cannulated
rats. Large amounts of label were not extractable.
In F344 rats receiving an oral dose of 0.6 mg/kg BDE-99, 43% of the dose was present in
the feces on day 1; the cumulative amount excreted in feces was 56% by day 10 after exposure.
Urinary excretion was 1.6% on day 1 and 2.9% by day 10 (Chen et al., 2006). Over the 10-day
period, about half of the single dose had been excreted; the remainder was retained by the
tissues.
Biliary and fecal excretion appear to be important in C57BL mice, based on the results
from the whole-body autoradiography study by Darnerud and Risberg (2006). Radiolabel was
observed in both the bile and intestinal contents after both oral and i.v. administration of
10 mg/kg BDE-99 or -85. Intestinal radiolabel was more intense for the oral route of exposure,
suggesting the presence of unabsorbed material in the feces as well as radiolabel derived from
bile.
Hakk et al. (2002b) published an extension of their study of BDE-99 in male Sprague-
Dawley rats to determine whether BDE-99 can bind to endogenous carrier proteins in the urine
and bile, either as the parent compound or as metabolites. Such binding may facilitate the
elimination of lipophilic xenobiotics. Because of the low amount of 14C in urine in the
conventional- and cannulated-treatment groups, urine collected from each rat group was pooled
over 3 days. Bile from cannulated rats was pooled on a daily basis. Chromatographic analysis
of urine revealed that the majority (76%) of the 14C in conventional rat urine was not associated
with protein, while 7% was bound to an 18-kDa monomeric protein characterized as
a2u-globulin. In the cannulated rat urine, none of the 14C was bound to protein (100% unbound).
Presumably, BDE-99 metabolites formed in cannulated rat urine were sufficiently polar and did
not require a carrier system for excretion via the urine.
The pooled bile sample 1 day after dosing indicated that 61% of the biliary 14C was
unbound, decreasing steadily to 43% by day 3 after dosing (Hakk et al., 2002b). Approximately
28% of the biliary 14C was associated with a 79-kDa protein, increasing steadily to 47% by day
3. Extractability of the bound radioactivity from rat bile protein ranged from 27-85%. In the
day 1 bile sample, 21% of bound, extractable radioactivity was parent compound and the
remainder was polar metabolites. At days 2 and 3, only metabolites were observed to be bound
to the 79-kDa protein. Metabolite identification was not possible. The authors concluded that,
although this study demonstrated the ability of BDE-99 and/or its metabolites to associate or
tightly bind with urine proteins, it was unknown whether BDE-99 exposure could lead to
nephropathy in rats.
In mice, urinary excretion of BDE-99 and -100 appears to involve binding to a major
urinary protein (MUP) (Staskal et al., 2006). These proteins are synthesized in the liver, secreted
22
-------
into serum, and eliminated in urine. Male mice secrete more protein than females. Analysis of
pooled urine samples from BDE-99 and -100 intravenously dosed female mice indicated that
59.6 and 55.1%, respectively, was protein bound to an MUP. The two congeners appeared to
bind to different MUP isoforms, with BDE-99 binding to MUP-1 and BDE-100 binding to
MUP-1 andMUP-3.
3.4.1. Half-life Determinations
In the study by Hakk et al. (2002a), BDE-99 preferentially deposited in lipophilic tissues.
BDE-99 was slowly mobilized from skin and fat deposits. In the conventional rat, an estimated
21% of the dose was deposited in skin at 3 days after exposure. At 6 days from exposure, over
18% of the dose to the conventional rat was still in the skin, declining to 12% by day 12. A
substantial portion of the dose in the conventional rat remained in adipose tissue, 14 and 10% for
the 6- and 12-day rats, respectively. Based on the disposition and excretion results obtained in
this study, the estimated whole-body half-life of BDE-99 in male Sprague-Dawley rats was about
6 days, indicating that BDE-99 has bioaccumulation potential.
3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS
Limited information is available on the absorption, distribution, metabolism, and
elimination of BDE-99 in experimental animals and in humans. A model for human metabolism
has not been established. Extrapolation of results from laboratory animals to humans by using
physiologically based pharmacokinetic models is not possible at this time.
23
-------
4. HAZARD IDENTIFICATION
4.1. STUDIES IN HUMANS—EPIDEMIOLOGY, CASE REPORTS, CLINICAL
CONTROLS
Epidemiological studies of BDE-99 are not available.
To assess whether PBDEs may be detrimental to neurodevelopment, Mazdai et al. (2003)
determined concentrations of PBDEs and total and free serum thyroxine (T4) and
triiodothyronine (T3) in human fetal and maternal sera (see also section 3.2.1). Twelve paired
maternal and cord blood samples were obtained from women 18-37 years old, presenting in
labor at an Indiana hospital. The PBDE congeners and their concentrations measured in fetal
and maternal serum samples are given in Table 3-2. There was no relationship between infant
birth weight and PBDE concentrations. No birth defects were documented. Thyroid hormones
were assayed in 9 of the 12 sample pairs. There was no correlation among total PBDEs and T3
or T4 concentrations (total or free). The authors cautioned that the sample size may have been
too small to detect an association between serum concentrations of PBDEs and thyroid hormone
levels.
In the study of PBDE levels in breast adipose tissue of 23 California women, described in
section 3.2.1 (She et al., 2002), there was no correlation between total concentrations of tetra- to
hexaBDE in breast adipose tissues and disease status (malignancies, benign tumors, or ductal
carcinomas in situ).
In summary, the available limited human studies do not permit any conclusions to be
made concerning a possible association between exposure to PBDEs or BDE-99 and adverse
health outcome in humans.
4.2. SHORT-TERM, SUBCHRONIC, AND CHRONIC STUDIES AND CANCER
BIOASSAYS IN ANIMALS—ORAL AND INHALATION
Short-term, subchronic, or chronic inhalation toxicological studies of BDE-99 are not
available.
4.2.1. Short-term and Subchronic Studies
4.2.1.1. Mice
The aim of the study by Skarman et al. (2005) was to determine the effects on plasma T4
levels and hepatic enzyme activities in juvenile mice following maternal gestational and
lactational exposure to BDE-99. Groups of 22 or 13 dams received 0 or 0.08 mmol/kg
(45 mg/kg) of BDE-99 (>99% purity) in corn oil every third day from GD 4 through PND 17, on
a total of 10 occasions. The total dose of BDE-99 administered was, therefore, 0.8 mmol/kg or
24
-------
450 mg/kg. Parallel groups of dams were similarly treated with a total dose of 0.8 mmol/kg
(-450 mg/kg) of a commercial pentaBDE (Bromkal 70-5DE with main constituents of 37%
BDE-99 and 35% tetraBDE-47) or with a total dose of 0.8 mmol/kg (260 mg/kg) Aroclor 1254.
On GD 17, four dams from each treatment group were sacrificed and liver and blood samples
collected. On PND 3, the size of the litters was adjusted to 10 pups. On PNDs 11, 18, and 37,
three to four pups from each litter were sacrificed and liver and plasma samples collected.
Dam and offspring body weights were not affected by BDE-99, Bromkal, or Aroclor
treatment. Significantly increased liver-to-body-weight ratio was seen on PND 20 in dams
treated with BDE-99 but not in their offspring on PND 11, 18, or 37. Pregnancy rate, gestation
length, and litter size were not statistically different from controls. Plasma total and free T4 in
the pregnant dams on GD 17, in the postweaning dams on PND 20, and in the offspring on PNDs
11, 18, and 37 were unaffected by BDE-99 treatment. On the other hand, plasma total and free
T4 were significantly reduced in the offspring of the Bromkal groups on PND 11 but returned to
control levels by PND 18.
Hepatic microsomal CYP-450 enzyme activity was measured by means of the EROD
activity assay, a marker of CYP-1A1 activity. Hepatic EROD activity in pregnant dams sampled
on GD 17 and in postweaning dams sampled on PND 20 was unaffected by treatment with
BDE-99. Induced EROD activity was seen in the Bromkal group in dams on GD 17 but returned
to control levels on PND 20, while Aroclor treatment increased EROD activity in the dams on
GD 17 and PND 20. Offspring sampled on PNDs 11 and 18 showed increased hepatic EROD
activity in all treatment groups relative to controls but returned to control levels by PND 37. The
increase in EROD activity was highest for the Aroclor group, while the increase was similar for
the Bromkal and BDE-99 groups.
Hepatic uridine diphosphoglucuronosyl transferase (UDPGT) activity was studied in
offspring on PNDs 11 and 18. UDPGT activity was not different from that in controls in the
BDE-99 group at both time points. A significant increase in UDPGT activity was observed in
the Aroclor group at both time points, while the Bromkal group showed an increase in enzyme
activity of borderline significance on PND 18 only.
Based on the above, the study of Skarman et al. (2005) shows that BDE-99 had no effect
on plasma T4 levels in dams and their offspring relative to controls at any sampling occasion,
suggesting that other components in Bromkal are responsible for the reduction of T4 levels in
offspring on PND 11. One of the components of Bromkal is tetraBDE-47, which has been
shown to cause a decrease in T4 levels in mice and rats (Hallgren and Darnerud, 2002; Hallgren
et al., 2001). These results indicate that interference with thyroid hormone homeostasis can vary
significantly between PBDE homologs.
25
-------
4.2.1.2. Rats
Hakk et al. (2002a) examined the effect of BDE-99 on total T4 plasma levels in young
adult male Sprague-Dawley rats. A single oral dose of 8 mg/kg of 14C-BDE-99 (>98% purity) in
corn oil was given to groups of conventional (three/group) and bile-duct-cannulated (five/group)
rats. The rats were housed in steel metabolism cages and sacrificed 3, 6, or 12 days after
exposure. Average total plasma T4 concentration (bound and free) was 1.7 |ig/dL in the control
rats. In the treated conventional rats, the average total T4 levels increased approximately twofold
to 3.2 |ig/dL at 3 days after exposure, remained elevated at 3.0 |ig/dL at 6 days after exposure,
but by day 12 returned to control levels at 1.9 |ig/dL. TetraBDE-47, which has been shown to
cause a decrease in T4 levels in mice and rats (Hallgren and Darnerud, 2002; Hallgren et al.,
2001), may be the component responsible for the reduction of T4 levels seen with commercial
pentaBDE mixtures (Skarman et al., 2005; Zhou et al., 2002).
4.2.2. Chronic Studies and Cancer Bioassays
Chronic toxicity/carcinogenicity studies of BDE-99 are not available.
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES
4.3.1. Mice
4.3.1.1. Eriksson et al. (2001)
This study was carried out to determine whether exposure to BDE-99 during the period
of rapid brain growth in neonatal mice could lead to disruption of the adult brain function.
Single doses of 0, 0.8, or 12 mg/kg BDE-99 (>98% purity) in a 20% fat emulsion (1:10 egg
lecithin to peanut oil in water) were administered by gavage to NMRI male mice on PND 10.
Mice serving as controls received 10 mL/kg of the 20% vehicle. Spontaneous motor behavior
tests (locomotion, rearing, and total activity) were measured over three 20-minute periods, at
ages 2 and 4 months, in groups of eight mice randomly selected from three to four different
litters, and the mice were tested once only. Habituation capability was evaluated in 2- and 4-
month-old mice. Habituation is defined as the ability of the animals to adapt to a new
environment and is characterized as initial investigation and exploration of their surroundings
followed by gradual aclimitization and acceptance of the new area. It was evaluated in terms of
the ratio of the motor behavior measures from the 40- to 60-minute observation period divided
by the measures from the 0-20 minute period and multiplied by 100 (habituation ratio). Swim
maze performance, a measure of learning and memory ability, was tested in groups of 16-18
mice, at age 5 months, given the high dose of BDE-99 (12 mg/kg).
There were no clinical signs of dysfunction throughout the experimental period nor any
significant deviations in body-weight gain in the BDE-99-treated mice compared with the
vehicle-treated mice. The spontaneous motor behavior data showed, for all three variables
26
-------
(locomotion, rearing, and total activity), a dose-related disruption in mice treated with BDE-99,
significant at both doses, and the aberrations were more pronounced in 4-month-old mice than in
2-month-old mice, indicating worsening with increasing age. Mice receiving 0.8 and 12 mg/kg
of BDE-99 displayed significantly less activity (hypoactive) during the first 20-minute period
(0-20 minutes), while during the third 20-minute period (40-60 minutes) they were significantly
more active (hyperactive) in relation to control animals and for all three behavioral variables.
The habituation capability significantly decreased with age in mice exposed to BDE-99 at 0.8
and 12 mg/kg. Performance of 5-month-old mice in the swim maze learning/memory test was
significantly worse in mice exposed to 12 mg/kg BDE-99 than in control mice. The lowest-
observed-adverse-effect level (LOAEL) in this study was 0.8 mg/kg for effects on spontaneous
motor behavior and decreased habituation capability.
4.3.1.2. Eriksson et al (2002)
This study was undertaken to investigate whether behavioral disturbances observed in
adult mice following neonatal exposure to BDE-99 are induced during a defined neonatal brain
developmental window of unique biological susceptibility. On PND 3, 10, or 19, male and
female NMRI mice were given a single oral dose of 0 or 8 mg/kg BDE-99 by gavage in a 20%
fat emulsion (1:10 egg lecithin to peanut oil in water). There were no effects on body weight or
body-weight gain nor clinical signs of dysfunction in the BDE-99-treated mice at any time during
the experimental period. Spontaneous motor behavior tests (locomotion, rearing, and total
activity) were measured over three 20-minute periods in 4-month-old male mice (10 mice
randomly selected from three to five different litters in each treatment group). Control mice
receiving the 20% fat emulsion on PND 3, 10, or 19 showed normal habituation (i.e., a decrease
in the variables locomotion, rearing, and total activity) in response to the diminished novelty of
the test chambers over a 60-minute period, divided into three 20-minute periods. Mice
neonatally exposed to BDE-99 on PND 3 or 10 showed decreased activity during the first
20-minute interval of the 60-minute period for all three behavioral variables compared with the
control groups. During the last 20-minute period, a significantly increased activity compared
with the controls was seen for all three behavioral variables. The most pronounced effects were
seen in mice exposed to BDE-99 on PND 10, with significant hypoactive behavior during the
first 20-minute period and significant hyperactive behavior for all three spontaneous behavior
variables (locomotion, rearing, and total activity) during the last 20 minutes of the 60-minute test
period. In mice neonatally exposed to BDE-99 on PND 19, there were no changes in the three
behavioral variables compared with controls (i.e., a decrease in activity in all variables was
evident over the 60-minute observation period). In conclusion, the behavioral disturbances
observed in adult mice, following neonatal exposure to BDE-99, are induced during a defined
27
-------
critical period of neonatal brain development, and mice exposed on PND 10 are most susceptible
to the neurotoxic effects of BDE-99.
Uptake and retention of radiolabeled BDE-99 in the mouse brain were also measured in
this study after exposure of NMRI male mice (five/group) to 8 mg/kg of 14C-BDE-99 on PND 3,
10, or 19; the animals were sacrificed 24 hours or 7 days after administration (see also section
3.1). The amount of radioactivity in the brain was between 0.4 and 0.5% of the administered
dose 24 hours after administration. Seven days after the administration, 14C-BDE-99 (or its
metabolites) could still be detected in the brain, decreasing to between 0.1 and 0.3% of the
administered dose. The amount of radioactivity in the brain was similar in mice exposed on
PND 3 or 10 compared with mice exposed on PND 19 and therefore does not appear to explain
the difference in behavioral effects seen in adult mice exposed to BDE-99 on PND 3, 10, or 19.
4.3.1.3. Viberg et al (2002)
A study was conducted to determine whether changes in spontaneous behavior in adult
mice neonatally exposed to BDE-99 would include effects on the cholinergic system and thereby
would alter the response in the adult animal to the cholinergic agent nicotine. On PND 10, male
NMRI mice received a single dose of BDE-99 (>98% purity) by gavage at 8 mg/kg in a 20% fat
emulsion. Control mice received fat emulsion in the same manner. At the age of 2 months,
12 mice per group, randomly picked from three to four different litters, were subjected to
spontaneous behavior testing (locomotion, rearing, and total activity) for a 60-minute period
(0-60 minutes), divided into three 20-minute periods. Directly after the spontaneous behavior
test, the mice were given a single s.c. injection of saline solution (control) or 0.08 mg/kg nicotine
base and were tested again immediately for spontaneous motor behavior during another
60-minute period (60-120 minutes). This amount of nicotine is known to cause an increased
activity in normal adult NMRI mice.
There were no clinical signs of toxicity in the treated mice at any given time during the
experimental period, and no difference was observed in body weights or body-weight gains
between controls and treated animals. A decrease in locomotion, rearing, and total activity over
the 60-minute test period was observed in control mice in response to the diminished novelty of
the test chamber, but BDE-99-treated animals displayed significantly less activity (hypoactivity)
for all three variables during the first 20-minute period (0-20 minutes), while during the last
20-minute period (40-60 minutes) the BDE-99-treated animals had significantly increased
activity (hyperactivity) for all three variables, compared with controls. Pair-wise testing between
the nicotine-injected and saline-injected mice showed, as expected, a significant increase in
response to nicotine in the neonatally vehicle-treated mice during the first 20-minute period
(60-80 minutes) for all three variables, locomotion, rearing, and total activity. In contrast,
animals treated with BDE-99 on PND 10 and injected with nicotine at the age of 2 months
28
-------
showed significantly decreased activity during the first 20-minute period (60-80 minutes)
compared with BDE-99-treated animals injected with saline. The authors concluded that
neonatal exposure to BDE-99 on PND 10 can affect the cholinergic system (see section 4.4.2.5),
seen as changes in the adult mouse response to the cholinergic agent nicotine.
4.3.1.4. Viberg et al (2004a)
This study was carried out to determine whether exposure to BDE-99 during a period of
rapid brain growth in neonatal mice could lead to disruption of the adult brain function. Single
oral doses of BDE-99 (purity >99%) of 0, 0.4, 0.8, 4.0, 8.0, or 16 mg/kg in a 20% fat emulsion
(1:10 egg lecithin to peanut oil) were given by gavage to male and female C57/B1 mice on
PND 10. Control mice received 10 mL/kg of the 20% fat emulsion only. Spontaneous motor
behavior was tested at ages 2, 5, and 8 months in eight male and eight female mice, randomly
selected from three to five different litters in each treatment group, at each testing occasion.
Spontaneous motor behavior was measured for a 60-minute period, divided into three 20-minute
periods, at each dose. Spontaneous motor behavior tests used measured locomotion (horizontal
movement), rearing (vertical movement), and total activity (all types of vibration within the test
cage [i.e., those caused by mouse movements, shaking/tremors, and grooming]). In order to
study time-dependent changes in habituation (2-month-old versus 8-month-old mice), data from
the spontaneous motor behavior tests were used.
The habituation ratio was used to analyze alteration in habituation between 2-month-old
and 8-month-old mice, within each treatment group, in comparison with their respective
controls. Data for the three spontaneous behavior variables (horizontal movement, vertical
movement and total activity) are only available in graphic form and could not be used for
quantitative assessment.1 Numerical values suitable for dose-response assessment are only
available for the habituation ratio.
There were no clinical signs of toxicity or effects on body-weight gain or body weight at
any of the dose groups. Control mice showed habituation over the three 20-minute test periods.
There were significant dose-related changes in spontaneous motor behavior (locomotion,
rearing, and total activity) at 0.8 mg/kg and above in male and female mice at ages 2, 5, and 8
months. These disturbances were also worse with increasing age. Male and female mice
receiving doses of 0.8 mg/kg and higher showed significantly decreased activity during the first
20-minute period (hypoactive) and significantly increased activity during the last 20-minute
period (hyperactive) compared with control animals. Male and female mice exposed to the
lowest dose (0.4 mg/kg) did not significantly differ in activity in any of the three behavioral
variables during any of the three 20-minute periods. The habituation capability for the
29
-------
locomotion and rearing variabiles was signicantly decreased in the 2- and 8-month-old male and
female mice at 0.8 mg/kg and above, as evidenced by dose-related increases in the habituation
ratio. The decline in habituation capability (i.e., the increase in the habituation ratio) was more
pronounced in the 8-month-old mice than in the 2-month-old mice.
The habituation ratio for rearing (ratio between the performance periods 40-60 minutes
and 0-20 minutes for rearing), which provided a good fit in the benchmark dose (BMD)
modeling (see section 5.1.2), was 0.24, 0.51, 1.49, 45.8, 94.2, and 217 for the control, 0.4, 0.8,
4.0, 8.0, and 16 mg/kg dose groups, respectively, in 2-month-old female mice; the habituation
ratio for rearing in 8-month-old female mice was 0.22, 0.33, 2.83, 43.8, 118, and 271 for the
control and five different doses, respectively, indicating that the capability of the animals to
habituate to a new environment decreased with increasing BDE-99 dose and with age. No major
gender differences in spontaneous motor behavior responses or habituation capability were seen
in this study.
The no-observed-adverse-effect level (NOAEL) for spontaneous motor behavior effects
in this study was 0.4 mg/kg. The LOAEL was 0.8 mg/kg for significant changes in spontaneous
motor behavior and decreases in the rearing and locomotion habituation capability in both male
and female mice, worsening with increasing age.
4.3.1.5. Viberg et al (2004b)
A study was conducted to determine effects on spontaneous behavior in adult male mice
neonatally exposed to BDE-99 and whether these effects would include changes in the density of
cholinergic nicotinic receptors in the hippocampus of the adult animal. Such changes have been
proposed to affect learning and memory functions. (See section 4.4.2.5 for the discussion of the
results of the receptor binding studies.) Single oral doses of 0, 0.2, 0.4, or 12 mg/kg of BDE-99
(>98%) in a 20% fat emulsion were given by gavage to male NMRI mice on PND 10.
Spontaneous motor behavior was measured over three 20-minute periods in groups of mice at
the age of 4 months. Ten mice were tested, randomly picked from three to five different litters
in each treatment group. Spontaneous motor tests evaluated locomotion, rearing, and total
activity behaviors. There were no clinical signs of toxicity or significant differences in body-
weight gain or adult weight between controls and mice treated with BDE-99 at any time during
the experimental period.
Habituation, defined as a decrease in the three behavioral variables (locomotion, rearing,
and total activity) in response to the diminished novelty of the test chamber over the 60-minute
period, was observed in the control animals. Mice exposed neonatally to 12 mg/kg of BDE-99
displayed significantly less activity (hypoactive) for all three behavioral variables during the first
Attempts to obtain numerical values and other information on the data from the neurobehavioral studies were not
30
-------
20-minute period (0-20 minutes) compared with the controls, while during the third 20-minute
period (40-60 minutes), they were significantly more active (hyperactive) than the control
animals in relation to all three behavioral variables. Mice receiving 0.2 or 0.4 mg/kg BDE-99
showed no significant differences in activity for any of the three behavioral variables compared
with the control animals at any of the 20-minute periods. The NOAEL in this study was
0.4 mg/kg and the LOAEL 12 mg/kg for effects on spontaneous motor behavior.
4.3.1.6. Ankarberg (2003)
The objective of the study by Ankarberg (2003) was to determine whether neonatal
exposure to nicotine could affect the susceptibility of adult mice to BDE-99. The motor
behavior response of adult mice after exposure to BDE-99 was used as a measure of the impact
of the nicotine on the neonatal nervous system. Ten-day-old male NMRI mice received s.c.
injections of saline (10 mL/kg) or nicotine base at 0.033 mg/kg, twice daily for 5 consecutive
days (total daily dose, 0.066 mg/kg-day). Studies of brain development in rodents previously
identified 10 days as the peak period for the developmental "brain growth spurt," during which
mice and rats acquire many sensory and motor functions. At the age of 5 months, the 8-10 mice
received 8 mg/kg BDE-99 in 20% fat emulsion or 10 mL/kg of 20% fat emulsion by gavage and
were tested for spontaneous motor activity (locomotion, rearing, and total activity) for three
20-minute periods, 24 hours after exposure to BDE-99. Control animals, animals that received
0.066 mg/kg-day nicotine base neonatally but were not given the BDE-99, and animals that
received only 8 mg/kg BDE-99 as adults showed a normal decrease in activity over the
60-minute test period, indicating a normal habituation pattern in response to the diminished
novelty of the test chamber. However, the mice that received nicotine on PND 10 and BDE-99
as adults showed a lack of habituation. They displayed hypoactive behavior in the beginning of
the test period (0-20 minutes) but became hyperactive toward the end of the period (40-
60 minutes), indicating that the neonatal nicotine exposure had affected their susceptibility to
BDE-99 as adults. At the age of 7 months, the animals were again tested for spontaneous motor
behavior. The lack of habituation in the nicotine-BDE-99-treated mice was even more
pronounced, indicating a disturbance that worsened with age. Overall, this study indicates that
neonatal nicotine exposure affected the response of the adult animals to BDE-99.
4.3.1.7. Branchi et al (2002)
In order to examine the neurobehavioral effect of perinatal exposure, BDE-99 dissolved
in corn oil was administered by gavage at 0, 0.6, 6, or 30 mg/kg-day to groups of female CD-I
Swiss mice (four/group) from GD 6 through PND 21, at which time the pups were weaned.
successful.
-------
Effects on pregnancy and somatic and neurobehavioral development of pups were assessed.
Body-weight gain of pregnant females, pregnancy duration, proportion of successful deliveries,
pup sex ratio, and body-weight gain of pups from birth to weaning were not affected by
treatment with BDE-99.
Six to eight male and female pups from each litter of each treatment group were used in a
series of tests to assess somatic and neurobehavioral development from PNDs 2-20. These tests
were carried out every 2 days and included hair growth; day of eyelid and ear opening and of
incisor eruption; righting, forelimb stick grasp, and forelimb-placing reflexes; level and vertical
screen tests; screen climbing test; and pole grasping test. Ultrasonic vocalization on PNDs 4, 8,
and 12 and homing tests on PND 11 were carried out on one male and one female not previously
handled or tested from each litter of each treatment group. In addition, an open-field apparatus
was used to test locomotion (horizontal movement), rearing (vertical movement), and
thigmotaxis (time and distance traveled close to the walls) in one male and one female from
each litter of each treatment group for 30-minute sessions on PNDs 22 and 34 and for 60-minute
sessions on PNDs 60 and 120.
In the battery of tests carried out from PNDs 2-20, BDE-99 treatment did not affect
somatic development (hair growth and day of eyelid and ear opening and incisor eruption).
