National Academies Press: OpenBook

Health Implications of Perchlorate Ingestion (2005)

Chapter: Appendix E Physiologically Based Pharmacokinetic Modeling

« Previous: Appendix D Sensitivity of Perchlorate-Induced Iodide Uptake Inhibition to Serum Iodide Concentrations
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 219
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 220
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 221
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 222
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 223
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 224
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 225
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 226
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 227
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 228
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 229
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 230
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 231
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 232
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 233
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 234
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 235
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 236
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 237
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 238
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 239
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 240
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 241
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 242
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 243
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 244
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 245
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 246
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 247
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 248
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 249
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 250
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 251
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 252
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 253
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 254
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 255
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 256
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 257
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 258
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 259
Suggested Citation:"Appendix E Physiologically Based Pharmacokinetic Modeling." National Research Council. 2005. Health Implications of Perchlorate Ingestion. Washington, DC: The National Academies Press. doi: 10.17226/11202.
×
Page 260

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Appendix E Physiologically Based Pharmacokinetic Modeling AS discussed in Chapter 4, physiologically based pharmacokinetic (PBPK) modeling is one of the methods of choice for determining human equivalent exposures and adjusting default uncertainty factors associated with the derivation of reference doses (RfDs) and reference concentrations (RfCs) for lifetime human exposures from animal studies (EPA 2002a). Thus, the U.S. Environmental Protection Agency (EPA) relied on a series of PBPK models developed by the Department of Defense (DOD) for perchlorate to facilitate interspecies extrapolations in its draft risk assessment (EPA 2002b,c). The PBPK models were initially developed to describe the dispo- sition (absorption, distribution, metabolism, and elimination) of perchlorate in adult rats (Fisher et al. 2000). As data became available and perchlo- rate-induced effects observed in animal studies and humans were shown to be mediated by interactions with iodide at the sodium-iodide symporter (NIS) in thyroid tissues, the initial model was expanded to include the disposition of iodide in the body and the inhibition of iodide uptake at the NIS in pregnant rats and fetuses, lactating rats and pups, and adult humans to address dose-response issues associated with potentially sensitive popula- tions (Clewell et al. 2001, 2003a,b; Merrill 2001b, Merrill et al. 2003). The purpose of the models was to facilitate extrapolation of internal, target-tissue doses from animals to humans, of high to low dose, and across routes of exposure in human health risk assessments. However, the chal- lenges associated with developing quantitative descriptions of the complex interactions and feedback mechanisms within the hypothalamic-pituitary- thyroid (HPT) axis prevented the development of a model that could de- scribe the dynamic interactions between the inhibition of iodide transport 219

220 Health Implications of Perchlorate Ingestion by the NIS and thyroid function. Thus, the PBPK models that were devel- oped by DOD focused on improving the ability to predict the kinetics of perchlorate and iodide with respect to their interaction at the NIS as the key event identified by EPA over a range of perchlorate exposure that encom- passes toxicity studies in animals, therapeutic uses in humans, and relevant environmental exposures. EPA used that approach in the derivation of an RfD for perchlorate, following established guidelines for the selection of key events, points of departure, and derivation of appropriate uncertainty factors (EPA 2002a,b,c). Although the committee has chosen a different point of depar- ture based on human data (see Chapter 5), it agrees with EPA that PBPK modeling constitutes the best approach to determining the human equivalent exposures and adjustments to default uncertainty factors when RfDs are based on data collected in animals. If future studies with perchlorate are conducted in laboratory animals, the PBPK models developed to date may be important tools for integrating the new data into the existing database on the exposure-dose-response relationships for perchlorate in rats and hu- mans. EPA provided a thorough description of the PBPK models in its analysis of the health risks associated with potential perchlorate exposures (EPA 2002b). The models were later reviewed by an external peer-review panel (EPA 2002c). In this appendix, the committee reviews only the general properties of the PBPK models developed by DOD, their underlying assumptions, and their general applicability in animal-to-human extrapola- tions. GENERAL APPROACHES TO PERCHLORATE PBPK MODEL DEVELOPMENT Perchlorate does not appear to be metabolized in the body. Once absorbed, it is rapidly distributed into all tissues except fat, with preferential uptake by tissues that contain the NIS, and ultimately cleared unchanged in urine. Thus, the development of the PBPK models for adult rats and hu- mans and potentially sensitive life stages (developing fetus and neonate) followed a logical progression of increasing complexity that linked expo- sure with key biochemical events as the toxicity and mode of action of perchlorate in animals and humans became better defined and methods of analyzing perchlorate in biologic fluids and tissues improved.

Appendix E 221 In deriving the RfD, EPA relied on the initial published models and a series of memoranda from DOD describing the continued updating of the suite of PBPK models (Clewell 2001a,b; Merrill 2000; Merrill 2001a,b). The final PBPK models were published shortly after the EPA review (Clewell et al. 2003a,b; Merrill et al. 2003) and reflected comments re- ceived by EPA's peer review (EPA 2002c) and journal peer reviewers. The models have similar base structures and sources for model parameter values and additional features where necessary to address specific life stages, such as growth and development and additional compartments associated with pregnancy and lactation in the rat. The first PBPK model of Fisher et al. (2000) was developed to describe the disposition of perchlorate in the adult male rat; it was based on rather sparse kinetic data that were available in the late 1990s. Tissues that were specifically included in the initial framework included lungs, kidneys, thyroid, and gastrointestinal (GI) tract; the remaining tissues were lumped into either poorly perfused or richly perfused compartments (Figure E-1). Each of those compartments consisted of mass-balance equations describing the rates of transfer of perchlorate into and out of each tissue on the basis of their known volumes, blood perfusion rates, partition coefficients, the presence (or absence) of any transport processes (such as that of the NIS in thyroid and GI tract), or the clearance of perchlorate into urine. Only sparse data on perchlorate after intravenous (IV) administration were available for model development. On the basis of the the preliminary analysis, it was observed that perchlorate clearance into urine after IV dosing at 0.01-3 mg/kg followed linear, first-order kinetics, whereas substantial nonlinear- ities were observed in systemic (serum and tissue) kinetics. Thus, the preliminary model was used as an initial framework for identifying critical data gaps for later model development and definitions of relevant internal- dose surrogates (as opposed to administered doses or external exposures) that could be used to enhance human health risk assessments that were based on toxicity studies in rats. Clewell et al. (2001) later proposed a suite of initial PBPK models that incorporated important steps in the mode of action of perchlorate (NIS iodide uptake inhibition) and key life stages for perchlorate toxicity (fetus and neonate) in addition to the adult male. To accomplish that, initial PBPK models were developed for perchlorate and iodide with interactions between them occurring at the NIS. The purpose of the preliminary PBPK models was to lay the groundwork for extrapolating the internal doses of perchlorate (such as blood concentrations and interactions with iodide in thyroid NIS) from rats to humans as a function of life stage (fetus, neonate,

222 Health Implications of Perchlorate Ingestion IV Lung Kidney Urine Ku Poorly Perfused Richly Perfused Kg GI Tract Vmaxg Kmg GI Blood Thyroid Lumen Vmaxt Kt Kmt Thyroid Blood FIGURE E-1 Diagram of preliminary PBPK model for perchlorate in adult male rat. Abbreviations: GI, gastrointestinal; IV, intravenous; Kg, rate constant for loss of perchlorate from GI tract; Kmg, rate constant for uptake of perchlorate in GI tract; Kmt, rate constant for uptake of perchlorate in thyroid; Kt, rate constant for loss of perchlorate in thyroid; Ku, rate constant for urinary excretion of perchlorate; Vmaxg, maximal capacity for NIS transport in GI tract; Vmaxt, maximum capacity for NIS transport in thyroid. Source: Adapted from Fisher et al. 2000. and adult), dose, and route of exposure with the parallelogram approach discussed in Chapter 4. The structures of the PBPK models across life stages shared physiologic compartments and biologic and chemical-specific sources of parameter