There was a statistically significant 2-day delayed appearance of screen-climbing response in the
high-dose group (30 mg/kg-day); all other responses based on neuromotor coordination from
PNDs 2-20 were not affected by BDE-99 treatment. No effects were seen in pups from any of
the treatment groups on ultrasonic vocalization or homing performance assessed on PND 11 both
for distance traveled and latency to reach the scent area. In the open-field test, there was no
statistically significant difference in activity between controls and treatment groups on PND 22.
However, BDE-99 exposure affected several behavioral/activity parameters in the open-field
arena on PNDs 34, 60, and 120, indicating that behavioral alterations due to perinatal BDE-99
exposure seem to worsen with increasing age, becoming clearly evident around 1 month of age.
On PND 34, mice were hyperactive in the 0.6 and 6 mg/kg-day dose groups but not in the high-
dose group. Mice exhibited a statistically significant increase in rearing frequency, with mice in
the 0.6 mg/kg-day dose group being more hyperactive than mice in the 6 mg/kg-day dose group.
The mice in the 6 mg/kg-day dose group also exhibited signs of increased locomotion as
evidenced by increases in distance traveled, although this was not apparent in the low- and high-
dose groups. Thigmotactic response, considered an index of anxiety, was not affected at any
dose. On PND 60, mice in the 0.6 mg/kg-day group, but not in the 6 and 30 mg/kg-day groups,
displayed significantly more locomotion compared with controls. Thigmotactic behavior on
PND 60, measured as percent of time spent near the walls, was significantly lower at the
medium dose only (6 mg/kg-day) in comparison with control mice, indicating a less marked
fearful response in this treated group.
32
-------
At adulthood (PND 120), the 0.6 and 6.0 mg/kg-day groups displayed significantly lower
levels of locomotion (in contrast to hyperactivity measured prior to PND 120) than controls
during the last part of the 60-minute test session. At this age, rearing and thigmotaxis were not
affected at any BDE-99 dose. The authors concluded that prenatal and postnatal exposure of
mice to BDE-99 produced a transient hyperactivity that was characterized by an inverted dose-
response relationship, ending around 4 months of age.
The behavioral/activity changes observed in mice treated with BDE-99 did not
consistently show a dose-response relationship. Effects were generally observed in mice
administered the medium dose and not observed in the others. The lack of a clear dose-response
relationship in the behavior/activity changes in this study does not permit clear identification of
the NOAEL/LOAEL for alteration in behavioral or activity parameters. Additionally, the
magnitude of variation in responses among the low, medium, and high doses cannot be
determined with any precision because all motor activity data are presented in graphic forms and
do not show a consistent relationship to dose.
4.3.1.8. Branchi et al (2005)
It has been reported that gavage administration of a test compound can in itself produce
stress in the animal. In this study, BDE-99 at 0 or 18 mg/kg-day was administered to CD-I
Swiss mice (nine/group) from GD 6 to PND 21, except for PND 0 (day of birth), when dams
were left undisturbed. Two modes of administration of BDE-99 to the dams were investigated
for their effects on neurobehavioral development in male offspring: by gavage in corn oil or by
self-administration, consisting of letting the mouse spontaneously drink BDE-99 dissolved in
corn oil from a modified syringe (without the needle and with a larger hole).
Pregnancy duration, body-weight gain of dams, proportion of successful deliveries, litter
size, pup weight, and sex ratio were not affected by treatment with BDE-99 or by the method of
administration in comparison with the respective control groups.
On PNDs 34, 60, 90, and 120, male mice (one from each litter, seven to nine mice/group)
were tested in an open-field apparatus that measured locomotion (horizontal movement), rearing
(vertical movement), and thigmotaxis (time and distance traveled close to the walls). Testing
sessions were 30 minutes (three 10-minute blocks) on PND 34 and 60 minutes (six 10-minute
blocks) on PNDs 60, 90, and 120. Each animal was tested only once. Distance traveled and
frequency of rearing were not affected by the exposure methods (gavage or self-administration),
and, therefore, the groups were pooled together. On PND 34, offspring of dams treated with
BDE-99 showed hyperactivity for distance traveled and frequency of rearing during the third
10-minute testing period. On PNDs 60, 90, and 120, behavioral parameters (distance traveled
and frequency of rearing) in treated mice were not different from those of controls at any time
point during the 60-minute testing period.
33
-------
With regard to thigmotactic behavior, an index of anxiety, mice administered BDE-99 by
gavage spent more time near the wall than self-administered BDE-99 mice, when tested on
PNDs 34, 60, and 90, with the percent of time spent near the wall in the gavage group reaching
statistical significance on PND 34 only. On PND 120, the difference in thigmotactic behavior
was minor, suggesting that the effect of the gavage route of administration was temporary.
On PND 22, two male mice from each of the control, BDE-99 gavage, and self-
administered groups were sacrificed, and BDE-99 levels were determined in the brain. No effect
of administration route was found on level of BDE-99 in the brain. The mean BDE-99 level in
treated animals was 640 ng/kg compared to 5 ng/kg for the controls.
Serum total and free T4 levels were also measured in BDE-99-treated male mice (eight to
nine/group) on PND 22 and were not found to be statistically different from control levels. No
effect of method of administration was found on T4 levels.
4.3.2. Rats
4.3.2.1. Kuriyamaetal (2005)
The effect of in utero exposure to BDE-99 on locomotor activity and male reproductive
health was investigated in rat offspring. Groups of 16-20 Wistar rats were given single doses of
0, 0.06, or 0.3 mg/kg of BDE-99 (98% purity) by gavage in peanut oil on GD 6. Since PBDEs
may interfere with thyroid hormone homeostasis, a reference group for thyroid-mediated effects
was included in which dams were treated from GDs 7-21 with 5 mg/L of the goitrogen
6-n-propyl-2-thiouracil (PTU) in drinking water (approximate dose: 0.9 mg/kg-day).
Developmental landmarks (eruption of incisors, fur development, eye opening, and testes
descent) and postnatal reflexes (development of spontaneous cliff-drop aversion reflex starting
on PND 3 and ability to stay on a rotating rod for 3 minutes at seven revolutions/minute starting
on PND 18) were evaluated in 163-218 male and female pups. The eruption of incisors was
significantly delayed in the PTU-treated group and in the 0.3 mg/kg BDE-99-treated group. The
development of spontaneous cliff-drop aversion reflex was significantly delayed in the PTU-
treated male and female offspring and in male offspring exposed to 0.3 mg/kg BDE-99. No
other effects on developmental landmarks or spontaneous reflexes were seen.
On PNDs 36 and 71, the circadian locomotor activity of one male and one female per
litter per group (16 litter groups, controls; 20 litter groups, 0.06 mg/kg; 19 litter groups,
0.3 mg/kg), housed individually, was evaluated over 24-hour periods. Locomotor activity was
measured in individual offspring by using a device that monitors the locomotion of the animal at
5-minute intervals over a 24-hour period. Locomotor activity included total activity measured as
light-beam interruption (LBI) counts per day, duration (hours) of activity per day, LBI counts per
active phase (an active phase is defined as the time period from when the animal begins to move
until it pauses moving), and duration of activity (minutes) per active phase. There was no
34
-------
difference between the sexes for all groups, and therefore the data for males and females were
pooled. On PND 36, the total activity (LBI count) was significantly increased in the offspring of
dams treated with 0.3 mg/kg BDE-99 and in the PTU-treated group. The number of active hours
per day was also higher in the 0.3 mg/kg group, an effect not seen in the PTU group. LBI
counts/active phase and duration of activity/active phase were also significantly increased in the
BDE-99 group at 0.3 mg/kg and in the PTU-treated group.
PTU-treated animals, while temporarily hyperactive on PND 36, restored to normal
levels on PND 71. On PND 71, both total activity and duration of activity per day were
significantly increased at 0.06 and 0.3 mg/kg BDE-99 but not in the PTU-treated groups. When
the locomotor activity was expressed as LBI counts/active phase and duration of activity/active
phase, hyperactivity was not seen on PND 71.
The effects of in utero exposure to BDE-99 on body and organ weights and the male
reproductive system of adult offspring (PND 140) were also investigated in this study (Kuriyama
et al., 2005). Twelve males per treatment group (from different litters) were sacrificed on
PND 140, and thymus, liver, spleen, testis, epididymis, ventral prostate, and seminal vesicle
weights were recorded. Spermatids and sperms were counted, sperm morphology were
examined, and testosterone and luteinizing hormone (LH) levels were measured. No effects
were seen on body weight or absolute and relative liver, thymus, seminal vesicle, or prostate
weights in BDE-99-treated animals. Sperm morphology, LH, and testosterone were unaffected.
Absolute and relative spleen weights were increased but not in a dose-dependent manner.
Relative testis weight was significantly decreased at 0.3 mg/kg BDE-99 and in the PTU group,
and relative epididymis weight was significantly decreased at 0.06 and 0.3 mg/kg BDE-99 and in
the PTU-treated group. Sperm numbers were significantly decreased compared with those in
controls at both BDE-99 doses but not in a dose-dependent manner, with sperm numbers (in
millions) being 190, 135, and 156 in the controls, 0.06, and 0.3 mg/kg dose groups, respectively.
Daily sperm production and spermatid count were significantly decreased at both doses in a
dose-dependent manner: daily sperm production (in millions) was 44, 30, and 29 and spermatid
counts (in millions) were 266, 183, and 175 in the control, low-, and high-dose groups,
respectively. The percentage of abnormal sperm was within normal limits in all groups.
Reproductive effects were also examined in this study. Adult male offspring
approximately 150 days old (5-19/group) from the BDE-99- and PTU-treated groups were mated
with untreated females (1:1) daily for 14 days to determine whether the males were fertile and
could produce normal offspring. The dams were sacrificed on GD 21. Uterine and fetal
weights, number of implantations, implantations per litter, viable fetuses per litter, percent total
resorptions, and male/female sex ratio were all within the normal range of control in all
treatment groups.
35
-------
Sexual behavior (ejaculatory, mounting and intromission latencies, intromission
frequency, and number of penetrations before the first ejaculation) in 160-day male offspring
(20/group) were also normal in all treatment groups compared with controls. The only effect
seen was a significant decrease in the 0.3 mg/kg BDE-99 group in the number of animals that
had two or more ejaculations during 20 minutes of mating. Approximately 50% of controls had
a second ejaculation, while 70% of the PTU-treated animals achieved a second ejaculation. In
the 0.06 and 0.3 mg/kg BDE-99 groups, only 39 and 21%, respectively, of the males achieved a
second ejaculation. Therefore, PTU treatment improved the sexual performance of mice, while
BDE-99 decreased it. The biological significance of this effect is uncertain.
In summary, treatment of rats with BDE-99 at 0.06 and 0.3 mg/kg on GD 6 resulted in a
dose-dependent decrease in daily sperm production, spermatid count, and relative epididymis
weight in adult male offspring on PND 140. No effects were seen on male fertility or sperm
morphology at these doses. However, in rodent species, sperm number has to be substantially
reduced before fertility is compromised, while relatively small changes in sperm production in
men may affect human reproduction. The decreased sperm production, spermatid count, and
epididymis weight warrant additional studies to determine their significance for reproductive
functions in humans.
The LOAEL in this study was 0.06 mg/kg, based on increases in certain locomotor
activity parameters on PND 71. A NOAEL for absence of hyperactivity was not identified in this
study. The LOAEL for decreased sperm production, spermatid count, and relative epididymis
weight on PND 140 was 0.06 mg/kg, the lowest dose tested.
4.3.2.2. Viberg et al (2005)
The objective of this study was to determine whether the changes in spontaneous
behavior and cholinergic receptors observed in adult mice neonatally exposed to BDE-99
(Viberg et al., 2004a, b) could also be induced in another species, namely the rat. Results from
the receptor assay are reported in section 4.4.2.5.
Single oral doses of 0, 0.8, 8.0, or 16 mg/kg BDE-99 (purity >98%) in a 20% fat
emulsion (1:10 egg lecithin to peanut oil) were given by gavage to male Sprague-Dawley rats on
PND 10. Control rats received 10 mL/kg of the 20% fat emulsion. Spontaneous motor behavior
was tested in nine 2-month-old rats randomly selected from three to five different litters in each
treatment at each testing occasion. Spontaneous motor behavior was measured for a 60-minute
period, divided into three 20-minute periods, at each dose. Spontaneous motor behavior tests
used measured locomotion (horizontal movement), rearing (vertical movement), and total
activity (all types of vibration within the test cage [i.e., those caused by rat movement,
shaking/tremors, and grooming]).
36
-------
There were no clinical signs of toxicity at any time during the experimental period nor
significant differences in body weight gain or adult weight between controls and rats treated with
BDE-99. Two-month-old control rats showed habituation (i.e., a distinct decrease in
locomotion, rearing, and total activity over the three 20-minute test periods) in response to the
diminishing novelty of the test chamber. Rats exposed on PND 10 to 8.0 and 16 mg/kg BDE-99
displayed significantly less activity for all three behavioral variables during the first 20-minute
period, while, during the third 20-minute period (40-60 minutes), they were significantly more
active than the control animals for all three behavioral variables. Rats receiving BDE-99 at 0.8
mg/kg did not show any difference from controls in locomotion or rearing activities over the
three 20-minute test periods. A slight decrease in the total activity variable was seen only during
the first 20-minute period that returned to control levels during the second and third 20-minute
periods.
The NOAEL in this study was 0.8 mg/kg. The LOAEL was 8.0 mg/kg for significant
changes in spontaneous motor behavior in 2-month-old rats exposed to BDE-99 on PND 10.
These changes in behavior were characterized by hypoactive behavior, followed by hyperactive
behavior, for all three variables (locomotion, rearing, and total activity) during the 60-minute test
period. The NOAEL/LOAEL values in this study indicate that rats are equally or perhaps less
sensitive than mice to the spontaneous motor behavior effects of BDE-99. In the study in mice
by the same research group (Viberg et al., 2004a), the NOAEL was 0.4 mg/kg and the LOAEL
was 0.8 mg/kg for significant changes in spontaneous motor behavior in 2-month-old mice
exposed to BDE-99 on PND 10.
4.3.2.3. Tahness et al (2005)
The effect of BDE-99 on the female reproductive system was evaluated by the same
research group as that of the Kuriyama et al. (2005) study. A single dose of 0.06 or 0.3 mg/kg
BDE-99 (98% purity) was administered by gavage to Wistar rats on GD 6. The controls received
the peanut oil vehicle. A reference control was treated with PTU at a concentration of 5 mg/L in
drinking water on GDs 7 through 21. At approximately 5 months of age, 20 virgin female F!
offspring from each group were mated with untreated males to evaluate fertility.
Pregnancy rate, total implantation sites, mean implantation sites per gravid dam, total
live fetuses per dam, resorption rate, and percentage of dams with resorptions in the F! females
were not statistically different from controls at both doses of BDE-99. The only statistically
significant effect noted was an increase in mean fetal weights at 0.06 mg/kg BDE-99, but not at
0.3 mg/kg BDE-99, and in the PTU-treated group. Pregnancy rate in the PTU-treated group was
also significantly lower than in controls.
Flistologic evaluation by electron or light microscopy of the ovary, uterus, and vagina
was performed in the Fj female offspring on PND 90. Electron micrographs and
37
-------
photomicrographs revealed qualitative ultrastructural changes in the ovaries and hyperplastic
vacuolar degeneration of the vaginal epithelium in the Fj offspring from the 0.06 and 0.3 mg/kg
BDE-99 and PTU-exposed groups. No significant changes were observed in the different
ovarian follicle types following exposure to either BDE-99 or PTU, indicating that follicle
numbers and maturation of follicles were unaffected. Skeletal anomalies were observed in two
animals from the F2 generation from two different litters, following exposure of the F0 dams to
0.3 mg/kg BDE-99 on GD 99. The possible causes for these anomalies remain unknown, and the
authors suggested they may be either spontaneous or substance related.
4.4. OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES
4.4.1. Subcutaneous Exposures
4.4.1.1. Lilienthal et al (2005)
BDE-99 was administered by s.c. injections to pregnant Long-Evans rats from GDs 10-
18 at doses of 1 or 10 mg/kg-day. Controls received the olive oil vehicle. For comparison, an
additional group was exposed to Aroclor 1254 at 30 mg/kg-day. Dissections were conducted on
GD 19 and in male offspring on PNDs 21 and 160 for organ examination and analyses of
circulating levels of estradiol and testosterone. Neurobehavioral measurements in male
offspring included sweet preference on PND 120 and haloperidol-induced catalepsy on PND
240. In addition, activity in the open field was studied in male offspring on PNDs 30, 90, and
400.
On PND 160 a slight but significant reduction in anogenital distance, a marker of sexual
development, was reported in male offspring (eight/group) after maternal exposure to the high
dose of BDE-99 (10 mg/kg-day) and also in the Aroclor-exposed group. Testes weights were not
affected by BDE-99 or Aroclor treatment. Circulating levels of estradiol and testosterone were
significantly decreased by exposure to BDE-99 at both doses of BDE-99 and by exposure to
Aroclor 1254. Sweet preference, measured as the ratio of saccharin to water consumption,
showed a significant increase (which may indicate behavioral feminization) at 10 mg/kg-day
BDE-99 but was not altered by Aroclor treatment (10-12/group). Locomotor activity on PND
30, 90, or 400 of male rats in the open field was not changed by BDE-99 or Aroclor treatment.
By contrast, in the catalepsy test, all exposed groups, in comparison with the control group,
exhibited increased retraction latencies of the hind legs in the box 30 minutes after the injection
with haloperidol. The authors concluded that these results, taken together, suggest an endocrine-
modulating activity of BDE-99.
4.4.1.2. Lilienthal et al. (2006)
This follow-up study used the same protocol as Lilienthal et al. (2005) and was
conducted to determine reproductive and developmental effects following BDE-99
38
-------
administration by s.c. injections to pregnant Long-Evans rats from GDs 10-18 at doses of 1 or 10
mg/kg-day. Body weight of dams and body weight gain during gestation were not influenced by
exposure to BDE-99 or Aroclor 1254. The numbers of implantations per litter and pups per litter
and the percentage of male pups were not different from controls. At weaning there were no
significant differences in weights of brain, thymus, testis, ventral prostate, and uterus between
BDE-99-exposed and control rats. There was a tendency for decreased pituitary weights in male
offspring on PND 21 at the high-dose level (10 mg/kg-day) compared with controls. In contrast,
females on PND 21 showed a statistically significant increase in pituitary weight at the low dose
only (1 mg/kg-day). The most notable effect seen was a significant reduction in thyroid weights
in adult male and female offspring, compared with controls, after exposure to 1 and 10 mg/kg-
day BDE-99.
Puberty onset was not affected in males but was delayed in females at 10 mg/kg-day.
A slight but significant reduction in anogenital distance was reported in adult males
(eight/group) on PNDs 21 and 160 after maternal exposure to the high dose of BDE-99 (10
mg/kg-day). A dose-related decrease in the number of secondary ovarian follicles was seen in
females exposed to BDE-99 that was significant at 10 mg/kg-day.
Concentrations of BDE-99 in the brain tissues of dams and offspring were highest on
GD 19 but decreased at weaning (PND 21) and returned to control levels on PND 160. A
decline in adipose tissue concentration occurred in dams during lactation and in male offspring
from weaning to adulthood. Circulating levels of estradiol on PNDs 21 and 160 were
significantly decreased in male offspring exposed to BDE-99 at both doses. Levels of
testosterone were significantly reduced in males on PND 160 at both doses of BDE-99. Sweet
preference, measured as the ratio of saccharin to water consumption in male rats, showed a
significant increase (which may indicate behavioral feminization) at 10 mg/kg-day BDE-99 but
was not altered by Aroclor treatment (10-12/group).
In summary, gestational exposure to BDE-99 did not affect reproductive success in dams
or development of body weights in offspring at the doses tested. The weights of reproductive
and nonreproductive organs were largely unchanged. There was a significant dose-related
decrease in circulating estradiol in male offspring at PNDs 21 and 160 and in testosterone on
PND 160. A marked effect was the decrease in thyroid weights in adult offspring, which was
more pronounced in the high-dose group.
4.4.1.3. Ceccatdli et al (2006)
Ceccatelli et al. (2006) examined the effects of prenatal exposure to BDE-99 (purity
>99%) on several developmental endpoints and gene expression in adult uteri of exposed female
rats. Groups of six to nine time-pregnant Long-Evans dams were given daily s.c. injections of 0,
1, or 10 mg/kg BDE-99 from GDs 10-18. There were no clinical signs of adverse effects during
39
-------
pregnancy, and the body weights of the dams were not significantly affected by the BDE-99
exposure. There were no significant differences between the control and exposed pups related to
litter size, sex ratio, and body weight/litter on PND 2 or 14 and anogenital distance on PND 2.
Litters were culled to 8-10 pups per litter on PND 2 and the male and female litter mates were
kept separated. There was no additional exposure to BDE-99 after birth. Monitoring of the day
of vaginal opening as the females reached puberty showed no significant delays on a per litter
basis but showed a slight dose-related significant delay (p < 0.05) for individual rats.
The female pups were sacrificed at 12 weeks of age. Body weights as well as absolute
and relative liver, uterine, and ovarian weights were measured. There were no significant
differences among groups except for a slight but significant increase in absolute and relative
ovarian weights in the animals prenatally exposed to 10 mg/kg-day. Analysis of plasma, brain,
and adipose tissue for the presence of BDE-99, 120 days after exposure ceased, identified small
amounts in the brain and plasma and substantially higher levels in the adipose tissues for both
dose groups.
The uteri were collected from the now adult female pups. Levels of mRNA for insulin-
like growth factor (IGF)-l, progesterone receptor (PR), estrogen receptor (ER)-a, and ER-P were
measured after amplification by using real-time PCR and targeting of the resultant cDNA by
using appropriate probes. There was a dose-dependent decrease in PR. In the case of ER-a and
ER-P, the levels were elevated compared with controls for the 1 mg/kg dose but were
comparable with controls for the 10 mg/kg-day dose. The IGF-1 results were more difficult to
evaluate. Distributions within dose groups were highly skewed. The levels for the 1 mg/kg dose
group were significantly higher than those for controls, while those for the 10 mg/kg group,
although still elevated compared with controls, were not significantly different and were lower
than those for the 1 mg/kg dose group.
The data from the first phase of the Ceccatelli et al. (2006) study suggested that BDE-99
might have subtle developmental impacts on the endocrine status of the uterus in female adult
rats exposed to BDE-99 only during prenatal development. The authors then conducted a second
phase to their experiment by examining the response of IGF-1, PR, ER-a, and ER-P biomarkers
in prenatally BDE-99-exposed adult females after a single s.c. injection of estradiol-l?p. The
rats used in this part of the study were ovariectomized at 10 weeks, injected with estradiol-l?p at
12 weeks, and sacrificed 6 hours later. The purpose of the ovariectomies was to reduce exposure
to endogenous estrogen.
The results of these analyses were complex because, when compared with untreated
controls, the ovariectomized controls had differing baseline levels of the hormonal biomarkers.
Baseline levels for all of the biomarkers except ER-P were decreased compared with the
nonovariectomized controls. Baseline levels of PR in the low-dose BDE-99-treated
ovariectomized animals were significantly higher than those in the ovariectomized controls. In
40
-------
the high-dose ovariectomized BDE-99-treated animals, IGF-1 and ER-P levels were significantly
higher than those in ovariectomized controls.
After treatment with the estradiol, the levels of IGF-1 were increased in all groups, but
the magnitude of the increase was less than that of the controls for both BDE-99-treated groups,
and the difference relative to the controls increased with BDE-99 dose. The levels of PR also
increased in response to estrogen; the increase in the high-dose BDE-99-treated group was
significantly higher than in the controls. In response to the estradiol, the levels of ER-a and ER-
P decreased in both the controls and the BDE-99-treated animals. There were no significant
differences between groups for ER-a, but for ER-P there was a significant, dose-related increase
in the magnitude of the response. Overall, the results of the estradiol challenge demonstrated
that there were significant differences in the hormonal responses of prenatally exposed BDE-99-
treated mice when they became adults.
4.4.2. Receptor Site Interactions
There is considerable evidence from studies of PCBs, chlorinated dibenzo-p-dioxins
(CDDs), and chlorinated dibenzofurans (CDFs) that halogenated aromatic compounds exert an
influence on cells by interacting with membrane receptor sites and activating cellular
transcription factors. Transcription factor complexes then initiate DNA synthesis, allowing the
cell to respond to the extracellular signal by producing a series of mRNAs that in turn produce a
variety of proteins. This process is termed signal transduction. The structural similarities
between PBDEs and PCBs suggest that PBDEs might activate the Ah receptor, ER, and
androgen receptor. Based on the data from the well-studied PCBs, CDDs, and CDFs, the
activation of these receptor sites is associated with immunosupression, reproductive effects, and
carcinogenesis (Klaassen, 1996; Bock, 1994), all endpoints of interest for PBDEs. Table 4-1
provides a summary of the pentaBDE congeners that have been evaluated in a variety of receptor
interaction studies.
Table 4-1. Receptor interaction studies of pentaBDE congeners
Congener
evaluated
Ah receptor
ER
Androgen
receptor
CARb
PXR/SXRb
85
Xa
X
99
X
X
X
X
X
100
X
X
X
105
X
X
119
X
X
126
X
X
Findings
Effect levels are l(T2to 1(T5 that of TCDDb. Receptor
binding and CYP-1A1/2 results are not always consistent
for the different assays that have been completed.
No activity for BDE-85 and -99. Weak activity for
BDE-lOOand-119.
BDE-100 has a greater antiandrogenic effect than
BDE-99.
Receptor interactions up-regulate expression of associated
CYP-450 isozymes. CAR activation stronger than PXRb.
Receptor interactions up-regulate expression of associated
41
-------
CYP-450 isozymes, but the up-regulation is not
concentration related.
3X indicates that the congener was tested for receptor effects; for most congeners several methodologies for
evaluation of receptor interactions were employed.
bCAR = constitutive androstane receptor; TCDD = 2,3,7,8-tetrachlorodibenzo-p-dioxin; PXR = pregnane X receptor;
SXR = steroid X receptor.
4.4.2.1. Aryl Hydrocarbon Receptors
Transcription of the genes for CYP-1 Al, -1A2, and -1B1 is linked to a signal
transduction cascade that is initiated by activation of the Ah receptor by an appropriate ligand.
The CYP-1 family of enzymes is highly conserved in mammals and is responsible for the
oxidative metabolism of a variety of planar and near-planar compounds (Lewis et al., 1998).
The CYP-1 family of enzymes metabolically activates and metabolizes polycyclic aromatic
hydrocarbons and aromatic amines as well as PBDEs. Many substrates for the CYP-1 family
enzymes are also Ah receptor ligands. Differences in Ah receptor affinity are correlated with
variations in CYP-1 induction. Receptor site affinity has been shown to reflect potency and the
potential for a xenobiotic to cause adverse health effects.
Chen et al. (2001) studied the affinity of several PBDE congeners for rat hepatic Ah
receptor through competitive binding assays and determined their ability to induce hepatic
CYP-450 enzymes by means of EROD assays (a biomarker for CYP-1 Al/2 induction) in chick
and rat hepatocytes, liver cell lines from rainbow trout, and rat and human tumor cell lines.
PentaBDE congeners BDE-85, -99, -100, -119, and 126 (>98% purity) had Ah receptor-binding
affinities approximately 2 x 10~2 to 8 x 10~5 that of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD).
BDE-85, which is not a major constituent in environmental samples, was the most active, but its
relative binding affinity was only 2 x 10~2 that of TCDD. The binding affinities of the
pentaBDEs were not influenced by the planarity of the molecule. The authors hypothesized that
the large atomic volume of bromine distorted the Ah-binding site so that the coplanarity of the
rings was less important in Ah binding than it was for the PCBs.
Quantitative measures of EROD induction were reported for BDE-85, -99, -100, -119,
and -126. EROD induction was strongest in all cell lines for BDE-100, 119, and 126, although
their relative induction potencies in the different cell cultures were approximately three to four
orders of magnitude lower than the potency of TCDD. BDE-85 was a very weak inducer in rat
hepatocytes and inactive in the other cells. The environmentally prominent congener BDE-99
was not an inducer in any cell line. These structurally related pentaBDE congeners were found
to have differing responses in the in vitro test systems studied, and all were considerably less
potent than TCDD, a strong Ah activator (Chen et al., 2001).
Peters et al. (2006) examined the interaction of BDE-99 and -100 as well as other PBDEs
on the Ah receptor in primary hepatocyte cultures from four healthy cynomolgus monkeys (three
42
-------
males, one female) by using EROD activation as a biomarker for receptor activation. Both
compounds were weak Ah agonists when coexposures of TCDD and the respective PBDE were
tested, as evidenced by a decrease in the activation caused by TCDD alone. The action of the
PBDEs was receptor mediated rather than through inhibition of the enzyme since no EROD
inhibition occurred if TCDD exposure preceded the PBDE exposure. Environmentally relevant
concentrations of PBDEs (1-10 jiM) were evaluated. There was variability in the response of
the primary hepatocytes from the four monkeys, likely reflecting individual differences in the
animals.
Using hepatocyte cultures from Sprague-Dawley rats, Chen and Bunce (2003)
investigated whether PBDE congeners, including pentaBDEs, act as Ah receptor agonists or
antagonists at sequential stages of the Ah receptor signal transduction pathway leading to
CYP-1A1. These issues are environmentally relevant because of the strong rank-order
correlation among strength of Ah receptor binding, CYP-1A induction, and toxicity for many
halogenated aromatic compounds.
There were four components to this study (Chen and Bunce, 2003): (1) the binding of the
PBDE congener to the Ah receptor, (2) the binding of the receptor/PBDE complex to an
oligonucleotide segment of the dioxin response element, (3) the induction of EROD, and (4) the
production of CYP-1 A mRNA and CYP-1 A protein. The pentaBDE congeners evaluated in the
study were BDE-85, -99, -100, -119, and -126.
BDE-119 and -126 were the most active of this group when compared to TCDD. They
were moderately active in dioxin response element binding and induced responses of both
CYP-1 Al mRNA and CYP-1 Al protein equivalent to the maximal response of TCDD in
primary Sprague-Dawley rat hepatocytes, although at concentrations three to five orders of
magnitude greater than TCDD. BDE-85 was inactive, and BDE-100 was a very weak activator
of dioxin response element binding. When tested in combination with TCDD, BDE-119 and
-126 tended to enhance the activity of a nonsaturating concentration of TCDD and slightly
inhibit a saturating TCDD concentration.
The environmentally prominent congener BDE-99 was inactive at all stages of signal
transduction. BDE-99 did not have an additive relationship with nonsaturating TCDD
concentrations and acted as an antagonist in combination with a saturating TCDD concentration.
The authors concluded that, at present, the current concentrations of PBDEs in the biota,
including those of the environmentally predominant congeners BDE-99 and -100, contribute
negligibly to dioxin-like toxicity compared with other environmental contaminants such as PCBs
and TCDD but cautioned that this may change as the concentrations of PCBs decline and those
for PBDEs increase.
Villeneuve et al. (2002) examined the ability of several pentaBDE congeners (BDE-99,
-100, -105, and -126) to induce Ah receptor-mediated gene expression in vitro, using H4IIE-luc
43
-------
(luciferase) recombinant rat hepatoma cells. The cells were grown in culture well-plates and
then exposed to PBDE concentrations ranging from 2-500 ng/mL. Luminescence was measured
and compared to the maximum response observed with a 1,500 picomolar TCDD standard
(%-TCDD-max). A positive response was defined as any response that was greater than three
standard deviations (SDs) above the mean value for the control. BDE-99, -100, and -105 failed
to induce Ah receptor-mediated gene expression in H4IIE-luc cells. BDE-126 induced
significant Ah receptor-mediated gene expression at 500 ng/mL, but the magnitude of induction
was only 1.7%-TCDD-max. These results are qualitatively consistent with those of Chen and
Bunce (2003).
Sanders et al. (2005) used an in vivo approach to study Ah receptor site activation by
BDE-99 as well as several other PBDE congeners. Groups of F344 male rats (three/group),
10-12 weeks old, were dosed by gavage once daily for 3 days with BDE-99 (96% purity) in corn
oil at 0, 1, 10, or 100 jimol/kg-day. The animals were sacrificed 24 hours after receiving the last
dose. The liver was removed, and RNA from a 100 mg liver sample was isolated, converted to
its cDNA, and amplified by using PCR. The resultant DNA samples were then analyzed to
determine the expression of CYP-1A1, a protein linked to Ah receptor activation.
BDE-99 had a significant effect on the level of CYP-1A1 (8.1 times the vehicle-treated
controls) only at 100 |imol/kg-day (57 mg/kg-day), making it a weak activator of the Ah
receptor. When BDE-99-induced CYP-1 Al expression was compared with induction by
tetraBDE-47 and hexaBDE-153, the impact on the Ah receptor seemed to be correlated to the
levels of contaminant polybrominated dibenzofurans in the mixtures, which in turn correlated
with increasing bromine content of the congeners.
The results from this study confirm in vitro data, suggesting that PBDEs are, at best,
weak activators of the Ah receptor. These results also raise the possibility that brominated
dibenzofuran impurities identified in the congeners studied may, in some cases, have
confounded the results from other studies.
4.4.2.2. Other CYP-450 Inducing Receptors
The study of CYP-450 mRNA expression in rat liver by Sanders et al. (2005) (see section
3.3) found that expression of CYP-2B was up-regulated by BDE-99 in F344 rats to a greater
extent than was CYP-1 Al, a biomarker for the activation of the Ah receptor. Up-regulation of
CYP-3 A was also observed. CYP-2B and -3 A are biomarkers for activation of the constitutive
androstane receptor (CAR) and pregnane X receptor (PXR), respectively. In the case of
BDE-99, the effect on CAR was greater than that on PXR. The CAR and PXR are both involved
in the metabolism of xenobiotics and are stimulated by phenobarbital. The CAR receptor is also
involved in steroid metabolism. The impact of BDE-99 on these receptors is similar to the
44
-------
impact of noncoplanar PCBs on the same receptors. Little is known about the physiological
effects of PXR and CAR receptors.
Pacyniak et al. (2007) carried out additional work with the PXR and its human
counterpart, the steroid X receptor (SXR), by using HepG2 cells transvested with the appropriate
cDNA, the receptor response elements, and a luciferase reporter vector. The cultured cells were
exposed to 0, 0.1, 1, 10, or 100 uM concentrations of BDE-99. With the PXR there was an
increase in relative luciferase activity that showed a significant increase above the control for all
tested concentrations; however, the increase above control was generally similar for all and was
not concentration related. With the SXR, there was a linear significant response to dose for both
the 10 and 100 uM concentrations but not for the 0.1 and 1 uM concentrations. The authors also
compared the response in PXR knock-out mice (10-12 weeks old) with the wild-type mice and
found that CYP-3 Al 1 was induced to a similar extent in the wild-type and control animals,
suggesting that the PXR was not activated by BDE-99. The authors suggested that the fact that
CYP-3 Al 1 protein was up-regulated even in the PXR knock-out mice could be related to
activation of the CAR.
4.4.2.3. Estrogen Receptors
Studies have also been conducted to evaluate the interaction between PBDEs and the ER
sites. Activation of ERs induces cell division in female reproductive organs, mammary glands,
and liver. Receptor-induced mitogenic activity has been linked to tumor formation in the
affected organs (Klaassen, 1996).
The in vitro estrogenic and antiestrogenic potencies of 17 PBDEs, including BDE-85,
-99, -100, and -119 and three hydroxylated PBDEs, were investigated in a human T47D breast
cancer cell line stably transfected with an ER-dependent luciferase reporter gene or human
embryonic kidney cells stably transfected with an ER-a or ER-P luciferase reporter gene (Meerts
et al., 2001). Cells were trypsinized and seeded in 96 well-plates for the ER-CALUX (Chemical-
Activated LUciferase expression) assay. After allowing for cell growth, the wells were exposed
to solutions containing the test compounds or estradiol and were incubated. The luciferase
activity was measured with a luminometer. BDE-100 and -119 showed estrogenic potencies in
the assay, with concentrations leading to 50% induction (median effective concentration [EC50])
of 2.5 and 3.9 jiM, respectively, in comparison to an EC50 value of 1.0 x 10~5 jiM for estradiol.
These pentaBDEs were, respectively, 250,000 and 390,000 times less potent than estradiol.
BDE-85 and -99 did not show any estrogenic activity in the ER-CALUX assay (Meerts et al.,
2001).
Several hydroxylated derivatives of PBDEs (>99% purity) were also evaluated in the
CALUX assay described above. 2,4,6,3',5'-Pentabromo-4'-hydroxy-DE (2,6-dibromo-4-(2,4,6-
tribromophenoxy-phenol), a T4-like hydroxylated-BDE, demonstrated no estrogenic activity up
45
-------
to concentrations of 10 jiM. Antiestrogenic potency was determined in the ER-CALUX assay by
treating T47D.Luc cells with various concentrations of PBDEs in the presence of estradiol. The
four pentaBDEs (BDE-85, -99, -100, and -119) and the T4-like hydroxylated-BDE compound did
not show antiestrogenic activity (Meerts et al., 2001).
Villeneuve et al. (2002) examined the ability of 10 different PBDEs, including BDE-99,
-100, -105, and -126 (99% purity), to initiate ER-mediated gene expression in vitro. At
concentrations up to 500 ng/mL, all pentaBDEs tested failed to induce ER-mediated gene
expression in MVLN recombinant human breast carcinoma cells, using a luciferase response
element for detection. Overall, the PBDEs tested were found to be 50,000 times less potent than
estradiol for inducing ER-mediated gene expression.
Villeneuve et al. (2002) also studied the ability of PBDEs to displace steroid hormones
from serum proteins. At concentrations up to 833 ng/mL, the pentaPBDEs tested in this study
did not show an appreciable capacity for displacing 3H-steroids from carp serum proteins that
had been stripped of hormones before testing. Unlabeled estradiol and testosterone also had a
limited effect on displacing the radiolabeled ligands, suggesting limited sensitivity of the assay
with carp serum.
Another aspect of the possible impact of PBDEs on estrogen (estradiol) was investigated
by Kester et al. (2002). In this instance, the authors studied the effect of hydroxylated PBDEs on
the activity of the human sulfotransferases that metabolically inactivate estrogen. Inhibition of
the sulfotransferases would increase the half-life of estradiol and facilitate increased
opportunities for receptor site stimulation. In this study, the human sulfotransferase that is active
in liver, endometrium, mammary gland, and testes was incubated with various concentrations of
4-hydroxy PBDE congeners. Tri-, tetra-, and pentaBDE hydroxy congeners were evaluated by
using concentrations of 0-1,000 nM. All three compounds tested acted as inhibitors of the
enzyme. The pentaBDE hydroxy congener (2,4,6,3',5'-pentabromo-4'-hydroxy-DE) was the most
effective inhibitor of the three tested compounds, causing approximately 90% inhibition at the
highest concentration. The authors hypothesized that the presence of bromine residues on the
two carbons adjacent to the hydroxyl grouping increased the likelihood of inhibition. A
Lineweaver-Burk analysis of the penta-compound data suggests that the inhibition was
noncompetitive (i.e., the interaction with the enzyme did not involve the active site). The
median inhibitory concentration (IC50) for the inhibition was 150 nM.
In summary, the mechanistic studies of the ER indicate that the activities of the
pentaBDEs are much lower than the activities of dioxin and PCBs. Receptor-site-mediated
activity via the ER site appears to be minimal for the pentaBDEs.
46
-------
4.4.2.4. Androgen Receptors
DE-71, a commercial pentaBDE mixture, was found by Stoker et al. (2004) to delay
puberty and suppress the growth of androgen-dependent tissues in male Wistar rats exposed to
doses of 30 or 60 mg/kg during the peripubertal period but not to doses of 0 or 3 mg/kg. In order
to examine which components of the mixture might be responsible for the observed effects,
androgen receptor binding by several of the individual congeners found in DE-71 was examined
in vitro (Stoker et al., 2005). The assays examined competitive binding of BDE-99 (98% purity)
and BDE-100 (100% purity) congeners in the presence of a tritium-labeled androgen agonist
(R1881) by using ventral prostate cytosolic extracts along with an assay in an MDA-kb2 cell line
containing the human androgen receptor and a transfected luciferase reporter element.
In the assay with the ventral prostate extract, 0.001, 1.6, 3.3, 16.7, or 33 jiM
concentrations of BDE-99 and -100 were incubated in the presence of 1.0 nM R1881 (an agent
that blocks the progesterone and glucocorticoid receptors ) and 10 jiM triamcinolone acetonide.
Both congeners acted as competitive inhibitors for the binding of R1881, but the activity of
BDE-100 was more potent than that of BDE-99. The approximate IC50 for BDE-99 was 33 |iM,
while BDE-100 had 98% inhibition at the same concentration.
In the assay using the MDA-kb2 cell line, BDE-99 and -100 were introduced at
concentrations of 10 pM, 10 nM, 1 jiM, or 5 jiM in the presence of 0.1 nM of the receptor
agonist dihydrotestosterone (DHT). BDE-100 demonstrated a concentration-dependent
antiandrogenic activity in this assay, with a 50% decrease in DHT activity at the 5 jiM
concentration. BDE-99 did not exhibit antiandrogenic activity in this assay.
4.4.2.5. Acetylcholine Receptors
Several studies have examined the impact of BDE-99 on acetylcholine receptors in the
hippocampus. Data have been collected on the activity of both the nicotinic and muscarinic
acetylcholine receptors. Nicotinic receptors are located in skeletal muscle and neurons. The
muscarinic receptors are found in smooth muscle, glands, and the central nervous system
(Klaassen, 1996). Interaction of acetylcholine with the appropriate receptor is responsible for
neuronal activation of muscle contraction along with learning and memory (Ankarberg, 2003).
In rats and mice, the most active period for development of the cholinergic system occurs
in the 3-week period after birth. There are several subfamilies of both muscarinic and nicotinic
receptors. Some display high-affinity binding properties and others low-affinity binding
(Ankarberg, 2003). Acetylcholine receptors are found in a number of areas in the brain,
including the cortex, cerebellum, hippocampus, striatum, and thalamus (Ankarberg, 2003). The
BDE-99 studies that have examined acetylcholine receptor binding have primarily utilized the
hippocampal tissues.
47
-------
A study of the impact of postnatal nicotine exposure by Ankarberg (2003) evaluated the
hypothesis that nicotine exposure during the developmental brain growth spurt period would
affect the development of the cholinergic system and change adult responses to cholinergic
agents. Using nicotine as the cholinergic binding agent, Ankarberg (2003) found that the
maximum impact on the cholinergic system occurred with exposures on PNDs 10-14 and not on
PNDs 3-7 or 19-23. Hippocampal tissues from the nicotine-treated animals were evaluated
24 hours after exposure of the pups and adult animals by using radiolabeled a-bungarotoxin, a
nicotinic receptor antagonist, and quinuclidinyl benzilate (QNB), a muscarinic receptor
antagonist. Nicotine is an acetylcholine receptor stimulant. There was a decrease in the low
affinity receptor-binding sites in all adult animals compared with pups exposed to nicotine
24 hours after exposure, suggesting a decline in receptors with age. However, only animals that
had been exposed to nicotine on PNDs 10-14 had receptor levels as adults that were
significantly lower than those of adult controls. Low-affinity-binding sites appeared to be
affected to a greater extent than high-affinity sites (Ankarberg, 2003).
The results from the study by Ankarberg (2003) and knowledge of the plasticity of the
cholinergic system in mice and rats during the postnatal period provided an incentive for
examining cholinergic receptors in BDE-99-exposed mice. Viberg et al. (2004b) evaluated the
nicotinic receptors in the hippocampus of adult mice after postnatal exposure to BDE-99 as part
of a neurobehavioral study. Single oral doses of 0, 0.2, 0.4, or 12 mg/kg of BDE-99 in a 20% fat
emulsion were given by gavage to male NMRI mice on PND 10, and the habituation response of
the animals was evaluated at 4 months of age. One week after completion of the behavioral
tests, the mice in the control and 12 mg/kg groups were sacrificed, and measurement of nicotine-
binding sites in the hippocampus was performed by using 3H-labeled a-bungarotoxin. Specific
binding was determined by calculating the difference in the amount bound in the presence versus
absence of a-bungarotoxin. There was significant decrease (31%) in a-bungarotoxin binding in
the hippocampus of adult mice given 12 mg/kg BDE-99 on PND 10 compared to the density in
control animals, indicating effects on the nicotinic receptors in the brain.
As part of another neurobehavioral study, Viberg et al. (2005) examined the binding of
tritium-labeled QNB to muscarine-like binding sites in the hippocampus. Single oral doses of 0,
0.8, 8.0, or 16 mg/kg BDE-99 (purity >98%) in a 20% fat emulsion (1:10 egg lecithin to peanut
oil) were given by gavage to male Sprague-Dawley rats on PND 10. Control mice received
10 mL/kg of the 20% fat emulsion. Behavioral tests were administered at 2 months of age. One
week after completion of the behavioral tests, the rats in the 0, 8.0, and 16 mg/kg BDE-99 groups
were sacrificed, and measurement of muscarine-like binding sites was performed. Specific
binding was determined by calculating the difference in the amount of QNB bound in the
presence versus absence of atropine, a known inhibitor of muscarinic receptors (Klaassen, 1996).
48
-------
There was a significant decrease in the density of specific [3H]-QNB binding sites in the
hippocampi in rats given 16 mg/kg, while no difference was seen in rats treated with 8.0 mg/kg.
The results of the studies by Viberg et al. (2005, 2004b) in conjunction with the work by
Ankarberg (2003) support the concept that exposure to BDE-99 during a critical window in
postnatal development may result in a decrease in selected acetylcholine receptors in the brain
and that these changes may contribute to some or all of the observed neurobehavioral responses
exhibited in exposed animals as adults. Additional studies of receptor responses to the PBDEs,
using the hippocampus and other regions of the brain, are warranted. The limited data available
indicate that the effects on habituation were only seen at doses that also cause decreased binding
of the cholinergic receptor antagonists.
4.4.3. Thyroid Effects
Because PBDEs have some structural similarity to the thyroid hormone T4, it has been
suggested that they may interfere with thyroid hormone transport by competitively binding with
transthyretin (TTR), one of the thyroid hormone-binding transport proteins in the plasma of
vertebrate species. The possible interference of several pentaBDEs with T4-TTR binding was
investigated in an in vitro competitive binding assay, using human TTR and 125I-labeled T4 as the
displaceable radioligand. The four pentaBDE congeners evaluated (BDE-85, -99, -100, and
-119) did not compete with T4-TTR binding (Meerts et al., 2000).
Meerts et al. (2000) also tested these four pentaBDEs before and after incubation with
differently induced hepatic microsomes to examine the ability of their hydroxylated metabolites
to displace T4 from TTR. The pentaBDEs were individually incubated with liver microsomes
prepared following induction with phenobarbital (a CYP-2B inducer), p-naphthoflavone (a
CYP-1A inducer), or clofibrate (a CYP-4A3 inducer). Incubation of the pentaBDEs with
CYP-2B-enriched rat liver microsomes resulted in the formation of metabolites that were able to
displace 125I-T4 from TTR. The metabolites of BDE-100 and -119 were able to displace more
than 60% of the 125I-T4 from TTR. BDE-85 and -99 showed a lower ability to displace 125I-T4
from TTR (20-60%). No T4-TTR displacement by pentaBDEs occurred after incubation with
liver microsomes enriched with CYP-1 A or -4A3. PentaBDEs are therefore able to compete
with T4-TTR binding only after metabolic conversion by induced rat liver microsomes,
suggesting an important role for hydroxylation. The relevance of this observation for humans
has yet to be resolved. T4-binding globulin, rather than TTR, is the major T4-binding protein in
humans.
As part of the Darnerud et al. (2005) study of the impact of a CB3 infection on the
distribution of BDE-99, plasma total T4 levels were monitored. On day 3 of the infection,
decreases were seen in T4 levels (33%). However, infections seem to be associated with a
49
-------
decreased release of circulating T4; thus, the decrease in T4 seen in this experiment could be
14,
unrelated to C-BDE-99 exposure.
4.4.4. Neurotoxicity
The effects of BDE-99 on the developing brain were investigated by Aim et al. (2006).
Neonatal NMRI mouse pups were given a single 12 mg/kg dose of emulsified BDE-99 or vehicle
(a 20% emulsion of egg lecithin and peanut oil, 1:10, in water) by gavage on PND 10, during the
rapid brain growth spurt. Both groups were sacrificed 24 hours after dosing, and brains were
rapidly excised. Striatal and hippocampal tissues from the brains of three mice were separately
pooled, homogenized, and cleaned to remove lipids and nucleic acids. Samples were analyzed
to determine total protein content and labeled using different cyanine dyes for the controls and
for the treated tissues. There were three pooled tissue samples from the controls and from
treated animals for both the striatal and hippocampal tissues. Two additional pooled replicate
samples were prepared from controls and treated animals with unlabeled protein from the pool
of control animals added to assist in protein identification. Four gels were run per brain region.
The proteins were separated by two-dimensional fluorescence difference gel electrophoresis.
Separation in the first dimension employed isoelectric focusing. Separation in the second
dimension was done with polyacrylamide gel electrophoresis. There was considerable similarity
in the gels from the striatum and hippocampus. There were 685 spots common to the four
striatal gels and 651 spots common to the four gels from the hippocampus. From these common
protein spots, 40 differentially expressed striatal proteins and 56 proteins from the hippocampus
were selected for further analysis.
The gel spots selected for further analysis were removed; the protein was extracted and
subjected to trypsin digestion. The resultant peptides were analyzed by using time-of-flight mass
spectrometry and identified by means of the National Center for Biotechnology Information
nonredundant database and the MASCOT search engine. Nine spots from the striatum and
10 from the hippocampus were identified in this fashion. Of the nine spots identified in the
striatum, three were up-regulated and six were down-regulated. All 10 proteins identified in the
hippocampus were up-regulated. Mortalin, a heat-shock protein up-regulated in the striatum and
in the hippocampus, was the only protein common to both tissues. Two of the striatal spots were
neuromodulins and three were stathmins. The neuromodulins play a role in guiding the growth
of axons and forming new neural connections. Both neuromodulins were up-regulated in the
BDE-99 exposed mice. Stathmins are also associated with neurite growth. They were down-
regulated in the BDE-99 exposed mice. Both of these proteins are substrates for protein kinase
C (PKC), an enzyme that functions in neuronal growth, learning, and memory. In the
hippocampus, two y-enolases, a-enolase, adenosine triphosphate synthase, a mitochondrial
hydrogen ion transporter, and isocitrate dehydrogenase were all up-regulated. Several of these
50
-------
proteins participate in PKC signaling complexes and/or are involved in cellular energy
production. The enolases have also been observed in brain synaptic terminals.
Identification of some of the differentially expressed proteins in the brains of BDE-99-
treated mice compared with controls does not fully explain the mode of action for the
neurodevelopmental effects of BDE-99. It does indicate that brain proteins in the hippocampus
and striatum of the exposed animals were different from those in the controls during the brain
growth spurt period. Several of the differentially expressed proteins are linked to the PKC
signaling cascade, a system that functions in neuronal growth, learning, and memory. The
authors suggested that the proteins identified may be biomarkers that will be useful in additional
studies of the early-life changes in brain development precipitated by BDE-99 exposure.
Activation of PKC has been suggested to be involved in the neurotoxicity of PCBs.
Madia et al. (2004), therefore, examined whether BDE-99 and Aroclor 1254 would cause
translocation of PKC a, s, and C, from cytosol to the membrane in astrocytoma cells. BDE-99
caused translocation of the three PKC isozymes present in astrocytoma cells, while Aroclor 1254
affected only PKC a and e translocation. The ability of BDE-99 and Aroclor 1254 to induce
apoptosis in astrocytoma cells was also investigated. BDE-99, but not Aroclor 1254, caused
apoptotic cell death in astrocytoma cells. These results indicate that the overall pattern of
cytotoxicity of BDE-99 to human astrocytoma cells is different from that of Aroclor 1254,
suggesting that these two compounds may also have different effects in vivo.
As described in section 3.2.2, Kodavanti et al. (2005) carried out a study using cultures of
cerebellar granule cells from 7- to 8-day-old Long-Evans rat pups. The cultures were treated
with 14C-labeled PBDE-99 (0.05 uCi/ml) combined with different concentrations of unlabeled
compound (0-30 uM) for 15 minutes to 1 hour. For each concentration tested, there was a linear
increase in percent accumulation over the 1-hour exposure period. When time was held constant
and concentration varied, the percent accumulation increased only at the low concentrations,
suggesting saturation of uptake. A similar pattern was observed for PKC translocation in the
cerebellar granule neurons where 3H-phorbol ester binding was increased at a PBDE-99
concentration of 10 jiM but then remained fairly constant at 30 and 60 jiM.