Appendix E 223 values. Each model contained specific descriptions of tissues involved in the uptake, distribution, and elimination of perchlorate and iodide and specific descriptions of the interactions between perchlorate and iodide via the NIS. Basic model structures thus included NIS-containing tissues, such as the thyroid, GI tract, and skin. The thyroid was divided into three - subcompartments representing the tissue capillary bed (referred to as the stroma in the perchlorate PBPK literature), follicle, and lumen (colloid), as described in more detail below. The GI tract was similarly divided into three subcompartments: the capillary bed, GI tissue, and GI contents; perchlorate and iodide are transported by the NIS from the tissue into the contents against a concentration gradient and diffuse between subcompart- ments on the basis of partition coefficients and electrochemical gradients. The skin was divided into two subcompartments: skin blood and skin tissue; perchlorate and iodide are transported from blood to tissue by the NIS and diffuse back and forth between the subcompartments. To model pregnancy and lactation, NIS-containing tissues, such as the placenta and mammary glands, were included. Those tissues were also divided into subcompart- ments as described below for each model. Other NIS-containing tis- sues—such as ovaries, choroids plexus, and salivary glands—were consid- ered too small to affect the concentrations of perchlorate and iodide in plasma or the thyroid and were therefore not specifically segregated from the poorly or richly perfused tissue groups. Remaining tissues in each model included the kidneys for clearance of perchlorate and iodide in urine, the liver for possible future modeling of hormone metabolism, a separate plasma and red-cell compartment for the distribution of perchlorate (plasma-protein-binding) and iodide, and fat because it is an exclusionary compartment (owing to the poor solubility of perchlorate and iodide anions) that is highly variable and changes during pregnancy and lactation. The lung compartment used in the preliminary model of Fisher et al. (2000) was combined with the richly perfused com- partment because neither perchlorate nor iodide is volatile; vapor inhalation was considered to be an irrelevant route of exposure, and neither anion is eliminated to any important degree in breath. Remaining tissues were grouped together on the basis of their similar kinetic properties and blood perfusion rates (as poorly vs richly perfused tissues). As the authors of the PBPK models discussed, it is important to note that the iodide PBPK models used for each life stage (adult, pregnant, lactating, fetus, or neonate) were rudimentary in that they described key physiologic and biochemical processes only in enough detail to reproduce the sparse radiolabeled-iodide kinetic data in animals and humans and the interactions with perchlorate at the NIS. With one exception, discussed

224 Health Implications of Perchlorate Ingestion below, the models did not include dietary or endogenous iodide kinetics, which could be important in interpreting human biomonitoring or clinical studies. And the models did not include formal descriptions of the pharma- codynamics of thyroid hormone control and the HPT biofeedback effects of alterations in iodide concentrations resulting from perchlorate exposure (although NIS up-regulation was modeled to fit the iodide kinetic data after multiple dosing kinetic studies with perchlorate, as discussed below). Such a combined PBPK-pharmacodynamic model would eventually be useful in predicting biologic responses (alterations in thyroid hormone homeostasis) to various perchlorate exposures or different life stages, as noted in EPA (2002c), and in providing a more quantitative basis for understanding species differences in potential toxicity. MODELING KEY EVENTS IN NIS-MEDIATED THYROID RESPONSES TO PERCHLORATE As discussed in Chapter 5, the committee considers the inhibition of iodide uptake at the NIS in the thyroid as one of the key biochemical events preceding the development of adverse responses, such as hypothyroidism, which may result in abnormal growth and development of the fetus and child or in metabolic sequelae at any age. It is important to emphasize that the inhibition of thyroid iodide uptake is not in itself an adverse response. However, because the inhibition of iodide uptake is the first definitive biochemical event that must occur before the chain of events leading to altered hormone homeostasis and then an adverse response, it is a critical component in establishing a relationship between exposure and an internal dose-response for perchlorate risk assessments. Thus, the PBPK models used by EPA in the derivation of an RfD for perchlorate focused on ade- quately describing the interaction between perchlorate, iodide, and the NIS in the thyroid gland even though the internal-dose metric used by EPA in calculating human equivalent exposures was limited to the blood perchlor- ate concentration. The working hypothesis underlying the development of the PBPK models for perchlorate and iodide across life stages of concern was that perchlorate competitively inhibits the thyroid uptake of iodide via the NIS and reduces the iodide available for the production of thyroid hormones (Figure E-2). The inhibition of iodide uptake results in decreases in the production of triiodothyronine (T3) and thyroxine (T4), which lead to an increase in thyroid-stimulating hormone (TSH) production that then stimu-

Appendix E 225 Hypothalamus TRH Pituitary Gland Increase TSH synthesis Increase of NIS synthesis of thyroid peroxidase Thyroid Gland organification NIS I- less iodide T 4 & T3 Decreased I- + T4 & T 3 Competitive Inhibition ClO4- ClO4- FIGURE E-2 Working hypothesis on mode of action of perchlorate on thyroid gland. Abbreviations: ClO4-, perchlorate ion; I-, iodide; NIS, sodium-iodide sym- porter; T3, triiodothyronine; T4, thyroxine; TRH, thyrotropin-releasing hormone; TSH, thyroid-stimulating hormone. lates an increase in the synthesis of NIS proteins and enzymes associated with hormone production. As reviewed in Chapter 4, there are important species differences in the distribution and disposition of thyroid hormones. In particular, the reduced protein binding of T3 and T4 in rat serum (rats lack a thyroxine-binding globulin that is normally present in humans) results in a greater rate of clearance of thyroid hormones from the body than in humans. That causes a compensatory increase in TSH release from the pituitary gland and a corresponding increase in the overall production rate of T3 and T4, making rats particularly sensitive to perchlorate’s disruptions in iodide uptake.

226 Health Implications of Perchlorate Ingestion During the first 10-12 weeks of gestation, the fetus depends on the mother for thyroid hormone production, so the fetus is affected by both alterations in the mother's hormone production and direct inhibition of iodide uptake by perchlorate that has transferred across the placenta. One of the major goals in the development of the PBPK models of Clewell et al. (2003a,b), Merrill (2001), and Merrill et al. (2003) was to simulate interac- tions between perchlorate and iodide at the NIS in the thyroid of rats and humans at key life stages (adult, pregnancy, lactation, fetus, and neonate) and thereby to facilitate high-to-low dose, route-to-route, and cross-species extrapolations based on an internal dose of perchlorate that is relevant to toxicity in both cancer and noncancer risk assessments. To accomplish that goal, a PBPK model for iodide was also developed that interfaced with the perchlorate PBPK model at the level of the NIS in the thyroid gland, GI tract, and skin. A fundamental component of the assumed competitive interaction between perchlorate and iodide is transport of perchlorate itself by the NIS. There has been some debate regarding that issue (Wolf 1998; Reidel et al. 2001a,b; EPA 2002c), but it has not been possible to measure perchlorate uptake directly in rat and human thyroid follicular cells, because of a lack of commercially available radiolabeled perchlorate or an analytic method that is sensitive enough. Thus, arguments have largely centered on the use and interpretation of indirect estimates of perchlorate transport (such as electrochemical gradient measurements vs measurement of radiolabeled uptake of chemicals similar to perchlorate in cells that contain the NIS). Until data become available to address the question directly, the committee agrees that the use of a competitive-inhibition model and its underlying assumption of perchlorate transport by the NIS, as reviewed by Clewell et al. (2004), is a reasonable approach and provides an adequate description of the available data. To model the interactions between perchlorate and iodide, Clewell et al. (2003a,b), Merrill (2001a,b), and Merrill et al. (2003) reduced the com- plexity in potential biologic interactions and maintenance of thyroid hor- mone homeostasis to a series of rate-limiting steps in a physiologically based compartment specific to each species or life stage (adult human, adult male rat, pregnant female rat and fetus, and lactating female rat and neo- nate). In their simplified models, the thyroid was divided into three com- partments representing the capillary bed (stroma), follicle cells, and colloid (lumen) after less complex descriptions (two compartments) failed to simulate the initial rapid phase of thyroid perchlorate uptake and equilib- rium and a slower phase of equilibrium and clearance observed in animal studies (Figure E-3).