4.4.5. Immunotoxicity
Mitogen-induced DNA synthesis and immunoglobulin G (IgG) synthesis by human
lymphocytes were examined after exposure to BDE-85 (>98% purity) in vitro in order to
determine the immunotoxic potential of this substance (Fernlof et al., 1997). Human peripheral
lymphocytes were isolated from blood donated by 15 healthy females. The lymphocytes were
cultured and utilized to assay radiolabeled-deoxythymidine uptake in response to pokeweed
mitogen stimulation. In addition, the supernatants from the culture media were examined for the
presence of immunoglobulin by using an antihuman IgG from goats. No effects on pokeweed
51
-------
mitogen-induced DNA proliferation or IgG synthesis were observed in human lymphocytes after
exposure of c
in this assay.
exposure of cells to 10 9 to 10 5 M BDE-85, indicating that this congener was not immunotoxic
4.4.6. Cytotoxicity
The cytotoxicity of BDE-99 was assessed in human astrocytoma cells and compared with
that of Aroclor 1254 (Madia et al., 2004). The mitochondrial activity that reduces 3-[4,5-
dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) was used to assess cell survival in
a quantitative colorimetric assay. Cytotoxicity was also evaluated by measuring the release of
lactate dehydrogenase (LDH) in the culture medium. To further determine the effects of
treatments on cell survival, cells were treated with trypan blue and counted with a
hemocytometer. BDE-99 and Aroclor 1254 caused comparable concentration-dependent
inhibition of MTT reduction. Aroclor 1254 caused significant release of LDH at the two highest
concentrations (50 and 100 |iM), while BDE-99 did not cause any change in this parameter.
Direct counting of dead cells with trypan blue staining provided similar results (i.e., Aroclor
1254 was toxic at high concentrations, while BDE-99 was not).
4.4.7. Genotoxicity
Evandri et al. (2003) studied the reverse mutation activity of BDE-99 by using
Salmonella typhimurium strains TA98 and TA100, Escherichia coli WP2 uvrA, and the
chromosome aberration test in Allium cepa. BDE-99 was nontoxic in bacteria at the highest
dose tested (0.305 mg/plate). The number of revertant colonies was comparable to the solvent
control group, both with and without S9. BDE-99 was also not clastogenic in the A. cepa test at
concentrations up to 100 jiM (56 mg/L). The number of structural chromosome aberrations
induced by BDE-99 was not significantly different from that of the control.
4.5. SYNTHESIS OF MAJOR NONCANCER EFFECTS
4.5.1. Oral
Alterations of behavioral parameters (i.e., impaired motor functions worsening with age)
have been shown to occur in male and female mice and rats orally exposed prenatally and
neonatally to BDE-99 (Kuriyama et al., 2005; Viberg et al., 2005, 2004a, b; Branch! et al., 2002;
Eriksson et al., 2002, 2001). Effects on spontaneous motor behavior were not species, gender, or
strain specific and were induced during a defined and narrow developmental window in which
rodents seemed to be uniquely susceptible to the neurodevelopmental effects of BDE-99
(Eriksson et al., 2002). Similar neurodevelopmental effects have been observed in studies of the
tetra (BDE-47), hexa (BDE-153), and deca (BDE-209) congeners. As indicated in the
Guidelines for Neurotoxicity Risk Assessment (U.S. EPA, 1998a), it is assumed that an agent that
52
-------
produces detectable adverse neurotoxic effects in experimental animal studies will pose a
potential hazard to humans. For BDE-99, in the absence of human evidence, behavioral changes
in experimental animal studies are assumed to indicate concern in humans.
Treatment of rats with BDE-99 on GD 6 resulted in a dose-dependent decrease in daily
sperm production, spermatid count, and relative epididymis weight in rat offspring at a dose as
low as 0.06 mg/kg BDE-99. However, no effects were seen on male fertility or sperm
morphology at these doses (Kuriyama et al., 2005). The decreased sperm production, spermatid
count, and epididymis weight warrant additional studies to determine their potential significance
to humans.
The effects of BDE-99 on the female reproductive system were evaluated in rats
(Talsness et al., 2005). Histologic changes in the ovaries and vaginal epithelium were seen at a
dose as low as 0.06 mg/kg BDE-99 but were not associated with statistically significant effects
on fertility (pregnancy rate, mean implantation sites per dam, live fetuses per dam, and
resorption rate).
Data from the studies of Ceccatelli et al. (2006) and Lilienthal et al. (2006), conducted
utilizing the subcutaneous exposure route and doses of 0, 1, or 10 mg/kg-day, suggest that
BDE-99 may have a subtle impact on sexual development in males and females, possibly
mediated through the steroid sex hormones. However, the effects observed were not consistently
dose related.
BDE-99 has been found in human milk, maternal and cord blood, and adipose tissues.
Concentrations found are high in all human biological samples in the U.S. relative to other
countries. Fetuses and infants are exposed to BDE-99. Whether such exposures constitute a
health risk at adulthood for neurodevelopmental dysfunction or adverse reproductive effects is
not known at this time. An association between prenatal or neonatal exposures to BDE-99 and
neurobehavioral or reproductive effects in humans has not been established.
4.5.2. Inhalation
No data are available on the toxicity of BDE-99 by the inhalation route of exposure.
4.5.3. Mode-of-Action Information
Researchers from the laboratory of Eriksson/Viberg have hypothesized that the observed
effects on locomotion and habituation are related to impaired development of the cholinergic
system during the postnatal brain growth spurt period (Viberg et al., 2005, 2004b, 2003a). They
have further hypothesized that the sensitivity of the cholinergic system occurs in the vicinity of
PND 10 and have tested this hypothesis by varying the time of dosing and observing differences
in the habituation effect for BDE-99 and -209 (Viberg et al., 2007; Eriksson et al., 2002). In rats
and mice the critical postnatal brain growth period occurs within the first few weeks after birth,
53
-------
while in humans it occurs in the last trimester of pregnancy and continues throughout the first
year of life. PNDs 10-14 appear to be a period of maximum vulnerability for the developing
cholinergic system that coincides with the most pronounced neurodevelopmental effects from
BDE-99 exposure.
There is evidence that BDE-99 interacts with brain tissues from studies reported by Aim
et al. (2006) in which the striatum and hippocampus of mice exposed to BDE-99 on PND 10
showed distinct differences in the proteins expressed from those seen in controls. Several of the
proteins that seemed to be biomarkers for BDE-99 exposure are linked to the PKC signaling
cascade. PKC signaling plays a role in neuron development, memory, and learning. Additional
neurological interactions were observed in an in vitro study of BDE-99 uptake by cultured
cerebellar granule neurons from Long-Evans rat pups that demonstrated that low BDE-99
concentrations caused translocation of PKC (Kodavanti et al., 2005). Although evidence exists
that BDE-99 and other PBDEs interact at the neurological level, data are inadequate to
determine the mode of action for BDE-99.
Exposure of rats to BDE-99 resulted in an increase of total T4 plasma levels 3-6 days
following exposure but returned to normal levels by 12 days after exposure (Hakk et al., 2002a).
Serum total and free T4 levels of male mouse offspring of dams treated with BDE-99 from GD 6
to PND 21 were not found to be statistically different from control levels on PND 22 (Branch! et
al., 2005). In a study in mice (Skarman et al., 2005), BDE-99 administered from GD 4 to
PND 17 had no effect on plasma T4 levels in dams and their offspring relative to controls at any
sampling occasion. The only effect noted in these in vivo studies was, therefore, a transient
increase in T4 plasma levels (Hakk et al., 2002a).
It is known that thyroid hormones are essential for brain development and that decreases
in thyroid hormone levels during fetal and early neonatal life may have profound adverse effects
on the developing brain (Morreale de Escobar et al., 2000). The limited data in humans (Mazdai
et al., 2003) and the available data in mice and rats (Branch! et al., 2005; Skarman et al., 2005;
Hakk et al., 2002a) do not seem to indicate that BDE-99 interferes with thyroid hormone
homeostasis. However, thyroid hormone levels and behavioral activity were not comeasured in
any of the developmental toxicity studies in mice or rats.
Hydroxylated pentaBDE metabolites have been shown in vitro to compete with T4 for
binding with high affinity to TTR. Meerts et al. (2000) indicated that pentaBDEs are able to
compete with T4-TTR binding only after metabolic conversion by induced rat liver microsomes,
suggesting an important role for hydroxylation. The relevance of this observation for humans
has yet to be resolved. T4-binding globulin, rather than TTR, is the major T4-binding protein in
humans. Despite the possibility that BDE-99 interacts with TTR, there are no mode-of-action
data that link thyroid effects to the reported neurobehavioral effects observed in rodents
54
-------
(Kuriyama et al., 2005; Branch! et al., 2005, 2002; Viberg et al., 2005, 2004a, b, 2002;
Ankarberg, 2003; Eriksson et al., 2002, 2001).
Other observations include the absence of effects on pokeweed mitogen-induced DNA
proliferation or IgG synthesis in human lymphocytes after exposure of cells to the pentaBDE-85
congener, indicating that this congener was not immunotoxic in this assay. BDE-99 was not
mutagenic in S. typhimurium or E. coll assays, with and without S9, or in the A. cepa
chromosome aberration test (Evandri et al., 2003). Studies of pentaBDE interactions with the
Ah, estrogen, and androgen receptors indicate that these compounds are considerably less potent
than dioxins and PCBs (Chen and Bunce, 2003; Villeneuve et al., 2002; Chen et al., 2001). The
implications of these results are unknown.
4.6. EVALUATION OF CARCINOGENICITY
There is inadequate information to assess the carcinogenic potential of BDE-99 (U.S.
EPA, 2005a,b). Epidemiological studies of exposure to BDE-99 and cancer occurrence in
humans are not available. Animal chronic toxicity/carcinogenicity studies have not been
conducted for BDE-99. BDE-99 was not mutagenic in S. typhimurium or E. coli assays, with
and without S9, or in the A. cepa chromosome aberration test (Evandri et al., 2003). Additional
in vitro or in vivo studies are not available to determine the full genotoxic potential of BDE-99.
4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
4.7.1. Possible Childhood Susceptibility
A population subgroup is susceptible if exposure occurs during a period of sensitivity, as
observed in mice and rats exhibiting alterations of neurobehavioral functions following prenatal
and neonatal exposure to BDE-99 (Kuriyama et al., 2005; Viberg et al., 2004a, b; Branch! et al.,
2002; Eriksson et al., 2001). The neonatal stage is a period of rapid development of the nervous
system and is considered a critical window for brain development. The animal model indicates a
potential for concern for early lifetime exposure (i.e., fetal or infant exposure) to the chemical.
The identification of BDE-99 in human milk, umbilical cord serum, and children's serum
(Mazdai et al., 2003; Schecter et al., 2003; Thomsen et al., 2002) implies humans are exposed to
BDE-99 during a critical window of rapid development of the brain, indicating a potential for
susceptibility. Whether such exposure constitutes a health risk for adverse neurodevelopmental
effects in children is not known at this time because of the limited toxicological database for
BDE-99. An association between prenatal or neonatal exposures to BDE-99 and
neurobehavioral dysfunction in humans has not been established.
55
-------
4.7.2. Possible Gender Differences
Most of the neurobehavioral studies were conducted in male rodents. In the
neurobehavioral studies conducted in both sexes of mice and rats (Kuriyama et al., 2005; Viberg
et al., 2004a; Branch! et al., 2002), there was no difference in neurobehavioral response in male
and female animals from exposure to BDE-99. There is no indication that susceptibility to
BDE-99 differs in male and female humans or in experimental animals.
56
-------
5. DOSE-RESPONSE ASSESSMENTS
5.1. ORAL REFERENCE DOSE (RfD)
5.1.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
Table 5-1 summarizes the oral reproductive and developmental toxicity studies on
BDE-99 that were candidates for use in the derivation of an RfD for this chemical. Other
toxicity studies were also considered for use in deriving the RfD but were rejected for various
reasons. These studies included a short-term toxicity study in mice (Skarman et al., 2005) and
an acute toxicity study in rats (Hakk et al., 2002a). In these studies, hepatic mixed-function
oxidase system enzyme activities and/or plasma thyroid hormone levels were measured
following exposure to BDE-99 (see section 4.2.1). Changes in the activity of the mixed-function
oxidase system enzymes, however, often accompany exposure to xenobiotic compounds and are
not definitively adverse since enzyme activities are inducible and usually return to normal levels
following cessation of exposure. In regard to hormone levels, the transient increase in T4 levels
seen in the single-dose study of Hakk et al. (2002a) was not confirmed in the longer duration
study of Skarman et al. (2005). Thus, neither of these studies is a good candidate for use in
dose-response assessment, and additional studies of the impact of BDE-99 on thyroid hormones
are warranted.
Of the studies summarized in Table 5-1, Viberg et al. (2004a) was selected as the
principal study, and neurobehavioral developmental effects were identified as the critical effect
for deriving an RfD for BDE-99. The principal study and critical effect were selected after
careful evaluation of all the available toxicity studies, including consideration of whether the
data were amenable to BMD modeling (see section 5.1.2). The primary reasons for selecting
Viberg et al. (2004a) were as follows: (1) several different dose levels of BDE-99 were
employed, (2) quantitative dose-response data were available with which to conduct BMD
modeling, (3) good model fits were obtained in subsequent BMD modeling, (4) a clear NOAEL
was identified from this study, and (5) the results of this study are supported by several other
studies in mice. In Viberg et al. (2004a), male and female mice were administered single oral
doses (0, 0.4, 0.8, 4.0, 8.0, or 16 mg/kg) of BDE-99 on PND 10. The NOAEL identified in this
study was 0.4 mg/kg. Adverse effects noted in 2-, 5-, and 8-month-old mice at the next highest
dose, 0.8 mg/kg, included hypoactive spontaneous motor behavior in the beginning of the test
period, hyperactive behavior at the end of the test period, and decreases in habituation capability,
with these behavioral disturbances becoming more pronounced with increasing age. By
administering BDE-99 on PND 10, animals were exposed during a period of maximum
vulnerability of the developing mouse brain (Eriksson et al., 2002). Exposures to BDE-99 at
57
-------
older ages (i.e., 2 months and up) did not cause such behavioral changes in mice (Branch! et al.,
2005).
Table 5-1. Summary of oral reproductive/developmental toxicity studies of
BDE-99
Species,
strain,
sex
Mouse,
C57/B1,
M&F
Mouse,
NMRI,
M
Mouse,
NMRI,
M
Mouse,
CD-I
Swiss,
M&F
Rat,
Wistar,
M&F
Rat,
Sprague-
Dawley,
M
Rat,
Wistar,
F
Duration,
purity
Single
dose on
PND 10,
>99%
Single
dose on
PND 10,
>98%
Single
dose on
PND 10,
>98%
GD6to
PND 21,
purity not
specified
Single
dose on
GD6,
98%
Single
dose on
PND 10,
>98%
Single
dose on
GD6,
98%
Dose levels
(mg/kg-day)
0,0.4,0.8,
4.0, 8.0, or
16
0, 0.2, 0.4,
or 12
0, 0.8, or 12
0, 0.6, 6, or
30
0, 0.06, or
0.3
0,0.8,8.0,
or 16
0, 0.06, or
0.3
NOAEL
(mg/kg-day)
0.4
0.4
Not
identified
Not
identified
Not
identified
Not
identified
0.8
Not
identified
0.3
LOAEL
(mg/kg-day)
0.8
12
0.8
Not
identified
0.06
0.06
8.0
0.06
Not
identified
Observed effects
Neurobehavioral developmental
effects; significant dose-related
changes in spontaneous motor
behaviors; decreased habituation
capability with increasing age
Neurobehavioral developmental
effects and effects on the
cholinergic system; significant
dose-related changes in
spontaneous motor behavior;
decreased habituation capability
with increasing age
Neurobehavioral developmental
effects; significant dose-related
changes in spontaneous motor
behavior & learning/memory
ability; decreased habituation
capability with increasing age
Transient hyperactivity in absence
of dose-response relationship
Neurobehavioral developmental
effects; effects on locomotion
(hyperactivity), increasing with
increasing age
Decreased daily sperm
production, spermatid count, and
relative epididymis weight in
offspring; no effect on fertility at
any dose
Neurobehavioral developmental
effects; significant dose-related
changes in spontaneous motor
behavior
Qualitative histologic changes in
ovaries and vaginal epithelium in
offspring
No significant effect on fertility at
any dose
Reference
Viberg et al.
(2004a)
Viberg et al.
(2004b)
Eriksson et
al. (2001)
Branchi et
al. (2002)
Kuriyama et
al. (2005)
Viberg et al.
(2005)
Talsness et
al. (2005)
Although Viberg et al. (2004a) was selected as the principal study, several concerns exist
regarding the design of this study. First, the protocol was unique and did not conform to current
58
-------
health effects testing guidelines for neurotoxicity screening batteries or developmental
neurotoxicity studies (U.S. EPA, 1998b). Second, the dosing regimen did not encompass
exposure via gestation and/or lactation (U.S. EPA, 1998a), with only single doses being
administered. In some respects, the fact that effects occurred with such limited dosing argues for
the sensitivity of this study. While dosing appears to have been performed during a critical
window of developmental susceptibility, this design is inadequate to determine the effects of
longer-term dosing. Extrapolating the results of this study to more traditional dosing regimens is
problematic, particularly with regard to the potential impact of in utero and postnatal exposure.
Another concern is that, based on the methodology provided in the published report, more than
one pup per litter was used for the behavioral testing (eight mice were randomly selected from
three to five different litters in each treatment group). This methodology may increase the
number of pups from the same litter, which may bias the analyses towards false positives,
resulting in observed neurobehavioral effects attributable to differences that are not treatment
related in pups born to a single dam. Another concern regarding the study design was the limited
number of neurobehavioral parameters that were assessed. The absence of a full functional
observational battery (FOB) limits the ability of this study to correlate the reported effects with
other FOB parameters, data which would be helpful in gauging the reliability of the limited
parameters that were measured. As indicated in the Guidelines for Neurotoxicity Risk
Assessment (U.S. EPA, 1998a), it is assumed that an agent that produces detectable adverse
neurotoxic effects in experimental animals will pose a potential hazard to humans. For BDE-99,
in the absence of human evidence, data from experimental animal studies are used as the basis
for the RfD.
While study design issues limit the utility of Viberg et al. (2004a), several additional
considerations support the use of these data. Supporting data that exposure in Viberg et al.
(2004a) occurred during the period of maximum vulnerability of the developing mouse brain
come from Eriksson et al. (2002), who demonstrated that vulnerability of adult mice to
neurodevelopmental effects of BDE-99 exposure occurred during a narrow phase of neonatal
brain development. Furthermore, acute exposure to a highly lipophilic chemical that possesses
an extended half-life, such as BDE-99, will result in systemic exposure that lasts much longer
than suggested by the exposure duration. In addition, there are a wide variety of brain structures
that have very limited critical windows during development. These critical windows result in
susceptible periods of exposure that can be very short in duration. The concept that exposure
during critical periods of development can induce functional neurological effects later in
development has been demonstrated with structurally related PBDE congeners, including tetra-,
hexa-, and deca-BDEs (Rice et al., 2007; Kuriyama et al., 2005; Viberg et al., 2004a, b, 2003a,
b; Branch! et al., 2002; Eriksson et al., 2001). Therefore, the observed neurobehavioral effects
are biologically plausible, and thus exposure to BDE-99 may pose a potential hazard to humans
59
-------
(U.S. EPA, 1998a). Taken together, these considerations support the use of Viberg et al. (2004a)
for deriving the RfD for BDE-99.
Supporting studies for neurobehavioral effects in mice from BDE exposure include
Viberg et al. (2004b) in which male mice were administered single oral doses (0, 0.2, 0.4, or
12 mg/kg) of BDE-99 on PND 10. The NOAEL identified in this study was 0.4 mg/kg and the
LOAEL was 12 mg/kg for dose-related changes in spontaneous motor behavior and effects on
the cholinergic system. A similar study by Eriksson et al. (2001) identified a LOAEL for
neurobehavioral developmental effects of 0.8 mg/kg in male mice treated with BDE-99 at doses
of 0, 0.8, or 12 mg/kg on PND 10. In a study by Branch! et al. (2002), transient hyperactivity in
the absence of a clear dose-response relationship occurred in male and female offspring of mice
treated with BDE-99 at 0.6, 6, or 30 mg/kg-day on GD 6 through PND 21.
A supporting study that observed the neurobehavioral developmental effects of BDE-99
in rats include Viberg et al. (2005). In this study, male rats were administered single oral doses
(0.8, 8.0, or 16 mg/kg) of BDE-99 on PND 10. The NOAEL in this study was 0.8 mg/kg and the
LOAEL was 8.0 mg/kg for dose-related changes in spontaneous motor behavior in adult rats.
In a study by Kuriyama et al. (2005), Wistar rat dams were exposed to 0, 0.06, or 0.3 mg/kg
BDE-99 on GD 6. Male and female offspring from these dams showed significant increases in
locomotor activity on PNDs 36 and 71 at both doses. The LOAEL for hyperactivity was
0.06 mg/kg.
Reproductive effects were also examined in the Kuriyama et al. (2005) study. Male
offspring of Wistar rat dams exposed to 0, 0.06, or 0.3 mg/kg BDE-99 on GD 6 showed
significant decreases on PND 140 in daily sperm production and spermatid count at both doses,
but fertility was not affected. However, in rodents, sperm number has to be substantially
reduced before fertility is affected.
The effects of BDE-99 on the female reproductive system have also been evaluated in
rats by Talsness et al. (2005). In this study, qualitative histologic changes in the ovaries and
vaginal epithelium were seen at both doses of BDE-99 tested, 0.06 and 0.6 mg/kg, but were not
associated with significant effects on fertility (i.e., pregnancy rate, mean implantation sites per
dam, live fetuses per dam, and resorption rate).
5.1.2. Methods of Analysis
The RfD for BDE-99 was derived by using the BMD approach by fitting the continuous
models available in BMD software (BMDS) version 1.3.2 to the neurobehavioral data in mice
from Viberg et al. (2004a). In the case of motor activity, no specific magnitude of change exists
that is generally regarded as indicative of an adverse effect. In the absence of consensus on the
level of response that is considered to be adverse, the benchmark response (BMR) selected was a
change in the treatment mean equal to 1 SD of the control mean. In addition, changes in the
60
-------
critical effect mean equal to 0.5 and 1.5 of the control SDs (0.5 and 1.5 SD, respectively) were
also employed as BMRs to evaluate the impact of BMR choice on the BMD and the 95% lower
bound on the BMD (BMDL).
Several BMD analyses were conducted by using relevant endpoints from Viberg et al.
(2004a). The specific data sets used included locomotion habituation ratio for males and
females at 2 and 8 months and rearing habituation ratio for males and females at 2 and 8 months.
The BMD modeling results based on Viberg et al. (2004a) are summarized in Table 5-2, with
further detail provided in Appendix B. No satisfactory model fits were obtained from the
habituation ratio data based on locomotion activity in male and female mice from Viberg et al.
(2004a). The BMD modeling results for the models that exhibited the best fit to the data for
rearing habituation ratio are displayed in Table 5-2.
Based on data from Viberg et al. (2004a), the best -fit model was the power model, fit to
data on rearing habituation in 8-month-old female mice following exposure to BDE-99. Based
on the power model, the resulting BMD1SD was estimated to be 0.41 mg/kg and the
corresponding BMDL1SD was estimated to be 0.29 mg/kg. The BMDs and BMDLs
corresponding to BMRs of 0.5 SD and 1.5 SD were also estimated in order to evaluate the
impact of BMR selection on these model -derived BMDs and BMDLs. Employing the same data
as before, the BMD0 5SD and BMDL0 5SD were estimated to be 0.29 and 0.20 mg/kg, respectively,
while the BMDi 5SD and BMDL! 5SD were estimated to be 0.50 and 0.37 mg/kg, respectively.
Table 5-2. Summary of BMD modeling output results with good data fit in
mice
Reference
Viberg et al. (2004a)
Viberg et al. (2004a)
Viberg et al. (2004a)
Endpoint
Rearing habituation in 2-month-old male mice
Rearing habituation in 2-month-old female mice
Rearing habituation in 8-month-old female mice
Model
Power
Power
Power
BMD1SD
(mg/kg-day)
0.59
0.70
0.41
BMDL1SD
(mg/kg-day)
0.44
0.47
0.29
Sand et al. (2004) also applied the BMD method to the Viberg et al. (2004a) spontaneous
motor behavior data observed in 2-, 5-, and 8-month-old male and female mice exposed orally
on PND 10 to different doses of BDE-99. Additional data not available in the published study of
Viberg et al. (2004a) were used to quantify spontaneous behaviors (i.e., locomotion, rearing, and
total activity) in terms of a fractional response defined as the cumulative response after
20 minutes of the test divided by the cumulative response produced over the whole 1-hour test
period. This fractional response contains information about the time-response profile (which
differed among treatment groups) and was found to have appropriate statistical characteristics.
In this analysis, male and female mice could be characterized by a common dose-response model
(i.e., each sex responded similarly to exposure to BDE-99). By fitting the Hill model to
61
-------
spontaneous motor behavior data observed in 2-month-old male and female mice, the BMDs and
BMDLs corresponding to a BMR of 10% were estimated. Total activity was found to be the
most sensitive neurobehavioral endpoint, with the BMD10 and BMDL10 for this endpoint
estimated to be 0.61 and 0.42 mg/kg, respectively. These values are similar to the BMD1SD of
0.41 and the BMDL1SD of 0.29 mg/kg estimated from the published data in Viberg et al. (2004a)
on rearing habituation in 8-month-old female mice. Values for the BMDo.os and BMDLo.05
corresponding to a BMR of 5% were also estimated by Sand et al. (2004) and were 0.33 and
0.21 mg/kg, respectively.
For comparison purposes, all relevant endpoints from Eriksson et al. (2001) and
Kuriyama et al. (2005) were also modeled. The specific data sets used and detailed BMD
modeling results are presented in Appendix B. No satisfactory model fits were obtained for any
of the endpoints from Eriksson et al. (2001). For the Kuriyama et al. (2005) study, additional
information on the data points, as well as standard deviations of the means for locomotor activity
on PNDs 36 and 71, were obtained from the study authors via e-mail (Ibrahim Chahoud, Charite
University Medical School, Berlin, to Mary Manibusan, U.S. EPA, dated 2004). The data
modeled as a continuous variable were locomotor activity on PND 36. Locomotor activity on
PND 71 was not amenable to modeling in BMDS. The best-fit model for duration of activity per
day (measured on PND 36) was the polynomial model, yielding a BMD1SD of 0.28 mg/kg and a
BMDL1SD of 0.22 mg/kg. The best-fit model for LBI counts per phase was the linear model,
yielding a BMD1SD of 0.16 mg/kg and a BMDL1SD of 0.11 mg/kg.
The BMDL1SD of 0.29 mg/kg estimated from data in mice (Viberg et al., 2004a) and the
BMDL1SD of 0.22 mg/kg estimated from data in rats (Kuriyama et al., 2005) are very similar and
suggest that mice and rats are equally susceptible to the neurobehavioral effects of BDE-99 and
that no significant difference is apparent when the animals are exposed in utero versus
perinatally to BDE-99. The mouse study by Viberg et al. (2004a) is selected for use in the
derivation of the RfD because, as indicated above, in this study: (1) several different dose levels
of BDE-99 were employed, (2) quantitative dose-response data were available with which to
conduct BMD modeling, (3) good model fits were obtained in subsequent BMD modeling, (4) a
clear NOAEL was identified from this study, and (5) the results of this study are supported by
several other studies in mice. Conversely, the Kuriyama et al. (2005) neurobehavioral study in
rats examined only two doses and is supported by only one other neurobehavioral study in rats,
Viberg et al. (2005).
For the reproductive effects observed in the Kuriyama et al. (2005) study, none of the
models in BMDS satisfactorily fit the data on spermatid count, sperm numbers, or daily sperm
production. The only data that could be adequately modeled were percent of adult rats with less
than two ejaculations (see Appendix B). However, no changes in fertility or other functional
reproductive endpoints were observed as a result of these decreased ejaculations. In general, rat
62
-------
fertility is less sensitive to changes in sperm count than human fertility. Therefore, without more
information, the biological significance of this effect is uncertain. For this reason, the BMD
modeling results based on the reproductive effects observed in Kuriyama et al. (2005) were not
used in the derivation of the RfD.
The studies by Viberg et al. (2005, 2004b) of rats and mice, respectively, and the data
from the Branch! et al. (2002) study were not amenable to a BMD approach because the data
needed for dose-response modeling in BMDS (i.e., means and SDs) were not available from the
published studies. Neurobehavioral data were only displayed graphically, and the needed means
and SDs cannot be read with any accuracy from the graphs. Data from the study by Talsness et
al. (2005) also could not be used for BMD modeling because the histologic changes observed in
the ovaries and vaginal epithelium of rats were only qualitatively described. Neither incidence
nor severity was quantified. In addition, no effect on fertility was observed in this study, and the
NOAEL was at the highest dose tested.
5.1.3. RfD Derivation
Through use of BMD modeling, the estimated BMDL1SD of 0.29 mg/kg based on a
decrease in rearing habituation in 8-month-old female mice exposed to BDE-99 on PND 10
(Viberg et al., 2004a) was selected as the point of departure for the RfD. To calculate the RfD, a
total uncertainty factor (UF) of 3,000 was applied. This total UF was comprised of a factor of 10
to account for extrapolating animal data to humans (UFA or interspecies variability), a factor of
10 to account for susceptible human subpopulations (UFH or intrahuman variability), a factor of
3 to account for extrapolating from a single-dose exposure to a lifetime exposure (UFS), and a
factor of 10 to account for database deficiencies (UFD). The rationale for application of each of
these UFs is described below.
A 10-fold UFA was used to account for laboratory animal to human interspecies
differences. Although the toxicokinetics of BDE-99 in animals have been evaluated, no
adequate description of toxicokinetics of BDE-99 in humans exists. The critical effect for
deriving the RfD, altered behavior due to exposure during development, is expected to be
relevant to humans. No quantitative data were identified to compare relative human and rodent
sensitivity to these changes. However, given the longer period of brain development in humans
as compared to rodents and the higher importance of cognitive function, it is appropriate to
consider that humans may be more sensitive than rodents in the absence of specific data. Based
on these considerations the default UFA value of 10 was applied.
A default intraspecies UFH of 10 was applied to account for variations in susceptibility
within the human population (intrahuman variability). This factor accounts for the segment of
the human population that may be more sensitive than the general population to exposure to
63
-------
BDE-99. A default value is warranted because insuffiencient information is currently available
to assess human-to-human variability in BDE-99 toxicokinetics or toxicodynamics.
A UFS of 3 was used to account for extrapolating from effects seen in a single-exposure
neurodevelopmental study to a lifetime exposure. Exposure on PND 10 occurred during a period
of rapid brain development in mice. Brain development does not continue at an equivalent rate
over a mouse's lifetime and is more quiescent during adult life stages. Many brain structures
have a critical window during early life development. Following BDE-99 exposure,
toxicokinetic data suggest that a mouse urinary protein becomes functional some time between
PNDs 28 and 40, which leads to a dramatic increase in BDE-99 urinary excretion, especially in
males. This increased excretion reduces the total body burden of BDE-99 in older mice
compared with that in younger mice, including the levels of radiolabel reaching the brain
24 hours after dosing. These data suggest that the risk of neurodevelopmental effects in neonatal
mice may be greater than in older mice because of rapid postnatal brain growth and coincident
increased retention of BDE-99 and/or its metabolites. Therefore, chronic exposure is not
expected to result in more serious effects. However, because the mice received only a single
dose rather than repeated doses over multiple days within the hypothesized critical window, a
threefold UF was applied.
A UFLfor LOAEL-to-NOAEL extrapolation was not used because the Agency's current
approach is to address this factor as one of the considerations in selecting a BMR for BMD
modeling. In this case, a change in the mean equal to 1 SD of the control mean was assumed to
represent a minimal biologically significant change.
A UFD of 10 was used to account for database uncertainty. The available oral database
for BDE-99 lacks a full prenatal developmental neurotoxicity study, a multigeneration
reproductive toxicity study, and conventional studies of subchronic and chronic toxicity. In
addition, uncertainties regarding the effects of exposures during the prenatal period, extended
postnatal exposures, and latent expression of early postnatal changes in the brain are addressed
as a component of this database UF.
In conclusion, application of a total UF of 3,000 to the BMDL1SD of 0.29 mg/kg results in
an RfD for BDE-99 of 1 x 10^ mg/kg-day or 0.1 |ig/kg-day. For comparison, by using a
NOAEL/LOAEL approach to derive the RfD, a total UF of 3,000 is applied to the NOAEL of
0.4 mg/kg for neurodevelopmental effects identified in the Viberg et al. (2004a) study, yielding
an RfD for BDE-99 of 1.3 x 10~4 mg/kg-day or 0.1 |ig/kg-day, essentially equivalent to the RfD
derived by using the BMD approach.
5.1.4. Previous RfD Assessment
A previous IRIS assessment of commercial grade pentaBDE (CASRN 32534-81-9) is
available (U.S. EPA, 1990). The composition of this commercial pentaBDE product was 58.1%
64
-------
penta-, 24.6% tetra-, 13.3% hexa-, 2.6% hepta-, 0.8% deca-, 0.3% octa-, and 0.2% nonaBDE
(Carlson, 1980a). An RfD of 2 x icr3 mg/kg-day (2 ng/kg-day) was derived, based on a NOAEL
of 1.8 mg/kg-day and a LOAEL of 3.5 mg/kg-day for induction of hepatic enzymes in a 90-day
oral gavage study in rats (Carlson, 1980b), employing a UF of 1,000. This UF of 1,000 is
comprised of a factor of 10 each for interspecies (i.e., animal to human) and intrahuman
variability in lieu of specific data on this chemical and a factor of 10 to account for extrapolation
from a subchronic effect level to its chronic equivalent. At the time of this previous assessment,
insufficient information was available to derive an RfC or to assess the carcinogenicity of this
commercial grade pentaBDE.
5.2. INHALATION REFERENCE CONCENTRATION (RfC)
No data are available for deriving an RfC for BDE-99.
5.3. CANCER ASSESSMENT
Inadequate information is currently available to assess the carcinogenic potential of
BDE-99 (U.S. EPA, 2005a, b).
65
-------
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD
AND DOSE RESPONSE
6.1. HUMAN HAZARD POTENTIAL
BDE-99 (CASRN 60348-60-9) is a component of commercial PBDE flame retardants.
BDE-99 has been found in human milk, adipose tissue, and blood. As a result, fetuses and
infants are potentially exposed to BDE-99. Although this information does not elucidate the
effects of BDE-99 on the human general population, it does demonstrate that exposure can
occur. For example, the presence of BDE-99 in human breast milk allows transfer of BDE-99
(and/or its metabolites) from a mother's body to her infant through breast-feeding.
No data are currently available regarding the potential toxicity of BDE-99 in humans
exposed via the oral route. However, the available animal data indicate that the nervous system
is a sensitive target. Neurobehavioral developmental toxicity has been identified as the critical
endpoint of concern in mice following pre- and neonatal oral exposure to BDE-99. Specifically,
BDE-99 appears to disrupt spontaneous behavior and causes hyperactivity in mice and rats,
which appear to be permanent effects that worsen with age (Kuriyama et al., 2005; Viberg et al.,
2004a, b; Branch! et al., 2002; Eriksson et al., 2001). Since fetuses and infants are exposed to
BDE-99 via maternal/cord blood and human breast milk, such exposures may constitute a risk of
adverse neurodevelopmental effects in these population groups. In addition to effects on
spontaneous motor behavior, Kuriyama et al. (2005) reported impairment of spermatogenesis in
adult rat offspring but no consequent effect on sperm morphology, sperm quality, testosterone,
LH levels, or the ability to sire offspring. Histological changes in the ovaries and vaginal
epithelium were seen in rats exposed to BDE-99. However, these changes were not associated
with statistically significant effects on fertility indices (Talsness et al., 2005).
No studies currently exist on the potential carcinogenicity of BDE-99 in humans or
experimental animals. Under the Guidelines for Carcinogen Risk Assessment (U.S. EPA,
2005a), there is "inadequate information to assess carcinogenic potential" of BDE-99 at this
time.
6.2. DOSE RESPONSE
The RfD of 0.1 jig/kg-day was derived from a BMDL1SD of 0.29 mg/kg-day, based on the
effects of BDE-99 on spontaneous motor behavior in mice (Viberg et al., 2004a). A total UF of
3,000 was applied to this BMDL1SD to derive the RfD. This total UF of 3,000 was comprised of
a factor of 10 to account for interspecies variability, a factor of 10 to account for intrahuman
variability, a factor of 3 to account for extrapolation from a single-dose to a chronic exposure,
and a factor of 10 to account for database deficiencies.
66
-------
No data are currently available regarding the potential toxicity of BDE-99 in humans
exposed via the oral route, and no suitable toxicokinetic or toxicodynamic models have been
developed to reduce uncertainty in extrapolating from mice to humans.
The extent of variability in susceptibility to BDE-99 among humans is unknown,
representing another important area of uncertainty in the RfD. However, subpopulations
expected to be more susceptible to BDE-99 toxicity include fetuses and children. Chronic
studies relevant to BDE-99 toxicity in these subpopulations have not yet been performed in
experimental animals.
The principal study used in the derivation of the RfD (Viberg et al., 2004a) evaluated a
number of behavioral parameters in a limited number of male and female mice ages 2, 5, and
8 months, exposed to BDE-99 on PND 10 at five different oral dose levels. Supporting studies
for the neurobehavioral developmental effects of BDE-99 include a study in male mice exposed
on PND 10 to three doses of BDE-99 (Viberg et al., 2004b), a study in which neonatal male mice
were exposed orally to BDE-99 at two dose levels (Eriksson et al., 2001), a study that examined
a number of neurobehavioral parameters in male and female neonatal mice pre- and postnatally
exposed to three oral dose levels of BDE-99 (Branch! et al., 2002), a study on effects on
locomotor activity in rats exposed in utero to two dose levels of BDE-99 (Kuriyama et al., 2005),
and a study in rats exposed on PND 10 to three oral doses of BDE-99 (Viberg et al., 2005).
The toxicological database for BDE-99 is sparse. No standard reproductive,
developmental, subchronic, or chronic studies exist in rats or mice, nor is there a much needed
developmental neurotoxicity study. In addition, several concerns regarding the experimental
design of the Viberg et al. (2004a) study used in deriving the RfD have been raised (see section
5.1.1). Thus, the overall confidence in the RfD is low.
67
-------
7. REFERENCES
Akutsu, K; Kitagawa, M; Nakazawa, H; et al. (2003) Time-trend (1973-2000) of polybrominated diphenyl ethers in
Japanese mother's milk. Chemosphere 53(6):645-654.
Aim, H; Scholz, B; Fischer, C; et al. (2006) Proteomic evaluation of neonatal exposure to
2,2',4,4',5-pentabromodiphenyl ether. Environ Health Perspect 114(2):254-259.
Ankarberg, E. (2003) Neurotoxic effects of nicotine during neonatal brain development. Comprehensive summaries
of Uppsala Dissertations from the Faculty of Science and Technology 907. Acta Universitatis Upsaliensis, Uppsala,
Sweden.
ATSDR (Agency for Toxic Substances and Disease Registry). (2004) Toxicological profile for polybrominated
biphenyls and polybrominated diphenyl ethers. Public Health Service, U.S. Department of Health and Human
Services, Atlanta, GA. Available online at http://www.atsdr.cdc.gov/toxpro2.html.
Bock, KW. (1994) Arylhydrocarbon of dioxin receptor: biologic and toxic responses. Rev Physiol Biochem
Pharmacol 125:1-42.
Bradman A; Fenster, L; Sjodin, A; et al. (2007) Polybrominated diphenyl ether levels in the blood of pregnant women
living in an agriculture community in California. Environ Health Perspect 115(l):71-74.
Braekevelt, E; Tittlemier, SA; Tomy, GT. (2003) Direct measurement of octanol-water partition coefficients of some
environmentally relevant brominated diphenyl ether congeners. Chemosphere 51:563-567.
Branchi, I; Alleva, E; Costa, LG. (2002) Effects of perinatal exposure to a polybrominated diphenyl ether (PBDE 99)
on mouse neurobehavioural development. Neurotoxicol 23(3):375-384.
Branchi, I; Capone, F; Vitalone, A; et al. (2005) Early developmental exposure to BDE 99 or Aroclor 1254 affects
neurobehavioural pro file: interference from the administration route. Neurotoxicol 26(2):183-192.
Carlson, GP. (1980a) Induction of xenobiotic metabolism in rats by short-term administration of brominated diphenyl
ethers. Toxicol Lett 5:19-25.
Carlson, GP. (1980b) Induction of xenobiotic metabolism in rats by brominated diphenyl ethers administered for 90
days. Toxicol Lett 6:207-212.
Ceccatelli, R; Faass, O; Schlumpf, M; et al. (2006) Gene expression and estrogen sensitivity in rat uterus after
developmental exposure to the polybrominated diphenyl ether PBDE 99 and PCB. Toxicology 220:104-116.
Cetin, B; Odabasi, M. (2005) Measurement of Henry's law constants of seven polybrominated diphenyl ether (PBDE)
congeners as a function of temperature. Atmos Environ 39:5273-5280.
Chen, G; Bunce, NJ. (2003) Polybrominated diphenyl ethers as Ah receptor agonists and antagonists. Toxicol Sci
76:310-320.
Chen, G; Konstantinov, AD; Chittim, BG; et al. (2001) Synthesis of polybrominated diphenyl ethers and their
capacity to induce CYP1A by the Ah receptor mediated pathway. Environ Sci Technol 35:3749-3756.
Chen, JW; Harner, T; Yang, P; et al. (2003) Quantitative predictive models for octanol-air partition coefficients of
polybrominated diphenyl ethers at different temperatures. Chemosphere 51:577-584.
Chen, LJ; Lebetkin, EH; Sanders, JM; et al. (2006) Metabolism and disposition of 2,2',4,4',5-pentabromodiphenyl
ether (BDE99) following a single or repeated administration to rats or mice. Xenobiotica 36(6):515-534.
68
-------
Choi, JW; Fujimaki, TS; Kitamura, K; et al. (2003) Polybrominated dibenzo-p-dioxins, dibenzofurans, and diphenyl
ethers in Japanese human adipose tissue. Environ Sci Technol 37(5):817-821.
Darnerud, PO; Risberg, S. (2006) Tissue localisation of tetra- and pentabromodiphenyl ether congeners (BDE-47,
-85 and -99) in perinatal and adult C57BL mice. Chemosphere 62:485-493.
Darnerud, PO; Wong, J; Bergman, A; et al. (2005) Common viral infection affects pentabromodiphenyl ether (PBDE)
distribution and metabolic and hormonal activities in mice. Toxicology 210:159-167.
Eriksson, P; Jakobsson, E; Fredriksson, A. (2001) Brominated flame retardants: a novel class of developmental
neurotoxicants in our environment? Environ Health Perspect 109(9):903-908.
Eriksson, P; Viberg, H; Jakobsson, E; et al. (2002) A brominated flame retardant, 2,2',4,4',5-pentabromodiphenyl
ether: uptake, retention, and induction of neurobehavioural alterations in mice during a critical phase of neonatal brain
development. Toxicol Sci67(l):98-103.
Evandri, MG; Mastrangelo, S; Costa, LG; et al. (2003) In vitro assessment of mutagenicity and clastogenicity of
BDE-99, a pentabrominated diphenyl ether flame retardant. Environ Mol Mutagen 42:85-90.
Fernlof, G; Gadhasson, I; Podra, K; et al. (1997) Lack of effects of some individual polybrominated diphenyl ether
(PBDE) and polychlorinated biphenyl (PCB) congeners on human lymphocyte functions in vitro. Toxicol Lett 90(2-
3):189-197.
Fischer, D; Hooper, K; Athanasiadou, M; et al. (2006) Children show highest levels of polybrominated diphenyl
ethers in a California family of four: a case study. Environ Health Perspect 114(10): 1581-1584.
Flemming, A; Moller, LM; Madsen, T. (2000) Brominated flame retardants: toxicity and ecotoxicity. Danish
Environmental Protection Agency, Denmark; Environmental Project No. 568.
Great Lakes Chemical Corporation. (2003) Voluntary Children's Chemical Evaluation Program (VCCEP). Tier 1
assessment of the potential health risks to children associated with exposure to the commercial pentabromodiphenyl
ether product. Prepared by Environ International Corporation, Ruston, LA, for Great Lakes Chemical Corporation
(now Chemtura, Middlebury, CT); 03-10607A.
Guvenius, DM; Bergman, A; Noren, K. (2001) Polybrominated diphenyl ethers in Swedish human liver and adipose
tissue. Arch Environ Contam Toxicol 40:564-570.
Guvenius, DM; Aronsson, A; Ekman-Ordeberg, G; et al. (2003) Human prenatal and postnatal exposure to
polybrominated diphenyl ethers, polychlorinated biphenyls, polychlorobiphenylols, and pentachlorophenol. Environ
Health Perspect 111(9):1235-1241.
Hakk, H; Larsen, G; Klasson-Wehler, E. (2002a) Tissue disposition, excretion and metabolism of
2,2',4,4',5-pentabromodiphenyl ether (BDE-99) in the male Sprague-Dawley rat. Xenobiotica 32(5):369-382.
Hakk, H; Larsen, G; Bergman, A; et al. (2002b) Binding of brominated diphenyl ethers to male rat carrier proteins.
Xenobiotica 32(12): 1079-1091.
Hakk, H; Huwe, J; Low, M; et al. (2006) Tissue disposition, excretion and metabolism of
2,2',4,4',6-pentabromodiphenyl ether (BDE-100) in male Sprague-Dawley rats. Xenobiotica 36(l):79-94.
Hallgren, S; Darnerud, PO. (2002) Polybrominated diphenyl ethers (PBDEs), polychlorinated biphenyls (PCBs) and
chlorinated paraffins (CPs) in rats—testing interactions and mechanisms for thyroid hormone effects. Toxicology
177(2-3):227-243.
Hallgren, S; Sinjari, T; Hakansson, H; et al. (2001) Effects of polybrominated diphenyl ethers (PBDEs) and
polychlorinated biphenyls (PCBs) on thyroid hormone and vitamin A levels in rats and mice. Arch Toxicol
75(4):200-208.
69
-------
Johnson-Restrepo, B; Kannan, K; Rapaport, DP; et al. (2005) Polybrominated diphenyl ethers and polychlorinated
biphenyls in human adipose tissue from New York. Environ Sci Technol 39:5177-5182.
Kester, MH; Bulduk, S; van Toor, H; et al. (2002) Potent inhibition of estrogen sulfotransferase by hydroxylated
metabolites of polyhalogenated aromatic hydrocarbons reveals alternative mechanism for estrogenie activity of
endocrine disrupters. J ClinEndocrinol Metab 87(3): 1142-1150.
Klaassen, CD; ed. (1996) Casarett and Doull's toxicology: the basic science of poisons. 5th edition. New York, NY:
McGraw-Hill; pp. 47-19, 373-376.
Kodavanti, PR; Ward, TR; Ludewig, G; et al. (2005) Polybrominated diphenyl ether (PBDE) effects in rat neuronal
cultures: 14C-PBDE accumulation, biological effects, and structure-activity relationships. Toxicol Sci 88(1): 181-192.
Kuriyama, SN; Talsness, CE; Grote, K; et al. (2005) Developmental exposure to low dose PBDE 99: effects on male
fertility and neurobehavior in rat offspring. Environ Health Perspect 113:149-154.
Lewis, DF; Watson, E; Lake, BG. (1998) Evolution of the cytochrome P450 superfamily: sequence alignments and
pharmacogenetics. Mutat Res 410(3):225-270.
Lilienthal, H; Roth-Harer, A; Hack, A; et al. (2005) Developmental neurotoxicity of PHAHs: endocrine-mediated and
general behavioral endpoints in adult male rats. Environ Tox Pharmacol 19:757-759.
Lilienthal, H; Hack, A; Roth-Harer, A; et al. (2006) Effects of developmental exposure to
2,2',4,4',5-pentabromodiphenyl ether (PBDE-99) on sex steroids, sexual development, and sexually dimorphic
behavior in rats. Environ Health Perspect 114(2): 194-201.
Lind, Y; Darnerud, PO; Atuma, S; et al. (2003) Polybrominated diphenyl ethers in breast milk from Uppsala County,
Sweden. Environ Res 93:186-194.
Madia, F; Giordano, G; Fattori, V; et al. (2004) Differential in vitro neurotoxicity of the flame retardant PBDE-99
and of the PCB Aroclor 1254 in human astrocytoma cells. Toxicol Lett 154:11-21.
Mazdai, A; Dodder, NG; Abernathy, MP; et al. (2003) Polybrominated diphenyl ethers in maternal and fetal blood
samples. Environ Health Perspect 111(9): 1249-1252.
Meerts, IA; van Zanden, JJ; Luijks, EA; et al. (2000) Potent competitive interactions of some brominated flame
retardants and related compounds with human transthyretin in vitro. Toxicol Sci 56:95-104.
Meerts, IA; Letcher, PJ; Hoving, S; et al. (2001) In vitro estrogenicity of polybrominated diphenyl ethers,
hydro xylated PDBEs, and polybrominated bisphenol A compounds. Environ Health Perspect 109(4):399-407.
Meironyte, D; Noren, K; Bergman, A. (1999) Analysis of polybrominated diphenyl ethers in Swedish human milk. A
time-related trend study, 1972-1997. J Toxicol Environ Health (Part A) 58(6):329-341.
Meneses, M; Wingfors, H; Schuhmacher, M; et al. (1999) Polybrominated diphenyl detected in human adipose tissue
from Spain. Chemosphere 3(13):2271-2278.
Morreale de Escobar, G; Obregon, MJ; Escobar del Rey, F. (2000) Is neuropsychological development related to
maternal hypothyroidism or to maternal hypothyroxinemia? J Clin Endocrinol Metab 85:3975-3987.
NRC (National Research Council). (1983) Risk assessment in the federal government: managing the process.
Washington, DC: National Academy Press.
Ohta, S; Ishizuka, D; Nishimura, H; et al. (2002) Comparison of polybrominated diphenyl ethers in fish, vegetables,
and meat and levels in human milk of nursing women in Japan. Chemosphere 46:689-696.
70
-------
Pacyniak, EK; Cheng, X; Cunningham, MK; et al. (2007) The flame retardants, polybrominated diphenyl ethers, are
pregnane X receptor activators. Toxicol Sci 97(1):94-102.
Palm, A; Cousins, IT; Mackay, D; et al. (2002) Assessing the environmental fate of chemicals of emerging concern: a
case study of the polybrominated diphenyl ethers. Environ Pollut 117:195-213.
Peters, AK; Sanderson, JT; Bergman, A; et al. (2006) Antagonism of TCDD-induced ethoxyresorufin-0-deethylation
activity by polybrominated diphenyl ethers (PBDEs) in primary cynomolgus monkey (Macaco fascicularis)
hepatocytes. Toxicol Letters 164:123-132.
Rice, DC; Reeve, EA; Herlihy, A; et al. (2007) Developmental delays and locomotor activity in the C57BL6/J mouse
following neonatal exposure to the fully-brominated PBDE, decabromodiphenyl ether. Neurotoxicol Teratol 29:511-
520.
Sand, S; von Rosen, D; Eriksson, P; et al. (2004) Dose-response modeling and benchmark calculations from
spontaneous behavior data on mice neonatally exposed to 2,2',4,4',5 -pentabromodiphenyl ether. Toxicol Sci 81:491-
501.
Sanders, JM; Burka, LT; Smith, CS; et al. (2005) Differential expression of CYP1 A, 2B, and 3 A genes in the F344
rat following exposure to a polybrominated diphenyl ether mixture or individual components. Toxicol Sci 88(1): 127-
133.
Schecter, A; Pavuk, M; Papke, O; et al. (2003) Polybrominated diphenyl ethers (PBDEs) in U.S. mothers' milk.
Environ Health Perspect 111(14): 1723-1729.
She, J; Petreas, M; Winkler, J; et al. (2002) PBDEs in the San Francisco Bay area: measurement in harbor seal
blubber and human breast adipose tissue. Chemosphere 46:697-707.
She, J; Holden, A; Sharp, M; et al. (2007) Polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls
(PCBs) in breast milk from the Pacific Northwest. Chemosphere 67:8307-8317.
Sjodin, A; Hagmar, L; Klasson-Wehler, E; et al. (1999) Flame retardant exposure: polybrominated diphenyl ethers in
blood from Swedish workers. Environ Health Perspect 107(8):643-648.
Sjodin, A; Patterson, DG, Jr; Bergman, A. (2001) Brominated flame retardants in serum from U.S. blood donors.
Environ Sci Technol 35(19):3830-3833.
Sjodin, A; Jones, RS; Focant, JF; et al. (2004) Retrospective time-trend study of polybrominated diphenyl ether and
polybrominated and polychlorinated biphenyl levels in human serum from the United States. Environ Health Perspect
112(6):654-658.