Appendix E 227 Colloid Passive Diffusion { } Active Transport Thyroid Follicle Passive Diffusion { } Active Transport (NIS) Capillary Bed Bound ClO4- I- Iodide (Stroma) Competitive Inhibition Venous Arterial Blood Blood FIGURE E-3 Diagram of three-compartment working model for perchlorate and iodide uptake in the thyroid. Abbreviations: ClO4-, perchlorate ion; I-, iodide; NIS, sodium-iodide symporter. Source: Adapted from Clewell et al. 2004. In the simplified model, both perchlorate and iodide were allowed to partition from arterial blood into the thyroid capillary bed on the basis of species-specific and life-stage-specific blood flow rates, tissue:blood partition coefficients, and tissue volumes. The physiologic model thus consisted of a series of mass-balance equations that described the overall rate of change in the amount of perchlorate or iodide in each tissue or tissue compartment as a function of the rate of chemical input less the rate of chemical output: Rate of Change Rate of Input Rate of Output of Chemical = of Chemical ! of Chemical (1) in Tissue “i” in Tissue “i” in Tissue “i”. For the thyroid capillary bed compartment—designated as “stroma” in Clewell et al. (2003a,b), Merrill (2001a,b), and Merrill et al. (2003) model equations described below—perchlorate and iodide that enter from the arterial blood can either passively diffuse or be actively transported by the NIS into the thyroid follicle and passively diffuse back from the follicle or partition back into venous blood draining the thyroid. Thus, the following mass-balance equation was used by Clewell et al. (2003a,b), Merrill (2001a,b), and Merrill et al. (2003) to represent the rate of change in the amount of perchlorate in the thyroid stroma (RATSP) according to

228 Health Implications of Perchlorate Ingestion CTFP RATS P = QT (CAP − CVTS P ) + PATFP ( − CTS P ) − RupTF P PTFP , (2) partitioning passive diffusion NIS transport with blood follicle to stroma stroma to follicle where QT is the thyroid blood flow, CAP is the concentration of perchlorate in arterial blood, CVTSP is the concentration of perchlorate in venous blood draining the thyroid stroma, PATFP is the permeability:area cross product for the diffusion of perchlorate across the membrane between the stroma and follicle, CTFP is concentration of perchlorate in thyroid follicle, PTFP is the thyroid stroma:follicle partition coefficient for perchlorate, CTSP is concentration of perchlorate in thyroid stroma, and RupTFP is the rate of active transport of perchlorate by the NIS from stroma to follicle. Identical equations were used to describe the rate of change in the concentration of iodide in the thyroid stroma with iodide-specific chemical parameters, such as the concentration of iodide in arterial blood, thyroid stroma, and thyroid follicle; permeability-area cross products for the diffusion of iodide across the thyroid follicle membrane; the thyroid stroma:follicle partition coeffi- cient for iodide; and the rate constants associated with the active transport of iodide by NIS. The rate of uptake of perchlorate into the thyroid follicle (RATFP) was then described by CTFP CTL P RATFP = RupTF P + PATFP (CTS P − ) − RupTL P + PATL P ( − CTFP ) PTFP PTL P , (3) NIS passive diffusion active passive diffusion transport stroma to follicle transport follicle to colloid stroma to follicle follicle to colloid where RupTLP is the rate of active transport of perchlorate from follicle into colloid (lumen), PATLP is the permeability:area cross product for the diffusion of perchlorate across the membrane between follicle and colloid, CTLP is the concentration of perchlorate in thyroid colloid, PTLP is the thyroid follicle:colloid partition coefficient for perchlorate, and CTFP is the concentration of perchlorate in thyroid follicle. For iodide, the same equa- tion (Equation 3) was used with the addition of terms describing the loss of inorganic iodide due to the production of thyroid hormones (see “Bound

Appendix E 229 Iodide,” Figure E-3) as a simple, first-order reaction lumping total hormone production: (Equation 3) ! (CLProdI)(CTFI), (4) where CLProdI is the first-order rate of organification of iodide and CTFI is the concentration of inorganic iodide in the thyroid follicle. The rate of change in the amount of perchlorate in the thyroid colloid (lumen) then becomes CTLP RATLP = RupTLP + PATLP (CTFP − ) , (5) PTLP active passive diffusion transport follicle to colloid follicle to colloid The active transport of perchlorate from the thyroid stroma to the follicle (RupTFP) and from the follicle to the lumen (RupTLP) was described by using Michaelis-Menten equations as the basis of saturable transport pro- cesses according to VmaxTFP * CTS P RupTFP = (6) K mTP + CTS P and Vmax TL P * CTF P (7) RupTL P = , K mTL P + CTF P where VmaxTFP and VmaxTLP are the maximal rates of transport from stroma to follicle via the NIS and follicle to lumen via other transporters (such as pendrin or apical iodide channels), respectively, and KmTP and KmTLP are the Michaelis constants for the transport of perchlorate from stroma to follicle and follicle to lumen, respectively. For the integration between the perchlorate and iodide models, the basic Michaelis-Menten equations for NIS transport (stroma to follicle) and active transport in the apical membrane (follicle to colloid) were modified by Clewell et al. (2003a,b), Merrill (2001a,b), and Merrill et al. (2003) to

230 Health Implications of Perchlorate Ingestion incorporate the competitive inhibition of iodide uptake by perchlorate, as follows: VmaxTFI * CTS I RupTFI = , CTS P KmTFI (1 + ) + CTS I (8) KmTFP competitive inhibition with ClO4! and VmaxTLI * CTFI RupTLI = , (9) CTFP KmTLI (1 + ) + CTFI KmTLP competitive inhibition with ClO4! where the rates of transport of perchlorate and iodide by the NIS and apical membrane transporter are mediated by the concentrations of each chemical in the stroma (CTSP and CTSI) or follicle (CTFP and CTFI) and the respec- tive Michaelis constants (KmTFP, KmTFI, KmTLP, and KmTLI). Because of the low Michaelis constants for NIS-based active transport of perchlorate (KmTFP = 1-1.8 × 105 ng/L) versus iodide (KmTFI = 4.0 × 106 ng/L) and for apical membrane transport of perchlorate (KmTLP = 1 × 108) versus iodide (KmTLI = 1 × 109), iodide has little effect on perchlorate kinetics. Thus, Merrill (2001a,b), Merrill et al. (2003), and Clewell et al. (2003a,b) did not include the corresponding competitive inhibition of perchlorate transport by iodide in their perchlorate models. Although the committee considers the inclusion of the competitive inhibition of thyroid perchlorate uptake by iodide as being more consistent with the underlying assumption of competi- tive inhibition, it agrees with the authors that this simplification has little or no effect on the PBPK simulations. Thus, the biochemical parameters controlling the uptake of perchlorate into the thyroid gland and within the three main regions (capillary bed or stroma, follicle, and colloid or lumen) include the tissue:blood partition coefficient; permeability-area cross products for passive diffusion between thyroid regions; the NIS between capillary bed and follicle, where perchlor- ate and iodide compete for transport; and a second apical membrane active