Skarman, E; Darnerud, PO; Ohrvik, H; et al. (2005) Reduced thyroxine levels in mice perinatally exposed to
polybrominated diphenyl ethers. Environ Tox Pharmacol 19:273-281.
Staskal, DF; Hakk, H; Bauer, D; et al. (2006) Toxicokinetics of polybrominated diphenyl ether congeners 47, 99,
100, and 153 in mice. Toxicol Sci 94(l):28-37.
Stenzel, JI; Markley, BJ. (1997) Pentabromodiphenyl oxide (PeBDPO): determination of the water solubility.
Prepared by Wildlife International Ltd., Easton, MD, for the Chemical Manufacturers Association's Brominated
Flame Retardant Industry Panel, Arlington, VA ; Project Number: 439C-109. Unpublished study.
Stoker, TE; Laws, SC; Crofton, KM; et al. (2004) Assessment of DE-71, a commercial polybrominated diphenyl
ether (PBDE) mixture, in the EDSP male and female pubertal protocols. Toxicol Sci 78:144-155.
Stoker, TE; Cooper, RL; Lambright, CS; et al. (2005) In vivo and in vitro anti-androgenic effects of DE-71, a
commercial polybrominated diphenyl ether (PBDE) mixture. Toxicol Appl Pharmacol 207:78-88.
71
-------
Talsness, CF; Shakibaei, M; Kuriyama, S; et al. (2005) Ultrastructural changes observed in rat ovaries following in
utero and lactational exposure to low doses of a polybrominated flame retardant. ToxicolLett 157:189-202.
Thomsen, C; Lundanes, E; Becher, G. (2002) Brominated flame retardants in archived serum samples from Norway:
a study on temporal trends and the role of age. Environ Sci Technol 36(7): 1414-1418.
U.S. EPA (Environmental Protection Agency). (1986a) Guidelines for the health risk assessment of chemical
mixtures. Federal Register 51(185):34014-34025. Available from: .
U.S. EPA. (1986b) Guidelines for mutagenicity risk assessment. Federal Register 51(185):34006-34012. Available
from: .
U.S. EPA. (1988) Recommendations for and documentation of biological values for use in risk assessment.
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment, Cincinnati, OH;
EPA/600/6-87/008. Available from the National Technical Information Service, Springfield, VA; PB88-179874/AS,
and online at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=34855.
U.S. EPA. (1990) Pentabromodiphenyl ether (CASRN 32534-81-9). Integrated Risk Information System (IRIS).
National Center for Environmental Assessment, Washington, DC. Available online at http://www.epa.gov/iris.
U.S. EPA. (1991) Guidelines for developmental toxicity risk assessment. Federal Register 56:63798-63826.
Available online at http://www.epa.gov/ncea/raf/rafguid.htm
U.S. EPA. (1994a) Interim policy for particle size and limit concentration issues in inhalation toxicity studies. Federal
Register 59(206):53799. Available from: .
U.S. EPA. (1994b) Methods for derivation of inhalation reference concentrations and application of inhalation
dosimetry. Office of Research and Development, Washington, DC; EPA/600/8-90/066F. Available from:
.
U.S. EPA. (1995) Use of the benchmark dose approach in health risk assessment. Risk Assessment Forum,
Washington, DC; EPA/630/R-94/007. Available from the National Technical Information Service, Springfield, VA,
PB95-213765, and online at http://cfpub.epa.gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive.
U.S. EPA. (1996) Guidelines for reproductive toxicity risk assessment. Federal Register 61:56274-56322. Available
online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA. (1998a) Guidelines for neurotoxicity risk assessment. Federal Register 63:26926-26954. Available online
at http://www.epa.gov/ncea/raf/rafguid.htm
U.S. EPA. (1998b) Health effects test guidelines: neurotoxicity screening battery. Office of Prevention, Pesticides
and Toxic Substances, Washington, DC; OPPTS 870.6200; EPA 712-C-98-238. Available online at
http://www.epa.gov/opptsfrs/publications/OPPTS_Harmonized/870_Health_Effects_Test_Guidelines/Series/870-
6200.pdf.
U.S. EPA. (2000a) Science policy council handbook: risk characterization. Office of Science Policy, Office of
Research and Development, Washington, DC. EPA/100-B-00-002. Available online at
http://www.epa.gov/OSA/spc/pdfs/prhandbk.pdf.
U.S. EPA. (2000b) Benchmark dose technical guidance document [external review draft]. Risk Assessment Forum,
Washington, DC; EPA/630/R-00/001. Available online at http://cfpub.epa.gov/ncea/cfm/
nceapubUcation.cM?ActType=PubUcationTopics&detype=DOCUMENT&subject=BENCHMARK+DOSE&subjtype
=TITLE&excCol= Archive.
U.S. EPA. (2000c) Supplementary guidance for conducting for health risk assessment of chemical mixtures. Risk
Assessment Forum, Washington, DC; EPA/630/R-00/002. Available from: .
72
-------
U.S. EPA. (2002) A review of the reference dose and reference concentration processes. Risk Assessment Forum,
Washington, DC; EPA/630/P-02/002F. Available online at
http://cfpub.epa.gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive.
U.S. EPA. (2004) Pentabromodiphenyl ether. Substance Registry System U.S. Environmental Protection Agency,
Washington, DC. Available online at http://www.epa.gov/srs.
U.S. EPA. (2005a) Guidelines for carcinogen risk assessment. Federal Register 70: 17765-18717. Available online at
http://www.epa.gov/cancerguidelines.
U.S. EPA. (2005b) Supplemental guidance for assessing susceptibility from early -life exposure to carcinogens. Risk
Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available online at http://www.epa.gov/cancerguidelines.
U.S. EPA. (2005c) Voluntary Children's Chemical Evaluation Program: data needs decision document of
pentabromodiphenyl ether. Office of Pollution Prevention and Toxics, Washington DC. Available online at
http ://www. epa. gov/chemrtk/vccep/pubs/chem22. htm.
U.S. EPA. (2006a) Science policy council handbook: peer review. 3rd edition. Office of Science Policy, Office of
Research and Development, Washington, DC; EPA/100/B-06/002. Available online at
http ://www. epa. gov/OS A/spc/2peerrev. htm.
U.S. EPA. (2006b) A framework for assessing health risk of environmental exposures to children. National Center
for Environmental Assessment, Washington, DC, EPA/600/R-05/093F. Available online at
ij_gy/ncci^
Viberg, H; Fredriksson, A; Eriksson, P. (2002) Neonatal exposure to the brominated flame retardant 2,2',4,4',5-
pentabromodiphenyl ether causes altered susceptibility in the cholinergic transmitter system in the adult mouse.
Toxicol Sci 67(1): 104-107.
Viberg, H; Frederiksson, A; Eriksson, P. (2003a) Neonatal exposure to polybrominated diphenyl ether (PBDE 153)
disrupts spontaneous behaviour, impairs learning and memory, and decreases hippocampal cholinergic receptors in
adult mice. Toxicol Appl Pharmacol 192(2):95-106.
Viberg, H; Fredriksson, A; Jakobsson, E; et al. (2003b) Neurobehavioral derangements in adult mice receiving
decabromodiphenyl ether (PBDE 209) during a defined period of neonatal brain development. Toxicol Sci 76: 1 12-
120.
Viberg, H; Fredriksson, A; Eriksson, P. (2004a) Investigations of strain and/or gender differences in developmental
neurotoxic effects of polybrominated diphenyl ethers in mice. Toxicol Sci 81:344-353.
Viberg, H; Fredriksson, A; Jakobsson, E; et al. (2004b) Neonatal exposure to the brominated flame -retardant,
2,2',4,4',5-pentabromodiphenyl ether, decreases cholinergic nicotinic receptors in hippocampus and affects
spontaneous behaviour in the adult mouse. Environ Toxicol Pharmacol 17:61-65.
Viberg, H; Fredriksson, A; Eriksson, P. (2005) Deranged spontaneous behavior and decrease in cholinergic
muscarinic receptors in hippocampus in the adult rat, after neonatal exposure to the brominated flame -retardant,
2,2',4,4',5-pentabromodiphenyl ether (PBDE 99). Environ Toxicol Pharmacol 20:283-288.
Viberg, H; Fredriksson, A; Eriksson, P. (2007) Changes in spontaneous behaviour and altered response to nicotine in
the adult rat, after neonatal exposure to the brominated flame retardant, decabrominated diphenyl ether (PBDE 209).
Neurotoxicology 28(1): 136-142.
Villeneuve, DL; Kannan, K; Priest, BT; et al. (2002) In vitro assessment of potential mechanism-specific effects of
polybrominated diphenyl ethers. Environ Toxicol Chem 21(1 1):243 1-2433.
Wong, A; Lei, YD; Alaee, M; et al. (2001) Vapor pressures of the polybrominated diphenyl ethers. J ChemEng Data
46:239-242.
73
-------
Zhou, T; Taylor, MM; DeVito, MJ; et al. (2002) Developmental exposure to brominated diphenyl ethers results in
thyroid hormone disruption. Toxicol Sci 66(1): 105-116.
74
-------
APPENDIX A: SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION
The "Toxicological Review" for pentabromodiphenyl ether (BDE-99) has undergone a
formal external peer review performed by scientists in accordance with EPA guidance on peer
review (U.S. EPA, 2006a, 2000a). The external peer reviewers were tasked with providing
written answers to general questions on the overall assessment and on chemical-specific
questions in areas of scientific controversy or uncertainty. The external peer review for BDE-99
was conducted in concert with the external peer review of other PBDE congeners (i.e., BDE-47,
-153, and -209), and some external peer review charge questions were specific to congeners
other than BDE-99. External peer reviewer comments on all of the PBDEs and the Agency
response are included below for completeness. A summary of significant comments made by the
external reviewers and EPA's responses to these comments follow. In many cases the comments
of the individual reviewers have been synthesized and paraphrased in development of Appendix
A. Synthesis of comments from individual peer reviewers resulted in summaries that combine
similar statements from peer reviewers that were mentioned in conjunction with more than one
charge question. In such cases, the comment and its response have been placed under the most
relevant charge question. Some of the peer review comments were not directly related to charge
questions. Those comments are categorized as miscellaneous and placed after those related to
the charge questions. EPA also received scientific comments from the public. These comments
and EPA's responses are included in a separate section of this appendix.
The peer review of the "Toxicological Review" of BDE-99 was coupled with the review
of the documents for BDE- 47, -153, and -209. Accordingly, most of the charge questions
address all four congeners. The responses to the charge questions in this appendix apply
primarily to comments related to BDE-99. The charge to the external peer reviewers and final
external peer review report (February 2007) pertaining to the toxicological reviews of the four
PBDE congeners are available at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=161970.
The public comments received can be found at
http;//wwwj;egulM^
0838.
EXTERNAL PEER REVIEWER COMMENTS
The reviewers made several editorial suggestions to clarify specific portions of the text. These
changes were incorporated in the document as appropriate and are not discussed further.
A. General Comments
A-l
-------
Charge Question 1. Are you aware of other published peer-reviewed toxicological studies not
included in these toxicological reviews that could be of relevance to the health assessment of
BDE-47, -99, -153, or -209?
Comment 1: Three reviewers stated that they were unaware of any other relevant studies that
would contribute to the BDE-99 IRIS assessment. One reviewer identified potentially relevant
additional literature:
Jones-Ortazo, HA; etal. (2005) Environ. Sci. Technol. 39:5121-5130
Wilford, BH; et al. (2005) Environ. Sci. Technol. 39:7027-7035
Schecter, A; et al. (2005) J. Toxicol. Environ. Health Part A 68:501-513
Kites, RA; et al. (2004) Environ. Sci. Technol. 38:4945-4949
Schecter, A; et al. (2006) Environ. Health Perspect. 114:1515-1520
Fischer, D; et al. (2006) Environ. Health Perspect. 114:1581-1584
Bradman, A; et al. (2007) Environ. Health Perspect. 115:71-74
Kodavanti, PRS; Derr-Yellin, EC. (2002) Toxicol. Sci. 68:451-457
Kodavanti, PRS; et al. (2005) Toxicol. Sci. 88:181-192
Reistad, T; Mariussen, E. (2005) Toxicol. Sci. 87:57-65
Reistad, T; et al. (2006) Arch. Toxicol. 80:785-796
Response: The Agency reviewed and evaluated the studies recommended by the reviewers and
has included the relevant studies for BDE-99. Fischer et al. (2006), Bradman et al. (2007), and
Kodavanti et al. (2005) were found to be relevant and were added to the document. The
remaining studies suggested by the reviewer fell outside the scope of the IRIS assessment (i.e.,
exposure data, commercial mixtures). Additionally, a new literature search was conducted to
ensure recently published, relevant studies are included in the IRIS assessment. A description of
these studies (Pacyniak et al., 2007; She et al., 2007) has been added to the "Toxicological
Review" in sections 3 and 4.
B. Oral Reference Dose (RfD) Values
Charge Question 2. Have the rationale and justification for deriving R/Ds on the basis of the
neurobehavioral toxicity studies been transparently and objectively described in the draft
toxicological reviews of BDE-47, -99, -153, and-209? Are there additional studies that should
be considered for deriving the RfDsfor any of the four PBDE congeners?
A-2
-------
Comment 1: Three reviewers stated that the rationale for deriving the RfD based on the
neurobehavioral toxicity studies was clearly and transparently described.
Response: No response needed.
Comment 2: Two reviewers stated that the neurobehavioral effects are the only toxic effects that
have been observed consistently in PBDE rodent studies. One of these reviewers stated that the
neurobehavioral studies appeared to provide the most appropriate dose-response data on which
to base the health assessment. Three reviewers felt that some consideration should be given to
other studies that provide data suitable for deriving an RfD, specifically the studies of Kuriyama
et al. (2005) and Branch! et al. (2002).
Response: The Agency reviewed and evaluated several studies of neurobehavioral endpoints in
rats (Kuriyama et al., 2005) and mice (Branch! et al., 2002) and reproductive endpoints in rats
(Kuriyama et al., 2005) in the IRIS health assessment for BDE-99. BMD analyses were
conducted using all relevant endpoints in the Viberg et al. (2004a) and Kuriyama et al. (2005)
studies. See the Agency's response to Comment 1 under Charge Question 4 for further rationale
for the selection of the study and endpoint on which to base the derivation of the RfD.
Charge Question 3. Do you agree or disagree with EPA basing the health assessment ofBDE-
47, -99, -153, and-209 to a large extent on the Eriksson/Viberg neurobehavioral studies?
Comment 1: One reviewer supported the use of the Eriksson/Viberg neurobehavioral studies as
the basis for the derivation of the RfDs, given the limited body of toxicological information
available. Two reviewers noted that the studies are limited by the fact that they originated from
the same laboratory. One reviewer was concerned that the experimental design of the principal
study selected more than one pup per litter, ignoring the "litter" effect. Treating littermates as
independent experimental units could confuse dose effects with litter effects. Another reviewer
was concerned with the specificity of the neurobehavioral data for developmental neurotoxicity
and suggested that independent confirmation of the endpoints is essential. One peer reviewer
identified the use of a single sex (male mice) as a limitation of the critical study that had not
been identified in the "Toxicological Review" discussion of study limitations. One of these
reviewers stated that these limitations do not hinder the derivation of the RfD for BDE-99 but
make the confidence low. Another reviewer noted that the neurobehavioral findings of this
laboratory have been corroborated in a study examining BDE-99 (Kuriyama et al., 2005). None
of the reviewers stated that the studies could not be used as the basis for the derivation of the
RfD.
A-3
-------
Response: The "lexicological Review" contains a detailed summary of the concerns with the
study design and methods utilized in the principal study (see section 5.1.1). A discussion of the
use of only male mice in the study by Viberg et al. (2004a) has been added to the discussion in
section 5.1.1 of the "lexicological Review." Additionally, the neurobehavioral effects reported
in Viberg et al. (2004a) are supported by an expanding body of literature (Rice et al., 2007;
Viberg et al., 2007, 2005, 2004b, 2003a, b, 2002; Kuriyama et al., 2005; Eriksson et al., 2002,
2001; Branch! et al., 2002) that details changes in motor and cognitive activity in rodents
following administration of single or repeated perinatal doses of PBDEs. Some of the concerns
associated with the methodology of the Eriksson/Viberg neurobehavioral studies are alleviated
by other studies of BDE-99 ( Kuriyama et al., 2005; Branch! et al., 2002) using more traditional
methodologies that have generated toxic effects similar to those reported by Viberg et al.
(2004a). Altogether, these studies support the findings of Viberg et al. (2004a) that exposure to
these PBDE congeners in early developmental stages can result in lasting changes in the
neurobehavioral activity of mice.
Charge Question 4. Are the Eriksson et al. (2001) (BDE-47), Viberg et al. (2004a) (BDE-99),
Viberg et al (2003a) (BDE-153), and Viberg et al (2003b) (BDE-2 09) studies appropriate for
determining the point of departure? Have the strengths and weaknesses of the Viberg and
Eriksson studies been appropriately characterized and considered?
Comment 1: All four reviewers believed that the Viberg et al. (2004a) study was appropriate for
determining the point of departure. One reviewer felt that the data were appropriate as long as
the document emphasizes that the neurochemical data also show alterations in normal
developmental patterns. Another reviewer noted that several candidate studies were modeled for
determining the point of departure, and the Viberg et al. (2004a) had the best dose-response data
and provided the best model fit. All four reviewers also stated that data in Kuriyama et al.
(2005) should be considered for determining the point of departure. Iwo of these reviewers
noted that the lowest effective dose in the Kuriyama et al. (2005) study was lower than those
reported in the Viberg and Eriksson studies. Another of the four reviewers stated that Kuriyama
et al. (2005) provided good dose-response relationships for neurobehavioral and male
reproductive effects. One of the reviewers stated that Branch! et al. (2002) should also be
considered as the principal study and that the lack of a dose-response relationship at the highest
dose was not a compelling reason to avoid modeling the data at the lower doses. One reviewer
stated that the histologic changes that occurred in the ovaries and vaginal epithelium in lalsness
et al. (2005) at very low doses of 0.06 mg/kg were not described with sufficient clarity to
A-4
-------
determine whether they could be used because neither the incidence nor the severity are
described.
Response: The Agency has reviewed and evaluated several neurobehavioral and reproductive
endpoints in rats from the Kuriyama et al. (2005) developmental study in the IRIS health
assessment for BDE-99. It is difficult to interpret the neurobehavioral effects reported by
Kuriyama et al. (2005) at PND 71 at the low dose of 0.06 mg/kg-day. While these effects are
similar to those observed in Viberg et al. (2004a), the LOAEL of 0.06 mg/kg-day is much lower
than the NOAEL of 0.4 mg/kg-day in Viberg et al. (2004a), and the point of departure, combined
with a UF of 3,000, produces an RfD that is the second lowest in the IRIS database. When BMD
analyses were conducted for the data collected on PND 36, where 0.06 mg/kg-day is a NOAEL
and 0.3 mg/kg-day is a LOAEL, the BMDs and BMDLs from the two studies are similar and
indicate that rats and mice may be equally susceptible to the neurobehavioral effects of BDE-99
and that no significant difference is apparent when the animals are exposed, in utero or
perinatally, to BDE-99. Because of the difficulty of interpreting the effects of the Kuriyama et
al. (2005) study at the low dose, the Agency decided to retain Viberg et al. (2004a) as the
principal study until such time as further research elucidates the effects of BDE-99 at very low
doses. Viberg et al. (2004a) was selected for use in the derivation of the RfD because a clear
NOAEL was identified, and the mouse study provided six data points compared with the four
data points provided in Kuriyama et al. (2005) collected at PND 36. Additionally, an adequate
fit to the data was obtained with BMD modeling and the critical effects are supported by several
other studies in mice. For the reproductive effects in Kuriyama et al. (2005), the only data set
that did not demonstrate significant lack of fit to the available models in the Agency's BMD
modeling software was the percent of adult rats with less than two ejaculations. These results
were not used to derive the RfD in the BDE-99 assessment, based on the lack of supporting data
and the uncertainty surrounding the biological significance of this effect. The effects of BDE-99
on the female reproductive system were evaluated in rats (Talsness et al., 2005). Histologic
changes in the ovaries and vaginal epithelium were seen at a dose as low as 0.06 mg/kg BDE-99
but were not associated with statistically significant effects on fertility (pregnancy rate, mean
implantation sites per dam, live fetuses per dam, and resorption rate). Data from the study by
Talsness et al. (2005) also could not be used for BMD modeling because the histologic changes
observed in the ovaries and vaginal epithelium of rats were only qualitatively described. Neither
incidence nor severity was quantified. In addition, no effect on fertility was observed in this
study, and the NOAEL was at the highest dose tested. The modeling output for Kuriyama et al.
(2005) and Viberg et al. (2004a) and a brief discussion of the results can be found in Appendix
B.
A-5
-------
The Branch! et al. (2002) study was not selected as the basis for the RfD based on the
lack of a clear dose-response relationship in the behavioral and activity parameters. While
changes in these parameters were observed, the changes were frequently noted in the low- and
mid-dose groups with no change in high-dose-treated animals or observed in only one of the
three dose groups. These responses do not allow for the identification of a NOAEL and/or
LOAEL. Additionally, the magnitude of variation in responses among the low-, mid-, and high-
dose animals cannot be determined with any precision because all motor activity data are
presented in graphic forms and, thus, are not amenable to modeling (even if the high dose was
dropped). The rationale for not selecting Branch! et al. (2002) as the basis for the RfD is
described in further detail in section 4.3.1.7
Charge Question 5. Have the most appropriate critical effect and point of departure been
selected? And has the rationale for the point of departure been transparently and objectively
described?
Comment 1: All four reviewers agreed with the selection of the neurobehavioral effects as
critical effects and that these effects were appropriate for identifying a point of departure. One
of the reviewers felt that the neurochemical data also provided critical information and should be
presented centrally rather than as supporting data. One reviewer stated that there was no
correlation between PND of exposure and the concentration of the chemical in the brain. One
reviewer added that decreased habituation might be as appropriate as or more appropriate than
the habituation ratio as an indicator of toxicity, while another believed that the actual behavioral
data, rather than the habituation ratio should have been presented in the document. It was not
clear to another reviewer why the actual data could not be recovered from the study authors to
allow for dose-response modeling and BMD estimation, given that the studies were published
fairly recently (2003). This reviewer recommended that the Agency attempt to recover the
neurobehavioral toxicity data from the study authors.
Response: Descriptions of the neurochemical data are fully summarized in section 4.4.2.5 and
are referenced in the principal study summary (section 4.3.1.4). The evidence of neurochemical
interactions and the potential relationship with the neurobehavioral effects are highlighted in the
mode-of-action section of the document (section 4.5.3.). The document presents the hypothesis
proposed by the Eriksson/Viberg group in which the observed effects on locomotion and
habituation are related to impaired development of the cholinergic system during the postnatal
brain growth spurt; however, data are unavailable to adequately determine the complete
relevance of the neurochemical effects or to establish a mode of action. Data by Eriksson et al.
(2002) demonstrated the presence of radiolabel in the brain after administration on PND 3, 10,
A-6
-------
or 19, and the data by Aim et al. (2006) showed that there were neurochemical differences in
animals treated with BDE-99 on PND 10 and controls. There are no measurements of levels of
any of the congeners in the brain or other tissues at the time of neurobehavioral testing at 2 or
4 months to show if any differences exist in the brain or other tissues at those time points.
The actual motor activity components (locomotion, rearing, and total activity) that gave
rise to the habituation ratios were reported in graphical form only and could not be reasonably
estimated as presented, and the Agency's attempts to obtain the raw data were unsuccessful.
Thus, the habituation ratios, rather than the actual habituation data, served as the basis for
determining the point of departure.
Comment 2: Two reviewers thought that the rationale for the point of departure had been
transparently and objectively described. Another reviewer felt the document provided clear
rationalization for the selection of the point of departure. None of the reviewers stated that the
rationale for the point of departure was not appropriately described.
Response: No response needed.
Charge Question 6. Have the rationale and justification for each UF selected in the draft
toxicological reviews of BDE-47, -99, -153, and-209 been transparently described? If the
selected UFs are not appropriate, what alternative UFs would you suggest and what are the
scientific rationales for those suggested? Does the database support the determinations of the
R/Ds for BDE-47, -99, -153, and -209?
Note: The peer reviewers provided fairly extensive comments about the individual components
of the combined UF. For that reason the following reviewer comments and EPA responses have
been grouped by the area of uncertainty to which they apply.
Comment 1: Two reviewers agreed that the document described the rationale and justification
for each UF and another reviewer noted that the selection of the UFs was described in detail.
The fourth reviewer felt that the document did not provide much explanation or justification for
applying the default interspecies and intraspecies UFs.
Response: There is little information available on the effects of BDE-99 in humans, and in the
absence of data there is no scientific rationale for moving away from the default values for the
interspecies and intraspecies UFs. Additional explanation for applying default interspecies and
intraspecies UFs was added to section 5.1.3.
A-7
-------
Comment 2: One reviewer suggested decreasing the interspecies UFA considering the relatively
specific and sensitive nature of the neurobehavioral and neurochemical measures compared with
conventional endpoints. However, another reviewer felt that the 10-fold UFA was justifiable
based on the lack of data on the mode of action in animals and humans.
Response: The 10-fold UFA for interspecies uncertainty is retained based on the lack of mode-
of-action, pharmacokinetic, and human data that would sufficiently illustrate the effects of BDE-
99 in animals and humans. Additional explanation for applying the default interspecies UFA was
added to section 5.1.3.
Comment 3: Two reviewers suggested lowering the intrahuman UFH. One of these reviewers
felt the relatively specific and sensitive nature of the neurobehavioral and neurochemical
measures compared with conventional endpoints warranted a decrease in the UFH. The other
reviewer recommended decreasing the 10-fold UFH to threefold based on the sensitivity of the
test species population (neonates).
Response: The 10-fold UFH for intraspecies uncertainty is retained based on the lack of
information concerning the pharmacokinetics and mode of action of BDE-99 in humans. In the
absence of human data, the effects in potentially susceptible populations exposed to BDE-99
cannot be determined. Additional explanation for applying the default intraspecies UFH was
added to section 5.1.3.
Comment 4: One reviewer disagreed with the treatment of a single-dose experiment as
equivalent to a subchronic exposure when applying a UF to account for differences in exposure
duration. This reviewer stated that the principal study needs to be treated as a single-dose study
and not a subchronic study. The reviewer also felt that the threefold UFS was inappropriate and
suggested raising the UFS from 3 to 10, to consider the extent to which the mother's pre-
pregnancy accumulated body burden would influence the developmental outcome, especially
since these data are unavailable. One reviewer felt that the accumulation of the chemical should
be considered in the calculation of the RfD. A third reviewer agreed with the application of a
threefold UFS, recognizing that for the observed neurobehavioral effects the timing of exposure
is more critical than the duration of exposure. This reviewer regards the UFS as accounting for
uncertainty from lack of prenatal exposure rather than uncertainty regarding potential effects of
chronic exposure. Another reviewer suggested the threefold UFS may not be necessary,
considering that exposure during a window of susceptibility indicates that chronic exposure may
not necessarily result in greater adverse effects.
A-8
-------
Response: For BDE-99, the principal study identified endpoints that, for the most part, reflect
specific aspects of developmental physiology. The hypothesized window of susceptibility,
proposed by the Eriksson/Viberg research group study authors, is based on the observation that
the developmental neurotoxic effects observed following postnatal exposure to BDE-99, with
peak vulnerability from PNDs on PND 10-14, will not occur once the toxicokinetics of intestinal
uptake and excretion have matured and the animal brain is developmentally less active (outside
the window of susceptibility). The Eriksson/Viberg et al. (2005, 2004b, 2003a) group has
suggested that the period of maximum vulnerability for the developing cholinergic system that
coincides with the most pronounced neurodevelopmental effects from BDE-99 exposure is from
PNDs 10-14. The UFS was viewed as a dosing duration adjustment rather than simply a
comparison of the effects of a subchronic to a chronic exposure, data that are lacking for
BDE-99. A threefold UFS was applied because the critical study dosed the animals only once
within the hypothesized critical window, not because the chronic exposures would have
exacerbated the impact on habituation.
In response to the comment about possible effects as the result of a maternal pre-
pregnancy body burden, the Agency notes that, although the principal study did not include
prenatal exposure, the maternal uptake and retention of BDE-99 during the prenatal period
would be lower than that of the pups during the postnatal period of vulnerability because of the
differences in toxicokinetics for mature versus neonatal animals. Support for the UFS is
provided by the Branch! et al. (2005, 2002) studies, where the effects levels were comparable to
those in the Viberg et al. (2004a) study and exposures extended from GD 6 to PND 21.
Comment 5: One reviewer recommended applying a threefold UF to the BMDL point of
departure to account for uncertainty in extrapolating from a dose of non-negligible toxicity
(BMDL10) to a dose of negligible toxicity.
Response: A UFLfor LOAEL-to-NOAEL extrapolation was not used because the Agency's
current approach is to address this factor as one of the considerations in selecting a BMR for
BMD modeling. In this case, a change in the mean equal to 1 SD of the control mean was
assumed to represent a minimal biologically significant change. The rationale for not applying
the UFLhas been added to section 5.1.3.
Comment 6: One reviewer disagreed with the use of a 10-fold database UFD, stating "if the
database is so uncertain as to require a UFD of 10, then the database is too limited to allow the
derivation of meaningful RfDs." This reviewer recommended a value of 1 based on the
relatively sensitive nature of the neurobehavioral endpoint, the consistent observation of the
neurobehavioral effects across the four PBDE congeners, and the availability of the dose-
A-9
-------
response data for deriving a BMDL. Another reviewer commented on the specificity and
sensitivity of the neurobehavioral and neurochemical measures and stated that it is inappropriate
to apply the database UFD. This reviewer observed that this UFD "addresses questions that go
beyond this endpoint and focus on risks that might occur, but there are no relevant data." The
reviewer felt that this UFD is more appropriately applied at the point of risk management. The
other two reviewers did not comment specifically on the database UFD but noted generally that
the database was poor and the overall confidence in the assessment is low (see next comment)
Response: EPA's practice is to apply a database UFD, generally ranging from 1-10, in the
health assessment to account for the potential for deriving an underprotective RfD as a result of
an incomplete characterization of a chemical's toxicity because of missing studies. In deciding
to apply this factor to account for deficiencies in the available data set and in identifying its
magnitude, EPA considers both the data lacking and the data available for particular organ
systems as well as life stages. EPA acknowledges that the principal study (involving postnatal
exposure) has identified what appears to be a relatively sensitive effect; however, the database
for BDE-99 lacks a two-generation reproduction study, as well as subchronic and chronic
toxicity studies. In light of the inadequate nature of the database, the Agency retains a 10-fold
UFD. The "IRIS Summary" and "Toxicological Review" contain sufficient information on the
rationale for the UFD to allow a risk manager to consider the impact of this UF during the risk
management process.
Comment 7: Two reviewers believed the database supports the determination of the RfD but
stated that the overall confidence in the RfD assessment is low. Another reviewer believed the
database is very poor and suggested that the RfDs be acknowledged as temporary while waiting
for additional studies that increase confidence.
Response: The statement that the overall confidence in the RfD is low is included in the
"Toxicological Review" in section 6.2. The Agency does not develop temporary RfDs for IRIS
assessments. However, the availability of new information is one of the factors considered in
selecting a chemical for reassessment.
C. Body Burden Approach
Charge Question 7. Are there adequate data for considering body burden as an alternative
dose metric to administered doses in any of the RfD derivations?
A-10
-------
Comment 1: All four reviewers agreed that the data were inadequate to consider body burden as
an alternative dose metric for the derivation of the RfD. Two of the reviewers stated that body
burden is a possible alternative but the data are too limited.
Response: EPA examined the data on BDE-99 to determine if a body burden approach could be
used for this congener during the development of the "lexicological Review." It was
determined that existing half-life, exposure, metabolite, and mode-of-action data could not
support a body burden calculation for this congener.
Charge Question 8. Do you agree with the rationale described in the "Toxicological Review "
of BDE-99 that the data on the window of susceptibility of the cholinergic receptors to BDE-99
tend to minimize body burden concerns?
Comment 1: Three reviewers stated that the question was unclear. One reviewer accepted the
concept as a basis for the experimental design, given the available information. A second
reviewer stated that there was no direct evidence that BDE-99 directly affects cholinergic
receptors and suggested that the mechanism of the interaction must be complex and indirect. A
third reviewer stated that, although there are no definitive data on mode of action, this
hypothesis is plausible. This reviewer acknowledges that the data on the window of
susceptibility of the cholinergic receptors to BDE-99 are suggestive but believes there are too
many other possibilities for mode of action for this rationale to minimize body burden concerns.
Response: Available mode-of-action data that describe the developmental neurotoxicity of
BDE-99 are limited. The Eriksson/Viberg group, the principal study authors, have hypothesized
that the observed effects on locomotion and habituation are related to impaired development of
the cholinergic system during the postnatal brain growth spurt period based on studies of BDE-
99 (Viberg et al., 2005, 2004b) and supported by studies with BDE-153 (Viberg et al., 2003a)
and BDE-209 (Viberg et al, 2007, 2003b). They have further hypothesized that the sensitivity of
the cholinergic system occurs in the vicinity of PND 10 and have tested this hypothesis by
varying the time of dosing and observing differences in the habituation effect for BDE-99 and -
209 (Viberg et al., 2007; Eriksson et al., 2002). The resulting deficit in cholinergic receptors
persisted across the duration of testing and could cause a hypoactive response to exposure to
cholinergic stimulants in adulthood. The following statement has been added to the mode-of-
action summary (section 4.5.3): "Although evidence exists that demonstrates BDE-99 as well as
other PBDEs interact at the neurological level, data are inadequate to determine the mode of
action for BDE-99."
A-ll
-------
Miscellaneous Comments
Commentsl: Three reviewers felt that the assessment would benefit from the combination of the
individual documents for the four congeners into one comprehensive document to compare and
cohesively present the similarities and differences among the congeners.
Response: The Agency has recently completed IRIS assessments for four individual PBDE
congeners: BDE-47, -99, BDE-153, and -209 (see Foreword). These congeners were selected
based on frequent detection in human tissues and the environment, availability in animal
toxicological studies suitable for human health assessment, and their common occurrence in
commercial PBDE mixtures. Although there is some repetition in the four documents, the
available database is sufficiently different from one congener to another to support the
separation of the four IRIS assessments. However, in response to the comments from the peer
reviewers, the Agency increased the text that compares the data on BDE-99 to that of the other
congeners evaluated using comparable methodological approaches.
Comment 2: One reviewer noted that the document failed to cite the purities of the radioactive
chemicals in most of the studies, the position of the label, location of radioactivity in the brain,
and the specific activities of the 14C compounds. Another reviewer was confused by the reliance
upon the 14C data and the intermixing with direct chemical measures. The reviewer felt that the
conclusions drawn were challenging.
Response: The requested data were added to the descriptions of the pertinent studies (in section
3) when they were provided by the authors of the paper. Frequently, the position of the
radiolabel was not specified. In a few cases the radiolabel was described as "uniform,"
suggesting that all carbons carried the radiolabel. If the authors of the paper used the term
"uniform," it was added to the discussion of the study. No change was made if the authors of the
paper did not comment on the position of the radiolabel.
Comment 3: One reviewer was concerned that the doses and concentrations of the compound
and the metabolites in biological tissues were presented in differing units (i.e., umol, ug, percent
of dose).
Response: Doses and concentrations are reported as given by the authors. If a dose was given in
molar or mole units per unit body weight, the doses have been provided parenthetically as mg/kg
body-weight values. Otherwise the units are those provided in the published papers.
A-12
-------
Comment 4: One reviewer acknowledged that developmental neurotoxicity is consistently
observed following exposure to the PBDEs despite very different patterns of metabolism,
distribution, and persistence within the body. This reviewer recommended rationalizing the
relative potency of the PBDEs, considering the differences in the extent of metabolism.
Response: Information is currently insufficient to adequately identify the relative potency of the
four congeners.
Comment 5: One reviewer suggested that the Agency provide conclusions on the extent of
metabolism and the presence of metabolites in excreta for the PBDEs or provide a statement if
conclusions cannot be drawn. One reviewer suggested the addition of a summary at the
beginning of the toxicokinetics section to reduce potential confusion.
Response: An overview to the toxicokinetics section and summary paragraphs have been added
to section 3. The data to determine the extent of metabolism are presently not available. There
do appear to be some differences in metabolism and excretion between mice and rats. The
available data are presented in sections 3.3 and 3.4.
Comment 6: One reviewer felt that the toxicokinetic information was not presented objectively.
Response: The toxicokinetic data were presented as they were reported. In synthesizing the
data, the authors consulted with Agency researchers when there was uncertainty in the
interpretation of results. Every attempt was made to accurately reflect the data in the published
papers.
Comment 7: One reviewer recommended presenting the receptor site interaction information in
a summary table.
Response: A summary table for the receptor studies has been added to section 4.4 of the
"Toxicological Review."
Comment 8: One reviewer recommended adding data for BDE-209 to Table 3-1, Median PBDE
congener concentrations in human biological media in the U.S.
Response: These data for BDE-209 were not available for inclusion in Table 3-2.
A-13
-------
Comment 9: One reviewer felt that the evolution of exposures that are different in the U.S.
compared with other countries and the pattern of exposures are important issues and the studies
need to be presented to support or refute these observations.
Response: The Agency has provided information on exposure in the U.S. and other countries for
comparison purposes. While the Agency agrees that exposure analysis is a critical component of
risk assessment, a comprehensive presentation and analysis of exposure data are outside the
scope of the IRIS health assessment.
Comment 10: One reviewer stated that the large number of bromine atoms of the PBDEs can
impart electrophilic and lipophilic properties to the aromatic ring of the chemical and also noted
that oily vehicles (e.g., corn oil) were used in most of the in vitro and in vivo animal studies.
This reviewer was concerned that the vehicle could significantly alter the distribution and tissue
uptake of the PBDEs between the oily vehicle and the biological system. These conditions could
lead to decreased absorption and distribution with subsequent alteration in metabolism and
excretion.
Response: The lipophilicity of BDE-99 is acknowledged in the "Toxicological Review" as part
of section 3 on toxicokinetics. It will be necessary to determine if absorption occurs via the
chylomicrons along with the body lipids or via direct membrane transport before the full impact
of the vehicle on absorption, distribution, metabolism, and excretion can be determined. The
data are currently inadequate to determine the impact of the oily vehicle on the distribution and
uptake of BDE-99.
Comment 11: One reviewer noted that, considering the antithyroid effects observed with DE-71
(a formulation that might be reasonably linked to neurodevelopmental effects observed with
BDE-47 and BDE-99) in Zhou et al. (2002) occurred at doses that are higher than those that
produce the neurodevelopmental effects of BDE-47 and BDE-99, then it must be concluded that
the neurodevelopmental effects cannot be linked to the antithyroid effects of these compounds.
Additionally, the antithyroid effects have not been substantiated.
Response: The Agency acknowledges that data are insufficient to determine the mode of action;
therefore, text was added to section 4.5.3 stating this conclusion.
Comment 12: One reviewer noted that the Hill model (utilized in modeling the rearing
habituation data in 8-month-old female mice [Viberg et al., 2004a]) fits best in terms of its
having the highest/* value; however, it also has the highest Akaike Information Criterion (AIC).
A-14
-------
This reviewer stated that this is not consistent with the approach utilized in the BDE-209
assessment, in which the lowest AIC is one of the criteria for model selection and suggested that
AIC be used consistently.
Response: The Agency reevaluated the approach utilized in analyzing the modeling results for
BDE-99 and determined that among the models that adequately fit the data, the power model
yielded the lowest AIC, indicating that it is the "best-fit" model for the rearing habituation data
in 8-month-old female mice (Viberg et al., 2004a). The rationale for this decision is explained
further in section 5.1.2 and Appendix B.
Comment 13: One reviewer felt that the use of a BMDL rather than a BMD as the point of
departure should be explicitly stated.
Response: The following statement can be found in section 5.1.3: "[T]he estimated BMDL1SD of
0.29 mg/kg based on a decrease in rearing habituation in 8-month-old female mice exposed to
BDE-99 on PND 10 (Viberg et al., 2004a) was selected as the point of departure for the RfD."
Comment 14: One reviewer did not understand why it was assumed that eight animals per sex
were tested behaviorally in Viberg et al. (2004a). Could this not be clarified by contacting the
study authors?
Response: The Agency reevaluated the study methods described in Viberg et al. (2004a). The
review of the study showed that the males and females were analyzed separately, and the
description of the study methods indicated that eight animals per analysis were randomly
selected from each treatment group. Consequently, eight males and eight females were analyzed
for each dose. Therefore, the statement that "we assume that there were eight males and eight
females exposed to each level" has been removed from the document.
Comment 15: One reviewer suggested raising the confidence in the RfD from low to medium,
considering the existence of multiple studies corroborating the main conclusions.
Response: The Agency acknowledges the existence of multiple studies that support the principal
study. However, considering the lack of a two-generation reproductive study and subchronic and
chronic toxicity studies, the Agency maintains the low confidence in the RfD for BDE-99.
PUBLIC COMMENTS
A-15
-------
The public commenters made several editorial suggestions to clarify specific portions of the text.
These changes were incorporated in the document as appropriate and are not discussed further.
Comments: One public commenter suggested that the Agency consider a body burden approach.
Response: The Agency presented this issue to the peer reviewers in the form of a charge
question. In response to the charge question about use of a body burden approach for dose
evaluation, the peer reviewers agreed that, whereas the body burden approach might be
appropriate for some of the congeners given their lipophilicity and distribution to adipose tissue,
data to support such an approach are not presently available.
Comments: One public commenter questioned the selection of Viberg et al. (2004a) as the
principal study for the derivation of the RfD and questioned the methods utilized by the principal
study authors.
Response: The Agency has included a detailed summary of the concerns with the study design
and methods utilized in the principal study (see section 5.1.1). These issues were raised during
the external peer review of the BDE-99 IRIS assessment. The peer reviewers acknowledged the
limitations and concerns with the study; however, all of the reviewers agreed that this study was
appropriate for derivation of the RfD for BDE-99 and that its limitations were transparently
discussed in the "Toxicological Review." Additionally, the neurobehavioral effects reported in
Viberg et al. (2004a) are supported by an expanding body of literature (Rice et al., 2007; Viberg
et al., 2007, 2005, 2004b, 2003a, b, 2002; Kuriyama et al., 2005; Eriksson et al., 2002, 2001;
Branch! et al., 2002) that details changes in motor and cognitive activity in rodents following
administration of single or repeated perinatal doses of PBDEs.
A-16
-------
APPENDIX B: BENCHMARK DOSE MODELING FOR BDE-99
METHODS
All dose-response modeling for this assessment was conducted by using EPA's BMDS.
Two different versions of this software were employed. BMDS version 1.3.2 was used to fit the
power and Hill models, while BMDS version 1.4 beta was used to fit the linear and polynomial
models. These two different versions were used because of specific types of errors encountered
with each version, as follows. When fitting continuous models in version 1.3.2 (i.e., linear,
polynomial, power, and Hill models), the degrees of freedom listed when the parameter
estimates hit a boundary are incorrect. This error appears to have been corrected in version 1.4
beta. However, preliminary analyses revealed that, in fitting the power and Hill models with
version 1.4 beta, the models occasionally failed to converge, while with version 1.3.2 no
convergence issues were encountered. Therefore, version 1.4 beta was used to fit linear and
polynomial models, while version 1.3.2 was used to fit the power and Hill models. With version
1.3.2, the degrees of freedom for the power and Hill models were corrected manually. ABMR
of one estimated SD (1.0 SD) from the control mean was used for all endpoints, consistent with
the Benchmark Dose Technical Guidance Document (U.S. EPA, 2000b).
A. Dose-Response Modeling Using the Data from Viberg et al. (2004a)
Rearing and locomotion habituation in male and female mice were each modeled
separately by using the continuous dose-response models available in BMDS (i.e., linear,
polynomial, power, and Hill models). All data sets failed the test for equality of variances across
dose groups, so the variance was modeled as a power function of the mean. In some cases, this
variance model failed to accurately estimate the observed variances; however, no alternative
variance model is currently available in BMDS. Therefore, in those cases where the modeled
variances yielded poor results, the differences between the observed and estimated standard
deviations were examined, especially in the region of the BMR. In all cases, these differences
were determined to be minimal, and thus the variance was modeled as a power function of the
mean in all modeling runs.
For both rearing and locomotion habituation, when data from all six dose groups were
included, all of the continuous models available in BMDS exhibited a significant lack of fit (i.e.,
goodness-of-fit/7 value <0.10) when default restrictions were placed on model parameters. For
some endpoints, satisfactory goodness-of-fit/> values (i.e.,/? > 0.10) were obtained by fitting a
polynomial model with no restrictions on the parameters (Figure B-l). However, this
unrestricted model yielded unreasonable behavior at low doses (i.e., in the region of the BMD)
as shown in Figure B-2. As Figure B-2 shows, the fitted model exhibited an initial decrease in
B-l
-------
the dose-response function, which is inconsistent with biological knowledge about the response
to PBDE exposure. Therefore, default parameter restrictions were employed in all models to
prevent such biologically implausible regions of decreasing dose response.
Polynomial Model with 0.95 Confidence Level
200
150
100
50
Polynomial
BMDLBMD
0
13:5508/052005
8
dose
10
12
14
16
Figure B-l. Unrestricted fourth-order polynomial model fit to rearing
habituation in 2-month-old male mice exposed to BDE-99.
Data source: Viberg et al. (2004a).
B-2
-------
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
Polynomial Model with 0.95 Confidence Level
Polynomial
BMDL
0 0.05
13:5808/052005
0.1
0.15 0.2
0.25
dose
0.3 0.35
0.4 0.45
0.5
Figure B-2. Low-dose behavior of unrestricted fourth-order polynomial
model fit to rearing habituation in 2-month-old male mice exposed to BDE-99.
Data source: Viberg et at. (2004a).
In most cases, when all six dose groups were included, the dose response trend at the
higher doses was different from that in the low-dose range. Therefore, to obtain a better fit to
the data at low doses (i.e., the region of interest), the highest dose group, 16 mg/kg body weight,
was dropped from the analysis. This is a reasonable approach given that the estimated BMD is
around 0.8 mg/kg, which is far below the high dose of 16 mg/kg. Among the four continuous
models available in BMDS, only the Hill model provided an adequate fit to these truncated data
sets, with adequate goodness-of-fit/> values (i.e.,/? > 0.10) obtained for the following endpoints:
rearing habituation in 2-month-old male mice and rearing habituation in 2- and 8-month-old
female mice (Tables B-l, B-3, and B-4, respectively). Even with the dropping of the high dose,
however, adequate model fits were not obtained for rearing habituation in 8-month-old male
mice, so for this data set the 8 mg/kg dose group was dropped as well. All of the continuous
models available in BMDS, except the Hill model, were fit to the remaining four dose groups.
This approach still seems reasonable because the retained high-dose group at 4 mg/kg is still
well above the region of the BMD. The Hill model was not fit to this truncated data set because
this model requires at least five dose groups in order to compute goodness-of-fit/> values.
Finally, adequate model fits were not obtained for any of the continuous models fit to the
locomotion habituation data employing either full or truncated data sets.
B-3
-------
For many of the endpoints, the mean response at low dose (i.e., 0.4 mg/kg) showed no
increase above the mean of the controls and also exhibited significantly smaller standard
deviations than the high-dose groups. As a result, the monotonically increasing dose-response
models in BMDS had difficulty fitting this low-dose mean, which resulted in poor goodness-of-
fit because of the very small SD at this low dose, even though a plot of the dose-response curve
appeared to fit the responses very well by visual examination. Despite the apparent good fit to
the data, BMD estimates from models that lacked goodness of fit (i.e.,/? < 0.1) were not
considered when determining the point of departure because the nonresponse at the 0.4 mg/kg
dose is so critical in determining the shape of the curve in the low-dose region.
BMD modeling results for all endpoints from Viberg et al. (2004a) are summarized in
Tables B-l through B-4. In addition to goodness-of-fit tests, AIC was employed for model
comparison and selection. The AIC is a function of the log-likelihood, and the numbers of
parameters in the model, as well as each data point, contribute to the log-likelihood. Thus, the
AIC can be used to compare model fits across models fit to the same data sets. The AIC was
employed to make comparisons across the linear, polynomial, and power models fit to the same
data sets in Tables B-l through B-4.
Modeling Results
1. Rearing habituation in 2-month-old male mice
The BMD modeling results for rearing habituation in 2-month-old male mice are
summarized in Table B-l. Based on the results of the chi-square goodness-of-fit tests, the Hill
and power models did not exhibit significant lack of fit, while the linear and polynomial models
did, and thus were not considered further in the derivation of the point of departure. Of the two
models that did not exhibit significant lack of fit, the power model provided the best fit to the
data because this model had a lower AIC value than the Hill model (i.e., 39 versus 96,
respectively). Thus, for rearing habituation in 2-month-old male mice, the BMDj OSD estimate
selected is 0.59 mg/kg, and its corresponding BMDL,! OSD is 0.44 mg/kg.
Table B-l. Rearing habituation in 2-month-old male mice
Model
Hill
Linear
Polynomial
Power
No.
Groups
5
4
4
4
p Value
0.84
0.01
0.029
0.71
AIC
96
86
44
39
BMD10SD
(mg/kg)
0.63
4.2
0.39
0.59
BMDL10SD
(mg/kg)
0.48
1.7
0.33
0.44
Source: Viberg et al. (2004a).
B-4
-------
Standard BMDS Output from Fitting the Power Model to Rearing Habituation in
2-Month-Old Male Mice (8 and 16 mg/kg-day Dose Groups Omitted)
Power Model
BMR = 1.0 SD
Power Model. $Revision: 2.1 $ $Date: 2000/10/11 20:57:36 $
Input Data File: S:\PROJECT FILES\EPA DECABDE\DBDE2\TETRA PENTA
BMD\VMRTRUNC.(d)
Gnuplot Plotting File: S:\PROJECT FILES\EPA DECABDE\DBDE2\TETRA PENTA
BMD\VMRTRUNC.pit
Thu Aug 04 13:53:06 2005
BMDS MODEL RUN
The form of the response function is:
Y[dose] = control + slope * doseApower
Dependent variable = MEAN
Independent variable = mg/kg
The power is restricted to be greater than or equal to 1
The variance is to be modeled as Var(i) = alpha*mean(i)Arho
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha = 5.05143
rho = 0
control = 0.18
slope = 1.46193
power = 2.47391
Asymptotic Correlation Matrix of Parameter Estimates
alpha
rho
control
slope
power
alpha
1
-0.38
-0.54
-0.053
0.056
-0
0
-0
0
rho
.38
1
.46
.12
.11
control
-0
0
-0
0
.54
.46
1
.37
.36
slope
-0.053
-0.12
-0.37
1
-0.99
power
0.056
0.11
0.36
-0.99
1
Variable
alpha
rho
control
slope
Parameter Estimates
Estimate Std. Err.
0.670595 0.21373
0.818561 0.129962
0.222551 0.112891
1.57485 0.418633
B-5
-------
power
2.41986
0.191752
Table of Data and Estimated Values of Interest
Dose N Obs Mean Obs Std Dev Est Mean Est Std Dev ChiA2 Res,
0
0.4
0.8
4
0.18
0.32
1.27
45.3
0.52
0.59
0.76
4.36
0.223
0.394
1.14
45.3
0.443
0.559
0.864
3.9
-0.0961
-0.132
0.15
-0.00486
Model Descriptions for likelihoods calculated
Model Al: Yij
Var{e(ij)}
Model A2: Yij
Var{e(ij)}
Model A3: Yij
Var{e(ij)}
Model R: Yi
Var{e(i)}
Mu (i) + e (ij )
SigmaA2
Mu (i) + e (ij )
Sigma(i)A2
Mu (i) + e (ij )
alpha*(Mu(i))Arho
Mu + e(i)
S i gma A 2
Likelihoods of Interest
Model Log(likelihood) DF AIC
Al -39.778224 5 89.556448
A2 -13.995306 8 43.990611
A3 -14.484425 6 40.968849
fitted -14.554490 5 39.108980
R -111.525662 2 227.051325
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adequately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
Tests of Interest
-2*log(Likelihood Ratio)
d.f
p-value
195.061
51.5658
0.978238
0.140131
6
3
2
1
<. 00001
<. 00001
0.6132
0.7082
Test
Test 1
Test 2
Test 3
Test 4
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is less than .05. A non-homogeneous variance
model appears to be appropriate
The p-value for Test 3 is greater than .05. The modeled variance appears
-------
to be appropriate here
The p-value for Test 4 is greater than .05. The model chosen seems
to adequately describe the data
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD = 0.591919
BMDL = 0.435548
B-7
-------
Power Model with 0.95 Confidence Level
50
40
CD
-------
2. Rearing habituation in 8-month-old male mice
The BMD modeling results for rearing habituation in 8-month-old male mice are
summarized in Table B-2. According to the results of the chi-square goodness-of-fit tests, none
of the continuous models in BMDS provide satisfactory fits to these data because all of the
models exhibited significant lack of fit (p < 0.1). This lack of fit may be due primarily to the
mean response reported in the 0.4 mg/kg dose group, which was zero. The published manuscript
from which these data were extracted did not indicate the basis for this zero value at 0.4 mg/kg
(i.e., whether the calculated values were very small and reported as zero or whether no data were
obtained for this dose group). Because this dose was in the vicinity of the BMD, modeling was
not conducted with this dose group dropped from the data set. Therefore, because of this failure
of the models to fit the data, no BMD or BMDL could be derived based on this endpoint.