Appendix E 231 transport process between follicle and colloid regions, where perchlorate also competitively inhibits iodide uptake. ADULT RAT PBPK MODEL Merrill et al. (2003) extended the preliminary PBPK model of Fisher et al. (2000) to provide a more detailed description of the disposition of perchlorate and its interactions with iodide in adult male rats. The model included oral gavage, IV injection, and drinking-water routes of exposure; target tissues (thyroid); tissues important to the distribution of perchlorate (GI tract, liver, kidney, skin, fat, blood, and remaining richly and poorly perfused tissues as lumped compartments); and elimination of perchlorate in urine, as shown in Figure E-4. The anatomic and physiologic structure of the iodide PBPK model is identical with that of the perchlorate model except for the blood and thyroid compartments. In the thyroid, inorganic iodide is used in the production of thyroid hormones (separate compartment representing total thyroid hor- mone-base iodide designated as “bound iodide” in the model structure [Figure E-4]), whereas perchlorate is nonreactive and, like inorganic iodide, can passively diffuse from the thyroid lumen and follicles back into the thyroid stroma, where it can partition into venous blood draining the thyroid tissues. This “bound iodide” compartment was necessary to describe the radioiodide kinetic studies (which do not differentiate the form of the radiolabel) because both free (inorganic) and bound (organic) iodide is taken up by tissues. Thus, the bound iodide was pooled with the inorganic iodide in the blood compartment for distribution of total iodide in the body. That simplification assumes that radioiodide behaved as inorganic iodide in all compartments other than the thyroid, including serum. In the thyroid, first-order rate constants were used to describe the formation of bound iodide and its later release into systemic circulation. An additional difference in the structure of the blood compartment was the need to explicitly include a description of protein binding of perchlorate in plasma but not iodide (iodide in blood already represented inorganic and organic iodide). Michaelis-Menten equations were used in the blood compartment (similar to the format used in Equations 6 and 7) to describe the association between free perchlorate and an unspecified plasma-protein (presumably albumin) binding site. Although such a simple description has been used to describe plasma-protein binding for many other chemicals, Merrill et al. (2003) also included a first-order rate constant for the dissocia- tion of perchlorate from the binding protein(s). It is not clear to the com- mittee why the extra parameter was needed in the model.

232 Health Implications of Perchlorate Ingestion Perchlorate Iodide IV Bound ClO4- IV Plasma Plasma Free ClO4- RBCs RBCs Oral GI Contents Oral GI Contents GI Tissue GI Tissue Capillary Bed Capillary Bed Liver Liver Richly Richly Perfused Perfused Kidney Kidney Urine Urine Colloid Colloid Thyroid Bound Thyroid Follicle Iodide Follicle Capillary Bed Capillary Bed Skin Skin Capillary Capillary Bed Bed Fat Fat Poorly Poorly Perfused Perfused FIGURE E-4 Diagram of PBPK models of perchlorate and iodide in adult rat. Bold arrows indicate active transport of perchlorate and iodide NIS in thyroid, skin, and gastric mucosa and by apical iodide channels in thyroid. Abbreviations: ClO4!, perchlorate ion; GI, gastrointestinal; IV, intravenous; NIS, sodium-iodide sym- porter; PBPK, physiologically based pharmacokinetic; RBCs, red blood cells. Source: Adapted from Merrill et al. 2003.

Appendix E 233 Both Merrill et al. (2003) and EPA (2002b) provided a detailed descrip- tion of the sources for each of the physiologic and chemical-specific model parameter values and the ability of the model to describe a variety of datasets. The basis of those model parameter values has been extensively peer-reviewed, so only selected parameters and simulation issues will be highlighted here. For instance, increases in thyroid perchlorate were ob- served in 2-week drinking-water exposure studies conducted by Yu et al. (2002) and Merrill et al. (2003) as a function of dose and duration of expo- sure (for example, at higher dose levels, thyroid perchlorate concentrations were significantly greater after 2 weeks of exposure than after a single dose). In the original PBPK model version summarized by EPA (2002c), such repeated-exposure (vs acute) studies were simulated by increasing the effective perchlorate tissue-partition coefficient between thyroid follicle and thyroid stroma from 0.13 at the lower drinking-water dose levels (0.01, 0.1, and 1.0 mg/kg per day) to 0.4, 1.25, and 2.0 at 3, 10, and 30 mg/kg per day, respectively. There is little biologic reason to justify a change in tissue partitioning as a function of perchlorate dose just to fit the observed data. Therefore, in the final publication, Merrill et al. (2003) presented a more biologically consistent alteration in the model to describe the nonlinear behavior in perchlorate kinetics as a function of dose and duration of exposure. The alteration consisted of increasing the maximum capacity for NIS transport, which is known to be up-regulated by increased TSH, rather than changing the tissue partitioning. No direct measures of the maximum capacity for NIS transport were available, so the maximum capacity for NIS transport in the thyroid (the only tissue where the NIS is up-regulated by TSH) was increased until predicted thyroid perchlorate and iodide corresponded with values mea- sured for each dose used in the drinking-water studies. The resulting maximum capacity for NIS transport was plotted against the corresponding free (unbound) serum perchlorate predicted by the model at each dose (although the authors did not specify the timing of the serum concentra- tions), and a Michaelis-Menten equation was derived from the resulting plot. The equation was then used as a surrogate of up-regulation of thyroid NIS transport as a function of free perchlorate in serum to simulate re- peated-dosing, drinking-water exposure scenarios. Such an adjustment of the model worked well for describing the kinet- ics of perchlorate in the thyroid when up-regulation reached a steady state but does not describe the time lag that occurs during the transition from basal to an up-regulated state. In essence, the PBPK model used either a basal transport level or an up-regulated level in its simulations without a transition phase. The lag for up-regulation could be an important difference

234 Health Implications of Perchlorate Ingestion between rats and humans that may warrant future studies because of the greater hormone production rates and low storage capacity in rats. Never- theless, the committee considers the approach used by Merrill et al. (2003) and later in the other PBPK models that simulated both acute and repeated dosing kinetics as a reasonable, biologically consistent simplification for describing the up-regulation of NIS transport in the thyroid. Furthermore, such up-regulation has little effect on serum concentrations of perchlorate, which were used as a basis of human equivalent exposure calculations in the derivation of the RfD for perchlorate. Merrill et al. (2003) performed a sensitivity analysis to determine which model parameters had the greatest effect on the simulations of serum perchlorate concentrations (measured as area under the curve) at drinking- water concentrations above (10 mg/kg per day) and below (0.1 mg/kg per day) saturation of the NIS. The results of the analysis indicated that chemi- cal-specific parameters associated with plasma-protein binding of perchlor- ate, uptake of perchlorate in the skin, and urinary clearance of perchlorate had the greatest influence on serum perchlorate simulations. Thus, it is particularly important to have independent data on those parameters. Merrill et al. (2003) optimized the perchlorate plasma-protein binding parameter values from in vivo kinetic data, such as Yu et al. (2002). In vitro equilibrium dialysis studies in rat plasma were available to confirm the parameter values. For urinary clearance of perchlorate, Merrill et al. (2003) likewise optimized a first-order clearance parameter value from available time-course kinetic data. The first-order clearance value, 0.07 L/hr per kilogram of body weight, is about 25% of the glomerular filtration rate for a male Sprague-Dawley rat, which supports the need to include protein binding in the model in that perchlorate should be readily filtered by the glomeruli. Given the importance of urinary clearance for the model to simulate serum perchlorate concentrations, and thus inhibition of thyroid iodide uptake, independent studies should be conducted to improve the estimation of urinary clearance of perchlorate (and iodide) in future models. Owing to its large size (19% of the body weight of the rat) and the presence of the NIS, the skin compartment was also identified in sensitivity analyses as an important tissue that affects the kinetics of both perchlorate and iodide. Yu et al. (2002) and Merrill et al. (2003) included skin in their pharmacokinetic studies with rats, but there appeared to be considerable variability in data in rats and a paucity of data in humans regarding the localization of perchlorate and iodide in skin, as discussed below. How- ever, the skin compartment does not represent as great a percentage of the

Appendix E 235 body weight of humans as of rats (19% vs about 3.7%). Thus, it may be more important to develop a thorough understanding of the kinetics of the two anions in rat skin than in humans. Although occasional datasets were not particularly well described, the committee agrees with EPA that the male rat PBPK model of Merrill et al. (2003) provided a reasonable description of a variety of datasets in rats after IV, subcutaneous, or intraperitoneal (IP) injection of drinking-water expo- sure to perchlorate and oral gavage or IV and IP injection of iodide. To facilitate comparisons among the final PBPK models, the final physiologic and biochemical values for the adult rat PBPK model and the other PBPK models of the pregnant rat and fetus, lactating rat and neonate, and adult human are summarized together in Table E-1. ADULT HUMAN PBPK MODEL EPA relied on a memorandum from Merrill (2001b) as the basis of human PBPK-model simulations for perchlorate and iodide. The structure of the adult human PBPK model was nearly identical with that of the male rat model of Merrill at al. (2003) from which it was derived (see Figure E-4) and will therefore not be reiterated here. Male and female human physiol- ogy constants were taken from data available in the literature and are summarized in Table E-1. Chemical-specific parameter values were devel- oped in a way analogous to that for the rat model; this often involved fitting of model parameters to available kinetic data on perchlorate and iodide at various dose levels. The parameter values are also summarized in Table E-1. EPA provided a thorough review of the sources of the model parame- ters; only the ones that are different from those for adult rat are highlighted in this appendix. For instance, slightly different partition coefficients and permeabil- ity-area cross products governing the diffusion of perchlorate and iodide were used for humans in tissues containing the NIS (stomach, skin, GI tract, and thyroid). Such differences are considered minor and probably reflected attempts to refine the human model on the basis of available kinetic data, inasmuch as the human model was being developed at the same time as the adult rat model. Similarly, several parameters governing the transport of perchlorate and iodide by the NIS were also adjusted from the rat model to improve the fits to the available human data. The binding affinity for the NIS appears to be similar across tissues and species, so values for each NIS-

TABLE E-1 Values of Physiologic and Biochemical Parameters for Disposition of Perchlorate and Iodide in Thyroid 236 Compartments of Rat and Human PBPK Models of Clewell (2003a,b), Merrill (2001b), and Merrill et al. (2003) Pregnancy (Rat, GD 0-21)a Lactation (Rat, PND 0-18) Adult Parameter Male Rat Dam Fetus Dam Neonate Human Physiologic Parameters for Perchlorate and Iodide Thyroid Compartments Body weight (kg) 0.3 0.28-0.361 0-0.0045 0.277-0.310 0.0075-0.1985 ~70.0 Tissues volumes Slowly perfused (% BW) 74.6 74.6 74.6 37.07-40.42 53.92-49.31 65.1 Rapidly perfused (% BW) 11.0 11.0 16 5.35 5.36 12.4 Fat (% BW) 7.4 10.0-11.0 0 12.45-6.9 0-4.61 21.0(M) 32.7(F) Kidney (% BW) 1.7 1.7 0.3-0.44 1.7 1.7 0.44 Liver (% BW) 5.5 3.4 8.5-7.2 3.4 3.4 2.6 GI tissue (% BW; VGI) 0.54b 3.6 2.0-3.0 3.9 3.9 1.7b GI contents (% BW) 1.68b 7.2 0.8-6.2 7.2 7.2 0.071b GI blood (% VGI) 4.1b 2.9 2.9 2.9 2.9 4.1b Skin Tissue (% BW; VSk) 19.0 19.0 8.8-19.3 19.0 19.0 3.7 Skin blood (% VSk) 2.0 2.0 2.0 2.0 2.0 8.0 Plasma (% BW) 4.1 4.7 4.7 4.7 4.7 4.4 Red blood cells (% BW) 3.3 2.74 2.74 2.74 2.74 3.5 Placenta (% BW) — 0-2.57 — — — — Mammary gland (% BW) — 1.0-5.5 — 4.4-6.6 — — Mammary blood (% mammary) — — — 18.1 — — Milk (L) — — — 0.002 — — Thyroid total (% BW) 0.0077 0.0105 0.058-0.038 0.0105 0.0125 0.03 Thyroid follicle (% thyroid) 59.9 45.9 61.4 45.89 61.4-37.2 57.3

Thyroid colloid (% thyroid) 24.4 45.0 18.3 45.0 18.3-32.5 15.0 Thyroid blood (% thyroid) 15.7 9.1 20.3 9.1 20.3-30.3 27.6 Cardiac output (L/hr per kg) 14.0 14.0 67.8 14.0-21.0 14.0 16.5 Blood flows (% cardiac output) Poorly perfused 24.0 24.0 24.0 7.9-1.9 16.9 13.0 Richly perfused 76.0 76.0 76.0 40.8 40.8 33.0 Fat 6.9 7-8.1 — 7.0 7.0 5.2 Kidney 14.0 14.0 3.6 14.0 14.0 17.5 Liver 17.0 18.0 4.5 18.0 18.0 22.0 GI 1.61b 13.6 4.6 1.61 1.61 1.0b Skin 5.8 5.8 10.4 0.058 0.058 5.8c Placenta — 0.0-12.3 — — — — Mammary — 0.2-1.2 — 9.0-15.0 — — Thyroid 1.6 1.6 1.6 1.6 1.6 1.6 Perchlorate-Specific Parameters Partition coefficients (unitless) Slowly perfused:plasma 0.31 0.31 0.31 0.31 0.31 0.31 Rapidly perfused:plasma 0.56 0.56 0.56 0.50 0.50 0.56 Fat:plasma 0.05 0.05 — 0.05 0.05 0.05 Kidney:plasma 0.99 0.99 0.99 0.99 0.99 0.99 Liver:plasma 0.56 0.56 0.56 0.56 0.56 0.56 Gastric tissue:gastric blood 0.70b 0.50 1.80 1.80 3.21 1.80b Gastric contents:gastric tissue 1.70b 1.30 2.30 2.30 5.64 2.30b Skin tissue:skin blood 1.00 1.15 1.15 1.15 1.15 1.15 Red blood cells:plasma 0.73 0.73 0.73 0.73 0.73 0.80 Placenta:plasma — 0.56 — — — — 237

TABLE E-1 (Continued) 238 Pregnancy (Rat, GD 0-21)a Lactation (Rat, PND 0-18) Adult Parameter Male Rat Dam Fetus Dam Neonate Human Mammary gland:plasma — 0.66 — — — — Mammary tissue:mammary blood — — — 0.66 — — Mammary tissue:milk — — — 2.39 — — Thyroid follicle:stroma 0.15 0.15 0.15 0.13 0.13 0.13 Thyroid lumen:follicle 8.00 7.00 7.00 7.00 7.00 7.00 Permeability area cross products (L/hr per kg) Gastric blood-gastric tissue 1.00b 1.00 1.00 1.00 1.00 0.60b Perchlorate-Specific Parameters Gastric tissue-gastric contents 0.80b 1.00 1.00 1.00 1.00 0.80b Skin blood-skin tissue 0.80 1.00 1.00 0.50 1.00 1.00 Plasma-red blood cells 1.00 1.00 1.00 1.00 1.00 1.00 Mammary blood-mammary — 0.04 — 0.01 — — tissue Mammary tissue-milk — — — 0.10 — — Placenta blood-placenta tissue — 0.10 — — — — Thyroid follicle-stroma 6.0 × 10-5 6.0 × 10-5 6.0 × 10-5 4.0 × 10-5 4.0 × 10-5 1.0 × 10-4 Thyroid lumen-follicle 0.01 0.01 0.01 0.01 0.01 0.01 Thyroid: NIS active transport Km (ng/L) 1.8 × 105 1.0 × 105 1.0 × 105 1.5 × 105 1.5 × 105 1.8 × 105 Vmax (ng/hr per kg)d 1.0 × 103 2.6 × 103 0-2.25 × 103 1.5 × 103 1.5 × 103 5.0 × 104 Thyroid:apical membrane transport Km (ng/L) 1.0 × 108 1.0 × 108 1.0 × 108 1.0 × 108 1.0 × 108 1.0 × 108