Table B-2. Rearing habituation in 8-month-old male mice
Model
Hill
Linear
Polynomial
Power
No. of
groups
5
4
4
4
p Value
0.01
<0.01
0.01
0.01
AIC
104
160
62
60
BMD10SD
(mg/kg)
0.57
0.055
0.42
0.56
BMDL10SD
(mg/kg)
0.45
0.019
0.36
0.42
Source: Viberg et al. (2004a).
3. Rearing habituation in 2-month-old female mice
The BMD modeling results for rearing habituation in 2-month-old female mice are
summarized in Table B-3. According to the results of the chi-square goodness-of-fit tests, the
Hill, polynomial, and power models do not exhibit significant lack of fit. Among these three
models that adequately fit the data, the power model yielded the lowest AIC, indicating that it is
the best-fit model. Thus, for the endpoint rearing habituation in 2-month-old female mice, the
! OSD estimate selected is 0.70 mg/kg, and its corresponding BMDLi OSD is 0.47 mg/kg.
Table B-3. Rearing habituation in 2-month-old female mice
Model
Hill
Linear
Polynomial
Power
No. of
groups
5
4
4
4
p Value
0.53
0.01
0.12
0.62
AIC
122
130
73
71
BMD10SD
(mg/kg)
0.71
0.065
0.45
0.70
BMDL10SD
(mg/kg)
0.50
0.034
0.38
0.47
Source: Viberg et al. (2004a).
B-9
-------
Standard BMDS Output from Fitting the Power Model to Rearing Habituation in
2-Month-Old Female Mice (8 and 16 mg/kg-day Dose Groups Omitted)
Power Model
BMR = 1.0 SD
Power Model. $Revision: 2.1 $ $Date: 2000/10/11 20:57:36 $
Input Data File: S:\PROJECT FILES\EPA DECABDE\DBDE2\TETRA PENTA
BMD\VFRTRUNC.(d)
Gnuplot Plotting File: S:\PROJECT FILES\EPA DECABDE\DBDE2\TETRA PENTA
BMD\VFRTRUNC.plt
Fri Aug 05 10:23:14 2005
BMDS MODEL RUN
The form of the response function is:
Y[dose] = control + slope * doseApower
Dependent variable = MEAN
Independent variable = mg/kg
The power is restricted to be greater than or equal to 1
The variance is to be modeled as Var(i) = alpha*mean(i)Arho
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha = 30.2634
rho = 0
control = 0.24
slope = 2.07452
power = 2.22846
Asymptotic Correlation Matrix of Parameter Estimates
alpha
rho
control
slope
power
alpha
1
-0.54
-0.64
-0.076
0.077
rho
-0.54
1
0.52
-0.097
0.068
control
-0.64
0.52
1
-0.41
0.4
slope
-0.076
-0.097
-0.41
1
-0.98
power
0.077
0.068
0.4
-0.98
1
Variable
alpha
rho
control
slope
Parameter Estimates
Estimate Std. Err.
1.22736 0.477381
1.13527 0.149123
0.37407 0.172425
1.52431 0.596555
B-10
-------
power
2.44986
0.283549
Table of Data and Estimated Values of Interest
Dose N Obs Mean Obs Std Dev Est Mean Est Std Dev ChiA2 Res.
0
0.4
0.8
4
8
8
8
8
0.24
0.51
1.49
45.8
0.69
0.95
0.93
10.9
0.374
0.536
1.26
45.9
0.634
0.777
1.26
9.72
-0.211
-0.0329
0.185
-0.00785
Model Descriptions for likelihoods calculated
Model Al: Yij
Var{e(ij)}
Model A2: Yij
Var{e(ij)}
Model A3: Yij
Var{e(ij)}
Model R: Yi
Var{e(i)}
Mu (i) + e (ij )
SigmaA2
Mu (i) + e (ij )
Sigma(i)A2
Mu (i) + e (ij )
alpha*(Mu(i))Arho
Mu + e(i)
S i gma A 2
Likelihoods of Interest
Model Log(likelihood) DF AIC
Al -68.422509 5 146.845019
A2 -29.014179 8 74.028357
A3 -30.262812 6 72.525625
fitted -30.387704 5 70.775407
R -112.660182 2 229.320365
Test 1:
Test 2:
Test 3:
Test 4:
Explanation of Tests
Does response and/or variances differ among Dose levels?
(A2 vs. R)
Are Variances Homogeneous? (Al vs A2)
Are variances adequately modeled? (A2 vs. A3)
Does the Model for the Mean Fit? (A3 vs. fitted)
Tests of Interest
-2*log(Likelihood Ratio)
d.f
p-value
167.292
78.8167
2.49727
0.249782
6
3
2
1
<. 00001
<. 00001
0.2869
0.6172
Test
Test 1
Test 2
Test 3
Test 4
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is less than .05. A non-homogeneous variance
model appears to be appropriate
The p-value for Test 3 is greater than .05. The modeled variance appears
B-ll
-------
Power Model with 0.95 Confidence Level
CD
CO
60
50
40
to" 30
CD
CO
CD
10
Power
BMDL BMD
0 0.5
10:2308/052005
1
1.5
2
dose
2.5
3.5
to be appropriate here
The p-value for Test 4 is greater than .05.
to adequately describe the data
Benchmark Dose Computation
Specified effect =
The model chosen seems
Risk Type
Confidence level =
BMD =
BMDL =
Estimated standard deviations from the control mean
0.95
0.699009
0.465291
Figure B-4. Dose-response relationship based on rearing habituation in 2-
month-old female mice.
Data source: Viberg et al. (2004a).
B-12
-------
4. Rearing habituation in 8-month-old female mice
The BMD modeling results for rearing habituation in 8-month-old female mice are
summarized in Table B-4. According to the results of the chi-square goodness-of-fit tests, the
Hill and power models do not exhibit significant lack of fit. Between these two models that
adequately fit the data, the power model yielded the lowest AIC, indicating that it is the "best-
fit" model. Thus, for the endpoint rearing habituation in 8-month-old female mice, the
! OSD estimate selected is 0.41 mg/kg, and its corresponding BMDLi OSD is 0.29 mg/kg.
Table B-4. Rearing habituation in 8-month-old female mice
Model
Hill
Linear
Polynomial
Power
No. of
groups
5
4
4
4
p Value
0.21
0.01
0.01
0.16
AIC
110
100
95
63
BMD10SD
(mg/kg)
0.42
1.9
0.34
0.41
BMDL10SD
(mg/kg)
0.32
0.56
0.093
0.29
Source: Viberg et al. (2004a).
B-13
-------
Standard BMDS Output from Fitting the Power Model to Rearing Habituation in
8-Month-Old Female Mice (8 and 16 mg/kg-day Dose Groups Omitted)
Power Model
BMR = 1.0 SD
Power Model. $Revision: 2.1 $ $Date: 2000/10/11 20:57:36 $
Input Data File: S:\PROJECT FILES\EPA DECABDE\DBDE2\TETRA PENTA
BMD\VFRTRUNC.(d)
Gnuplot Plotting File: S:\PROJECT FILES\EPA DECABDE\DBDE2\TETRA PENTA
BMD\VFRTRUNC.plt
Fri Aug 05 10:29:26 2005
BMDS MODEL RUN
The form of the response function is:
Y[dose] = control + slope * doseApower
Dependent variable = MEAN
Independent variable = mg/kg
The power is restricted to be greater than or equal to 1
The variance is to be modeled as Var(i) = alpha*mean(i)Arho
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha = 14.0337
rho = 0
control = 0.22
slope = 1.46284
power = 2.44841
Asymptotic Correlation Matrix of Parameter Estimates
alpha
rho
alpha
1
-0.51
rho
-0.51
1
control
-0.59
0.51
slope
-0.13
-0.13
power
0.13
0.11
control
slope
power
-0.59
-0.13
0.13
0.51
-0.13
0.11
1
-0.2
0.18
-0.2
1
-0.96
0.18
-0.96
1
Parameter Estimates
Variable
alpha
rho
control
slope
Estimate
0.931829
0.985437
0.268288
2.91602
Std. Err.
0.321871
0.137401
0.124922
0.525713
B-14
-------
power
1.95332
0.131911
Table of Data and Estimated Values of Interest
Dose N Obs Mean Obs Std Dev Est Mean Est Std Dev ChiA2 Res,
0
0.4
0.8
4
8
8
8
8
0.22
0.33
2.83
43.8
0.61
0.83
0.56
7.4
0.268
0.755
2.15
44
0.505
0.841
1.41
6.23
-0.0957
-0.506
0.48
-0.0323
Model Descriptions for likelihoods calculated
Model Al: Yij
Var{e(ij)}
Model A2: Yij
Var{e(ij)}
Model A3: Yij
Var{e(ij)}
Model R: Yi
Var{e(i)}
Mu (i) + e (ij )
SigmaA2
Mu (i) + e (ij )
Sigma(i)A2
Mu (i) + e (ij )
alpha*(Mu(i))Arho
Mu + e(i)
S i gma A 2
Likelihoods of Interest
Model Log(likelihood) DF AIC
Al -56.126826 5 122.253652
A2 -19.791783 8 55.583565
A3 -25.553315 6 63.106629
fitted -26.518839 5 63.037677
R -110.453252 2 224.906503
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adequately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
Tests of Interest
-2*log(Likelihood Ratio)
d.f
p-value
181.323
72.6701
11.5231
1.93105
6
3
2
1
<. 00001
<. 00001
0.003146
0.1646
Test
Test 1
Test 2
Test 3
Test 4
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is less than .05. A non-homogeneous variance
model appears to be appropriate
The p-value for Test 3 is less than .05. You may want to consider a
B-15
-------
Power Model with 0.95 Confidence Level
50
40
CD
CO
o 30
Q.
CO
CD
CD
20
10
0
Power
BMDL BMP
0 0.5
10:2908/052005
1.5
2
dose
2.5
3.5
different variance model
The p-value for Test 4 is greater than .05. The model chosen seems
to adequately describe the data
Benchmark Dose Computation
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
Estimated standard deviations from the control mean
0.95
0.407446
0.292715
Figure B-5. Dose-response relationship based on rearing habituation in 8-
month-old female mice.
Data source: Viberg et al. (2004a).
B-16
-------
5. Locomotion habituation in BDE-99-treated male and female mice
All of the continuous dose-response models available in BMDS (i.e., Hill, linear,
polynomial, and power models) exhibited significant lack of fit (p < 0.1) when fit to the
locomotion habituation data in either male or female mice from Viberg et al. (2004a), even when
dose groups were dropped. Thus, BMD modeling results for this endpoint are not presented
here.
6. Summary
The selected BMD and BMDL estimates based on each of the four endpoints under
consideration are summarized in Table B-5. Based on these results, the lowest BMD1SD and
BMDL1SD estimates were derived from rearing habituation response in 8-month-old female mice,
and the corresponding estimates were 0.41 and 0.29 mg/kg, respectively. As described above,
the power model employing the rearing habituation data from four of the six dose groups in
8-month-old female mice from Viberg et al. (2004a) yielded the best fit. Figure B-5 displays the
power model fit to the rearing habituation data in 8-month-old female mice. Following this
graph, the standard BMDS output from the power model is presented.
Table B-5. Summary of BMD and BMDL results
Endpoint
Rearing
habituation
Rearing
habituation
Rearing
habituation
Rearing
habituation
Age
2-mo nth-old
8-month-old
2-mo nth-old
8-month-old
Species, sex
Mouse, male
Mouse, male
Mouse, female
Mouse, female
BMD10SD
(mg/kg)
0.59
-
0.70
0.41
BMDL10SD
(mg/kg)
0.44
-
0.47
0.29
Source: Viberg et al. (2004a).
B. Dose-Response Modeling Using the Data from Eriksson et al. (2001)
The locomotion, rearing, and total activity habituation data in male mice from Eriksson
et al. (2001) were modeled in BMDS, using three of the four available continuous dose-response
models (i.e., linear, polynomial, and power models). The Flill model was not used in modeling
these data because this model requires at least five dose groups, and the Eriksson et al. (2001)
study employed only three dose groups.
Similar to when the Viberg et al. (2004a) data were modeled, the test for equality of
variances across dose groups failed for all data sets modeled, and so the variance was modeled
as a power function of the mean. In some cases, this variance model failed to accurately
B-17
-------
estimate the observed variances; however, no alternative variance model is currently available in
BMDS. Therefore, in those cases where the modeled variances yielded poor results, the
differences between the observed and estimated SDs were examined, especially in the region of
the BMR. In all cases, these differences were determined to be minimal, and thus the variance
was modeled as a power function of the mean in all modeling runs.
As was the case in modeling the Viberg et al. (2004a) data, restrictions were placed on
model parameters to avoid biologically implausible dose-response relationships. More
specifically, the parameter estimates for the polynomial model were restricted to be nonnegative,
while the power parameter in the power model was restricted to be greater than or equal to 1.
These restrictions were applied to the parameter estimates to prevent nonmonotonic dose-
response for the polynomial model and supralinear curves for the power model. A supralinear
dose-response curve was not considered plausible, given the responses observed in the Viberg et
al. (2004a) study that employed tighter dose spacing (i.e., more data points in the low-dose
region).
Modeling Results
All of the continuous models available in BMDS (i.e., linear, polynomial, power, and
Hill models) exhibited significant lack of fit (p < 0.1) for all endpoints evaluated from the
Eriksson et al. (2001) study. Although plots of the fits of the polynomial and power models to
the 2-month rearing data appear good, with only three data points, these models possess too
many parameters to evaluate goodness of fit. Also, Viberg et al. (2004a) employed an additional
low dose at 0.4 mg/kg at which there was no response, while Eriksson et al. (2001) only tested
two doses above controls (i.e., 0.8 and 12 mg/kg). Eriksson et al. (2001) observed responses at
both of these doses. Therefore, Eriksson et al. (2001) likely missed the initial flat portion of the
dose-response curve that was captured by Viberg et al. (2004a), leading to nearly linear fits
through the data points using the polynomial and power models. Because no adequately fitting
models were found among the available continuous models in BMDS, no BMD or BMDL
estimates are reported from this study.
C. Dose-Response Modeling Using the Data from Kuriyama et al. (2005)
Kuriyama et al. (2005) exposed rat dams to a single PBDE dose on GD 6 (at dose levels
of 0, 60, or 300 |ig/kg) and observed the response in the male offspring of these exposed dams.
Although this was a nested experimental design, the data were not presented in a form amenable
to a nested analysis. Therefore, the data were simply modeled using the standard continuous and
quantal models available in BMDS.
The effects modeled as continuous variables were locomotor activity on PNDs 36 and 71
(i.e., LBI counts per day, duration of activity [in hours] per day, LBI counts per active phase, and
B-18
-------
duration of activity [in minutes] per active phase), spermatid count, daily sperm production, and
sperm number on PND 140. Other endpoints assessed in the study were not evaluated using
BMD modeling because inspection of the data revealed no clear dose-response relationship.
In fitting the continuous models, the test for equality of variances across dose groups
failed for the locomotor activity data, and so the variance was modeled as a power function of
the mean. In some cases, this variance model failed to accurately estimate the observed
variances; however, no alternative variance model is currently available in BMDS. Therefore, in
those cases where the modeled variances yielded poor results, the differences between the
observed and estimated SDs were examined, especially in the region of the BMR. In all cases,
these differences were determined to be minimal, and, thus, the variance was modeled as a
power function of the mean in all modeling runs. For the spermatid data, however, variances
were determined to be homogeneous across dose groups; therefore, a constant variance model
was deemed appropriate for use in dose-response modeling.
As was the case in modeling the Viberg et al. (2004a) and the Eriksson et al. (2001) data,
restrictions were placed on model parameters to avoid biologically implausible dose-response
relationships. More specifically, the parameter estimates for the polynomial model were
restricted to be nonnegative, while the power parameter in the power model was restricted to be
greater than or equal to 1. These restrictions were applied to the parameter estimates to prevent
nonmonotonic dose response for the polynomial model and supralinear curves for the power
model.
The quantal data from this study were observations of the percent (or proportion) of adult
animals with two or more ejaculations. In order to make these data amenable to modeling with
quantal models, these data had to be converted so that adversity increased with increasing dose
(i.e., the percent [or proportion] of adult animals with less than two ejaculations).
Modeling Results
1. Effects modeled as continuous variables
The effects modeled as continuous variables were locomotor activity on PNDs 36 and 71
(i.e., LBI counts per day, duration of activity [in hours] per day, LBI counts per active phase, and
duration of activity [in minutes] per active phase [these data were presented in the published
manuscript in graphical form, but the raw data were obtained from the authors]) as well as
spermatid count, daily sperm production, and sperm number on PND 140. Among all of these
continuous endpoints, only duration per day and LBI counts per phase on PND 36 were
adequately fit by the linear and polynomial models. These modeling results are summarized in
Table B-6.
For the endpoint duration per day and of the two models that showed adequate fit, the
polynomial model provided the best fit based on its lower AIC value (Figure B-6); therefore, the
B-19
-------
BMD and BMDL estimates of 0.28 and 0.22 mg/kg, respectively, were selected based on this
endpoint. For the endpoint LBI counts per phase, only the linear model provided a satisfactory
fit to the data. The BMD1SD and BMDL1SD estimates based on this endpoint are 0.16 and
0.11 mg/kg, respectively.
Table B-6. BMD modeling results for the endpoints duration per day and
LBI counts per phase on PND 36 for male rats exposed to BDE-99 in utero on
GD6
Endpoint
Duration per day
Duration per day
LBI counts per phase
Model
Linear
Polynomial
Linear
No. groups
3
o
J
o
J
p Value
0.15
0.67
0.92
AIC
196
194
415
BMD10SD
0.25
0.28
0.16
BMDL10SD
0.17
0.22
0.11
Source: Kuriyamaet al. (2005).
Figure B-6 displays the polynomial model fit to the duration per day data in male rats
exposed to BDE-99 in utero on GD 6 (Kuriyama et al., 2005). Following this graph, the standard
BMDS output from the fitting of the polynomial model to these data is presented.
Polynomial Model with 0.95 Confidence Level
8.5
8
7.5
7
6.5
Polynomial
BMDL
BMD
0
09:32 08/08 2005
50
100
150
dose
200
250
300
Figure B-6. Dose-response relationship based on activity duration per day on
PND 36 in male rats exposed to BDE-99 in utero on GD 6.
Data source: Kuriyama et al. (2005).
B-20
-------
BMR = 1.0 SD
Polynomial Model. (Version: 2.3; Date: 6/21/2005)
Input Data File: S:\PROJECT FILES\EPA DECABDE\DBDE2\TETRA PENTA
BMD\KURIYAMAPND36.(d)
Gnuplot Plotting File: S:\PROJECT FILES\EPA DECABDE\DBDE2\TETRA PENTA
BMD\KURIYAMAPND36.plt
Mon Aug 08 09:32:55 2005
BMDS MODEL RUN
The form of the response function is:
Y[dose] = beta_0 + beta_l*dose + beta_2*doseA2 + ...
Dependent variable = MEAN
Independent variable = dose
The polynomial coefficients are restricted to be positive
The variance is to be modeled as Var(i) = alpha*mean(i)Arho
Total number of dose groups = 3
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha = 2.80847
rho = 0
beta_0 = 7
beta_l = 0
beta_2 = 0
!!! Warning: optimum may not have been found. !!!
!!! Bad completion code in maximum likelihood optimization routine !!!
i i i
You may want to try choosing different initial values.
Parameter Estimates
i i i
95.0% Wald Confidence Interval
Variable
alpha
rho
beta 0
beta 1
beta 2
Estimate
9.41543e-006
6.30301
6.74761
1.91982e-041
1.63437e-005
Std. Err.
3.00252e-005
1.61672
0.154227
NA
5.22144e-006
Lower Conf.
Limit
-4.94328e-005
3.13429
6.44533
6.10987e-006
Upper Conf.
Limit
6.82637e-005
9.47173
7.04989
2.65775e-005
NA - Indicates that this parameter has hit a bound implied by some inequality
constraint and thus has no standard error.
Asymptotic Correlation Matrix of Parameter Estimates
alpha
rho
beta 0
beta 2
alpha
1
-1
-0.016
0.005
rho
-1
1
0.016
-0.0041
beta 0
-0.016
0.016
1
-0.38
beta 2
0.005
-0.0041
-0.38
1
The following parameter(s) have been estimated at a boundary point or have
been specified. Correlations are not computed:
beta 1
B-21
-------
Table of Data and Estimated Values of Interest
Dose
0
60
300
N
32
40
29
Obs Mean
7
6.6
8.2
Obs Std Dev
1.1
1.4
2.4
Est Mean
6.75
6.81
8.22
Est Std
Dev
1.26
1.29
2.34
ChiA2
Res .
1.13
-1.01
-0.0426
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = alpha*(Mu(i))Arho
Model R: Yi = Mu + e(i)
Var{e(i)} = SigmaA2
Likelihoods of Interest
Model
Al
A2
A3
fitted
R
Log (likelihood)
-101.125573
-90.874249
-92.783358
-92.871731
-108.721847
d.f .
4
6
5
4
2
AIC
210.251146
193.748498
195.566717
193.743462
221.443695
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adequately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
(Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
Tests of Interest
Test
Test 1
Test 2
Test 3
Test 4
-2*log (Likelihood
Ratio)
35.6952
20.5026
3.81822
0.176745
Test df
4
2
1
1
p-value
<.0001
<.0001
0.0507
0.6742
The p-value for Test 1 is less than .05. There appears to be a difference
between response and/or variances among the dose levels. It seems appropriate
to model the data
The p-value for Test 2 is less than .1.
appears to be appropriate
The p-value for Test 3 is less than .1.
variance model
The p-value for Test 4 is greater than
adequately describe the data
A non-homogeneous variance model
You may want to consider a different
1. The model chosen seems to
Benchmark Dose Computation
B-22
-------
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD = 277.537
BMDL = 223.634
B-23
-------
2. Effects modeled as quantal variables
The percent (or proportion) of adult animals with less than two ejaculations was modeled
using the dichotomous models currently available in BMDS. A BMR of 10% extra risk was
chosen, consistent with the Benchmark Dose Technical Guidance Document (U.S. EPA, 2000b).
Prior to modeling, all doses were converted from |ig/kg to mg/kg. Table B-7 summarizes the
BMD modeling results based on this endpoint. None of the dichotomous models exhibited a
significant lack of fit (p > 0.1). Of the models fit to the data, the best-fitting model is the log-
logistic, as it has the lowest AIC, although the AICs for all of the models in Table B-7 are not
significantly different from each other. The fit of the log-logistic model to the data is shown in
Figure B-7, followed by the standard BMDS output generated by this model. Based on this
model, the estimated BMD is 0.019 mg/kg and its corresponding BMDL is 0.0067 mg/kg.
Table B-7. Summary of the quantal dose-response modeling results based on
the percent of animals with less than two ejaculations in male rats exposed to
BDE-99 in utero on GD 6
Model
Gamma
Log-logistic
Logistic
Log-probit
Probit
Multistage
Quantal linear
Quantal quadratic
Weibull
Chi2 Res
0.28
0.041
0.38
0.50
0.39
0.13
0.28
0.62
0.28
p Value
0.72
0.96
0.63
0.52
0.61
0.72
0.72
0.40
0.72
AIC
78.6
78.5
78.7
78.9
78.7
78.6
78.6
79.2
78.6
BMD10
(mg/kg)
0.031
0.019
0.041
0.058
0.043
0.031
0.031
0.10
0.031
BMDL10
(mg/kg)
0.016
0.0067
0.025
0.027
0.027
0.016
0.016
0.071
0.016
Data source: Kuriyama et al. (2005).
B-24
-------
Log-Logistic Model with 0.95 Confidence Level
1
Log-Logistic
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2 &• BMDL BMD
0 0.05 0.1 0.15 0.2 0.25 0.3
dose
09:03 08/08 2005
Figure B-7. Log-logistic model fit based on percent of animals with less than
two ejaculations in male rats exposed to BDE-99 in utero on GD 6.
Data source: Kuriyama et al. (2005).
B-25
-------
Logistic Model $Revision: 2.3 $ $Date: 2003/09/03 11:53:20 $
Input Data File: C:\MY DOCUMENTS\BMD\LL1.4.(d)
Gnuplot Plotting File: C:\MY DOCUMENTS\BMD\LL1.4.pit
Sun Aug 07 21:25:09 2005
HMDS MODEL RUN
The form of the probability function is:
P[response] = background+(1-background)/[1+EXP(-intercept-siope*Log(dose))]
Dependent variable = Column2
Independent variable = COLUMN1
Slope parameter is restricted as slope >= 1
Total number of observations = 3
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
User has chosen the log transformed model
Default Initial Parameter Values
background = 0.45
intercept = 1.7933
slope = 1
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -slope
have been estimated at a boundary point, or have been
specified by the user,
and do not appear in the correlation matrix )
background intercept
background 1 -0.6
intercept
Interval
Variable
Upper Conf. Limit
background
0.657517
intercept
3.244
slope
-0.6
Parameter Estimates
Estimate Std. Err.
0.452005 0.104855
1.77733 0.730579
1 NA
95.0% Wald Confidence
Lower Conf. Limit
0.246492
0.345417
14556+^+0924
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
B-26
-------
Model
Full model
Fitted model
Reduced model
AIC:
Analysis of Deviance Table
Log(likelihood)
-37.2311
-37.2324
-39.9403
78.4649
Deviance Test d.f.
0.002753
5.41857
P-value
0.9582
0.06658
Goodness of Fit
Dose
0.0000
0.0600
0.3000
Est. Prob.
0.4520
0.5955
0.8025
Expected 01
9.040
11.911
16.049
^served J
9
12
16
Size
20
20
20
Scaled
Residual
-0.01801
0.04075
-0.02772
ChiA2 = 0.002753
d.f. = 1
P-value = 0.9582
Benchmark Dose Computation
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
0.1
Extra risk
0.95
0.0187877
0.00670333
B-27
------- |