VmaxC (ng/hr per kg) 2.0 × 104 1.0 × 104 1.0 × 104 1.0 × 104 1.0 × 104 2.5 × 105 Gastric: NIS active transport Km (ng/L) 1.7 × 105 b 1.0 × 105 1.0 × 105 1.5 × 105 — 2.0 × 105 b VmaxC (ng/hr per kg) 2.0 × 104 b 8.0 × 105 1.0 × 105 1.0 × 106 1.0 × 106 1.0 × 105 b Perchlorate-Specific Parameters Skin:NIS active transport Km (ng/L) 1.8 × 105 1.0 × 105 1.0 × 105 1.5 × 105 1.5 × 105 2.0 × 105 VmaxC (ng/hr per kg) 5.0 × 105 6.0 × 105 4.0 × 105 8.0 × 105 8.0 × 105 1.0 × 106 Mammary: active transport Km (ng/L) — 1.0 × 105 — 1.5 × 105 — — VmaxC (ng/hr per kg) — 2.2 × 104 — 2.0 × 104 — — Milk: active transport Km (ng/L) — — — 1.0 × 106 — — VmaxC (ng/hr per kg) — — — 2.0 × 104 — — Placenta active transport Km (ng/L) — 1.0 × 105 — — — — VmaxC (ng/hr per kg) — 6.0 × 104 — — — — Transfer placenta to fetus (L/hr per kg) — 0.065 — — — — Transfer fetus to placenta (L/hr per kg) — 0.12 — — — — Plasma-protein binding Km (ng/L) 1.1 × 104 1.0 × 104 1.5 × 104 1.0 × 104 1.0 × 104 1.8 × 104 VmaxC (ng/hr per kg) 3.4 × 103 4.0 × 103 1.5 × 103 9.0 × 103 2.0 × 103 5.0 × 102 Dissociation constant (L/hr per kg) 0.032 0.034 0.01 0.034 0.01 0.025 Urinary clearance (L/hr per kg) 0.07 0.07 — 0.07 0.0075 0.1265 239

TABLE E-1 (Continued) 240 Pregnancy (Rat, GD 0-21)a Lactation (Rat, PND 0-18) Adult Parameter Male Rat Dam Fetus Dam Neonate Human Fraction pup urine ingested by dam — — — 0.80 — — Iodide-Specific Parameters Partition coefficients (unitless) Poorly perfused:plasma 0.21 0.21 0.21 0.21 0.21 0.21 Richly perfused:plasma 0.40 0.40 0.40 0.40 0.40 0.40 Fat:plasma 0.05 0.05 — 0.05 0.05 0.05 Kidney:plasma 1.00 1.09 1.09 1.09 1.09 1.09 Liver:plasma 0.44 0.44 0.44 0.44 0.44 0.44 Gastric tissue:gastric blood 1.00b 1.00 1.00 1.00 1.20 0.50b Gastric contents:gastric tissue 3.50b 2.00 2.00 1.00 1.00 3.50b Skin tissue:skin blood 0.70 0.70 0.70 0.70 1.00 0.70 Red blood cells:plasma 1.00 1.00 1.00 1.00 1.00 1.00 Placenta:plasma — 0.40 — — — — Mammary gland:plasma — 0.66 — — — — Mammary tissue:mammary blood — — — 0.80 — — Mammary tissue:milk — — — 1.00 — — Thyroid follicle:stroma 0.15 0.15 0.15 0.15 0.15 0.15 Thyroid lumen:follicle 8.00 7.00 7.00 7.00 7.00 7.00 Permeability area cross products (L/hr per kg) Gastric blood-gastric tissue 1.00b 0.80 0.10 0.80 0.04 0.20b Gastric tissue-gastric contents 0.10b 0.60 0.30 0.60 0.09 2.00b

Skin blood-skin tissue 0.10 0.10 0.02 0.20 0.02 0.06 Plasma-red blood cells 1.00 1.00 1.00 1.00 1.00 1.00 Placenta blood-placenta tissue — 0.005 — — — — Mammary blood-mammary tissue — 0.01 — 0.02 — — Mammary tissue-milk — — — 0.02 — — Thyroid follicle-stroma 1.0 × 10-4 1.0 × 10-4 1.0 × 10-4 1.0 × 10-4 1.0 × 10-4 1.0 × 10-4 Thyroid lumen-follicle 4.0 × 10-7 4.0 × 10-7 4.0 × 10-4 1.0 × 10-4 1.0 × 10-4 1.0 × 10-4 Thyroid: NIS active transport Km (ng/L) 4.0 × 106 4.0 × 106 4.0 × 106 4.0 × 106 4.0 × 106 4.0 × 106 VmaxC (ng/hr per kg) 5.4 × 106 4.4 × 104 0-5.0 × 104 5.0 × 104 1.3 × 104 1.5 × 105 Thyroid: apical membrane transport Km (ng/L) 1.0 × 109 1.0 × 109 1.0 × 109 1.0 × 109 1.0 × 109 1.0 × 109 VmaxC (ng/hr per kg) 4.0 × 106 4.0 × 106 4.0 × 106 6.0 × 107 6.0 × 107 1.0 × 108 Thyroid:hormone production (L/hr per kg) 0.1 0.03 — 0.1 0.06 — Thyroid: hormone secretion (L/hr per kg) 1.2 × 106 1.0 × 10-6 — 7.0 × 10-7 1.0 × 10-6 — Deiodination of thyroid hormone (L/hr per kg) — — — 0.02 0.025 — Gastric:NIS active transport Km (ng/L) 4.0 × 106 b 4.0 × 106 b 4.0 × 106 b 4.0 × 106 b — 4.0 × 106 b VmaxC (ng/hr per kg) 2.0 × 106 b 1.0 × 106 2.0 × 106 2.0 × 106 2.0 × 106 9.0 × 105 b Skin: NIS active transport Km (ng/L) 4.0 × 106 4.0 × 106 4.0 × 106 4.0 × 106 4.0 × 106 4.0 × 106 VmaxC (ng/hr per kg) 5.0 × 105 6.0 × 104 7.0 × 105 4.0 × 105 2.5 × 105 7.0 × 105 Mammary: active transport Km (ng/L) — 4.0 × 106 — 4.0 × 106 — — VmaxC (ng/hr per kg) — 4.0 × 104 — 8.0 × 105 — — 241

TABLE E-1 (Continued) 242 Pregnancy (Rat, GD 0-21)a Lactation (Rat, PND 0-18) Adult Parameter Male Rat Dam Fetus Dam Neonate Human Milk: active transport Km (ng/L) — — — 1.0 × 107 — — VmaxC (ng/hr per kg) — — — 4.0 × 105 — — Placenta active transport Km (ng/L) — 4.0 × 106 — — — — VmaxC (ng/hr per kg) — 5.5 × 104 — — — — Transfer placenta to fetus (L/hr per kg) — 0.06 — — — — Transfer fetus to placenta (L/hr per kg) — 0.12 — — — — Plasma protein binding Km (ng/L) — — — 1.0 × 105 1.0 × 105 — 2e 3 VmaxC (ng/hr per kg) 1.0 × 10 — — 1.5 × 10 5 × 102 — Dissociation constant (L/hr per kg) — — — 0.09 0.05 — Urinary clearance (L/hr per kg) 0.05 0.03 — 0.06 0.012 0.1 Fraction pup urine ingested by dam — — — 0.8 — — a Pregnancy models covered both embryonic and fetal development up to gestation day 21; tissue samples were taken only from developing fetuses (gestation day 20-21) in pharmacokinetic studies. Earliest day for fetal tissue volume data is gestation day 11, so exponential curves were used to simulate tissue growth during embryonic development. Final equations for fetal growth were adjusted by measured pup body weights at birth (or when samples were collected from kinetic studies). b Male rat and adult human GI refers only to stomach. c Data from Williams and Leggett (1989). d VmaxC is Vmax scaled by body weight according to equation Vmax (mg/hr) = [VmaxC (mg/hr per kg)][body weight (kg)0.70]. e It was unclear why value for capacity of plasma-protein binding of iodide (Vmax) was reported in adult male rat model but protein binding was included only in lactating rat PBPK model.

Abbreviations: BW, body weight; F, female; GD, gestation day; Km, Michaelis-Menten constant; hr, hour; kg, kilogram; L, liter; M, male; NIS, sodium-iodide symporter; ng, nanogram; PND, postnatal day; VGI, volume of gastrointestinal tract; Vmax, maximum capacity for binding or transport; VmaxC, maximum capacity for binding or transport scaled by body weight; VSk, volume of skin. 243

244 Health Implications of Perchlorate Ingestion containing tissue were not substantially different between the various rat and human PBPK models summarized in Table E-1. However, when attempting to simulate the radioiodide-uptake data of Greer et al. (2000, 2002), Merrill (2001b) determined that a nearly 10-fold range in NIS capacity existed between subjects, which is not unusual for many biochemi- cal processes. In fact, the baseline thyroid radioiodide uptake varied by a factor of 3-5 among the volunteers. In addition to potential differences in NIS protein expression between subjects, there are potential environmental influences on the variability of the kinetic data, including differences in endogenous iodide concentrations resulting from differences in diets, the timing of blood collections relative to meals, the presence of other potential inhibitors of NIS, and other factors that were not controlled for. As dis- cussed in Chapter 3 and Appendix D, the differences in absolute capacities of the NIS or differences in basal iodide concentrations between humans by themselves should not have an important effect on intersubject sensitivity to the rates of perchlorate inhibition of iodide uptake. Regardless, varia- tions in thyroid parameter values have little effect on serum concentrations of perchlorate (and iodide) that EPA used in deriving the RfD, because the thyroid is small. Plasma-protein binding is often a source of substantial species differ- ences in chemical disposition, as was evident for perchlorate. On the basis of the serum data of Greer et al. (2000), humans had a lower capacity for binding perchlorate than did rats (for example, the capacity for plasma- protein binding was 500 ng/hr per kilogram in humans vs 3,400 ng/hr per kilogram in rats), whereas the binding affinities and dissociation constants were similar. Likewise, urinary clearance may reflect substantial species differences; in the human PBPK model, the urinary clearances of both anions were about twice as high as in the rat, possibly partially because of decreased plasma-protein binding, at least for perchlorate. The chemical-specific parameters were either scaled from the rat or developed from the iodide kinetic data of Hays and Solomon (1965) and the perchlorate- and iodide-uptake data and inhibition measurements of Greer et al. (2000, 2002). Model validations (simulations of data from independ- ent experiments that were not used in model development) were based on simulations of the urinary clearance of perchlorate in healthy males after oral doses of 9.07-20 mg/kg (Eichen 1929; Durand 1938; Kamm and Drescher 1973), serum perchlorate concentrations after drinking-water exposures at 12 mg/kg per day (unpublished results provided to E. Merrill by Dr. Brabant, Hanover, Germany [Merrill 2001b]), and iodide uptake in the thyroid of a male with Graves disease when the maximum capacity of

Appendix E 245 the NIS was increased about 10 times that in healthy subjects (Stanbury and Wyngaarden 1952). However, the model underpredicted the degree of inhibition of thyroid radioiodide uptake after dosing with 100 mg of potas- sium perchlorate; this suggests that the increased inhibition of iodide uptake in Graves disease may not be simply due to the affinity or capacity of the NIS (no perchlorate or iodide kinetics in serum or urine were evaluated in this study). Overall, the adult male human PBPK model provided a reasonable description of the available human perchlorate and iodide kinetics data over doses spanning several orders of magnitude (0.02-12 mg/kg). As discussed above for the male rat, the human skin compartment may be important in controlling the kinetics of both anions, although to a much smaller extent than in rats because of its smaller fraction of total body weight. Regardless, including a skin compartment was critical to the human model simulations of serum perchlorate concentrations, so the paucity of human data on the disposition of perchlorate and iodide in skin remains a concern for future research. Although no formal sensitivity analysis was performed on the human PBPK model, it is likely that, in addition to the skin compartment, urinary clearance of both anions and the plasma-protein binding of per- chlorate may be important for additional future research. Furthermore, the PBPK model was developed for adult males and females (primarily healthy subjects although one subject with Graves disease was simulated) but not for pregnant or lactating females, human fetuses, neonates, or children. To assist in validating the parallelogram approach used to derive human equiv- alent exposures, discussed in Chapter 4, consideration should be given to refining the iodide PBPK model to incorporate data from biomonitoring studies, such as Soldin et al. (2003) and Hollowell et al. (1998), that include the analysis of iodide in pregnant women. PBPK MODEL FOR PREGNANT RATS AND FETUSES During gestation and early infancy, thyroid hormones are needed for normal development (see Chapter 2). Although the critical periods for effective disruption due to the inhibition of thyroid iodide uptake by per- chlorate are not known, it is likely that the developing fetus becomes directly susceptible to perchlorate when the thyroid begins to sequester iodide and secrete hormones. That occurs by 12 weeks of gestation in humans or around 17-20 days in rats (Clewell et al. 2003a). The fetus is potentially indirectly susceptible earlier if the mother's hormone production

246 Health Implications of Perchlorate Ingestion is compromised. Because perchlorate can competitively inhibit iodide uptake in the maternal thyroid, placental iodide transfer, or fetal thyroid uptake, Clewell et al. (2003a) developed a PBPK model covering the full gestation period in rats. Clewell et al. (2003a) extended the male rat PBPK model to the preg- nant female rat by using pregnant female rat-specific physiology over gestation days 2-20 and adding placental and mammary tissue compart- ments. The maternal model was linked with a fetal model that was similar in structure to the adult model (without the fat, placenta, and mammary tissues), as shown in Figure E-5. To simplify the model, all fetuses from a single litter were lumped together in the model structure (for example, each tissue compartment was multiplied by the number of fetuses in each experi- ment to simulate the available data). That simplification was necessary to simulate the kinetic data in fetal samples that were too small and had to be pooled across a litter for analyses. Although a kidney compartment was included in the fetal model, urinary excretion was not included, because urine production is not well developed until after birth. Pregnancy is a remarkably dynamic process in which changes occurr rapidly in both the pregnant animal and its developing offspring. Those changes have the potential for affecting the delivery of perchlorate (and iodide) to its target site at the appropriate time for effects to occur. In the PBPK models just described for adult rats and humans, the volume of each tissue and its corresponding blood flow were treated as constants. How- ever, during pregnancy, some tissue volumes and blood flow rates change in the mother and the embryo and fetus and must be treated as variables rather than constants in the PBPK model. As a result, Clewell et al. (2003a) based their physiologic model on the work of O'Flaherty et al. (1992) and Fisher et al. (1989), who derived detailed equations for each tissue’s unique and dynamically changing growth rates and blood perfusion rates over the entire gestation period in rats. Those models have been widely used as the structural basis of most rat pregnancy PBPK models developed over the last 10 years (Corley et al. 2003). FIGURE E-5 Diagram of PBPK model for perchlorate and iodide distribution in pregnant rat and developing fetus (adapted from Clewell et al. [2003a]). Bold arrows indicate active transport of perchlorate and iodide by NIS in thyroid, skin, gastric mucosa, mammary gland, and placenta and by apical iodide channels in thyroid. Model compartments that were added to adult rat PBPK model of Merrill et al. (2003) are designated by shading. Abbreviations: ClO4-, perchlorate ion; GI, gastrointestinal; IV, intravenous; NIS, sodium-iodide symporter; PBPK, physiologi- cally based pharmacokinetic; RBCs, red blood cells.

Appendix E 247 Pregnant Female Fetus IV Bound ClO4- Bound ClO4- Plasma Plasma Free ClO4- Free ClO4- RBCs RBCs Oral GI Contents GI Contents GI Tissue GI Tissue Capillary Bed Capillary Bed Liver Liver Richly Richly Perfused Perfused Kidney Kidney Urine Colloid Colloid Bound Thyroid Thyroid Iodide Follicle Follicle Capillary Bed Capillary Bed Skin Skin Capillary Capillary Bed Bed Poorly Fat Perfused Poorly Perfused Mammary Gland Capillary Bed Capillary Bed Placenta

248 Health Implications of Perchlorate Ingestion As with the adult male rat model of Merrill (2001b), up-regulation of maternal thyroid NIS was included as a heuristic approximation of feedback control by the HPT axis because of the inhibition of thyroid iodide uptake by perchlorate (that is, no attempts were made to describe the time-depend- ent up-regulation or other physiologic changes associated with it). Al- though fetal thyroids begin to secrete hormone during gestations days 17-20, organification of thyroid iodide was not included in the fetal thyroid model, as it is in the mother, because of a lack of quantitative data on production and secretion rates. The mammary gland, which also has NIS, was not shown to concentrate perchlorate and iodide over peak serum concentrations during gestation although the concentrations of both anions were observed by Clewell et al. (2003a) to remain higher in mammary tissue than in serum during the clearance phase. Therefore, the mammary gland was divided into two compartments: mammary blood and mammary tissue with both NIS trans- port and passive-diffusion equations included, as for the skin compartment. Likewise, the placenta, which also contains the NIS, was modeled as two compartments. Transport of nutrients and xenobiotics to the embryo and fetus occurs across both the yolk sac and chorioallantoic placentas, which vary substan- tially among species in function and development. For example, iron and proteins are transported across the yolk sac in rats and across the chorio- allantoic placenta in humans. Thus, several PBPK models that describe the transfer of xenobiotics between the mother and the developing embryo and fetus include both types of placenta, which develop at different rates. Because of a lack of appropriate data, Clewell et al. (2003a) combined both types of placenta into one placental compartment with empirical first-order rate constants governing the transfer of perchlorate and iodide between the placenta and the fetal plasma compartments (see Figure E-5). Future research could include a more realistic understanding of the placental transfer of perchlorate and iodide. To determine values of several chemical-specific parameters for their model, Clewell et al. (2003a) conducted a series of experiments, including drinking-water kinetic studies with perchlorate in pregnant Sprague-Dawley rats over gestation days 2-20, radioiodide kinetic studies after IV adminis- trations on gestation day 20, and radioiodide-inhibition studies on gestation day 20. Those studies were used to develop and refine estimates of NIS transport parameter values, thyroid apical membrane transport parameter values, partition coefficients, permeability-area cross products for diffusion, plasma-protein binding (perchlorate only), and urinary clearance. For several tissues, there was a biologic basis for expecting differences in

Appendix E 249 perchlorate and iodide kinetic parameter values in male versus pregnant female rats (such as plasma-protein binding, thyroid transport, and skin transport). For other tissues, model parameter values were the same (or nearly so) between adult male and pregnant female rats. The committee does not see such differences as detracting from the utility of the PBPK models for simulating elements important to the disposition of perchlorate and iodide as a function of life stage. The predictions of the PBPK model were validated with datasets that were not used to develop the model. Several published datasets describing the kinetics of iodide in pregnant rats were reasonably well described by the PBPK model, especially given the rapidly changing dynamics of gestation. The ability of the model to predict the perchlorate-induced inhibition of thyroid iodide uptake in published studies and those conducted by Clewell et al. (2003a) requires accurate simulations of both perchlorate and iodide kinetics. Thus, those datasets served as an additional validation of the perchlorate and iodide PBPK models. Clewell et al. (2003a) also performed a sensitivity analysis to determine which model parameters had the greatest effect on simulations of serum perchlorate concentrations and thyroid iodide uptake. As with the adult male rat PBPK model, maternal serum perchlorate is sensitive to serum- protein binding and urinary-clearance parameters. Fetal serum simulations were sensitive to plasma transfer rates, placental NIS, placental diffusion, fetal serum-protein binding, and maternal urinary-clearance parameters. As discussed above for the male rat, parameters identified by the sensitivity analysis present important research opportunities. The inhibition of thyroid iodide uptake was sensitive to many more parameters. The authors appro- priately concluded that that was because thyroid iodide uptake simulations were for a specific point in time and were thus more sensitive to the chang- ing kinetics of both perchlorate and iodide. PBPK MODEL FOR LACTATING RATS AND NEONATES Neonatal hormone feedback control is independent of the dam, and the thyroid gland continues to develop after birth in the rat. But, the dam still controls both perchlorate and iodide transfer to the offspring via nursing. Therefore, Clewell et al. (2003b) extended their rat gestation PBPK model to include lactation transfer over postnatal days 0-10. As with gestation models, lactation-transfer PBPK models must contend with rapid changes in both maternal and neonatal physiology. There are fewer PBPK models that cover this period of development than there are gestation models

250 Health Implications of Perchlorate Ingestion (Corley et al. 2003). Therefore, Clewell et al. (2003b) relied on the initial framework described by Shelley et al. (1988) and the earliest model that incorporated physiologic changes of the rat dam and neonate by Fisher et al. (1990) as the structural and physiologic basis of the perchlorate and iodide PBPK models. Additional compartments that were necessary to describe perchlorate and iodide kinetics and interactions were either taken from the adult rat and pregnant rat PBPK models described above or modi- fied to specifically describe the data available on lactation transfer of perchlorate and iodide. The latter modifications are described below. During gestation, the NIS in the mammary gland delays the clearance of iodide (and perchlorate) from mammary tissues although these anions typically do not rise above peak serum concentrations. During lactation, however, iodide is concentrated in the milk. Inhibition of iodide uptake by perchlorate not only decreases the iodide available in milk but also transfers perchlorate, which can inhibit the uptake of iodide in the thyroid of the nursing pup (Yu et al. 2001). Thus, the two-compartment model for the mammary gland used in the pregnancy model of Clewell et al. (2003a) was expanded to three compartments: mammary blood, mammary tissue, and milk at parturition (see Figure E-6). Similar to NIS activity in the thyroid, NIS transport of iodide and perchlorate occurs between mammary blood and mammary tissue. A second mechanism reportedly exists to transport iodide against a concentration gradient into milk, so Clewell et al. (2003b) incorporated equations for the competitive transport of perchlorate and iodide (designated by bold arrows in Figure E-6) between compartments with passive diffusion that is driven by concentration gradients between compartments. Perchlorate and iodide are then transferred from the milk to the GI contents of the pup with a first-order rate of transfer, and both anions are returned to the dam (GI contents) via pup urine with a first-order trans- fer rate to simulate grooming by the dam. The remaining maternal tissue compartments, except plasma, were structured similarly to the way they were in the pregnancy model with changing physiology over postnatal days 0-10 based on the lactation model of Fisher et al. (1990). The neonatal model was similar to the maternal model except for the absence of mammary tissues. In contrast with the developing fetus, a fat compartment was needed because this tissue begins to grow after birth. To simplify the model, all pups from a given litter were combined in the model structure (that is, total body weight was multiplied by the number of pups in each litter). Thyroid secretion of hormones was modeled in a fashion similar to that used in the adult rat and pregnancy models with one major exception. In

Appendix E 251 the lactation model, Clewell et al. (2003b) added deiodination of thyroid hormones to the plasma compartment as a surrogate of tissue deiodination and included competitive protein binding between iodide and perchlorate in plasma (Figure E-7). It is not clear why that is needed in the lactation model but not in the adult rat, adult human, and pregnant rat models. To develop the lactation-transfer PBPK model, Clewell et al. (2003b) conducted perchlorate drinking-water exposure studies in Sprague-Dawley rats covering the period from gestation day 2 through postnatal day 5 or 10 and radioiodide kinetic and inhibition studies in lactating rats and neonates on postnatal day 10. Kinetic parameter values for the perchlorate model were estimated from the maternal drinking-water study for both dams and neonates. For the iodide model, chemical-specific parameter values were obtained for the dams and neonates that were each dosed independently. Chemical-specific model parameter values (Table E-1) in tissues other than the mammary gland and blood compartment were kept as similar as possible to those of the adult rat and pregnant rat models although some were ad- justed to fit the perchlorate and iodide kinetic studies conducted by Clewell et al. (2003b). Up-regulation of thyroid NIS to simulate the drinking-water kinetic studies was performed as described for the adult male rat PBPK model. Model simulations of neonatal exposure via nursing were assumed to be continuous; in gavage experiments, the dose was introduced as a bolus directly into GI contents. Drinking-water exposure for the perchlorate studies was modeled as 12-hr/day exposure to reflect nocturnal feeding behaviors (6 p.m. to 6 a.m.); simulations of dietary iodide studies, which were reported only for lactating rats and their pups, were based on the assumption of 24-hr/day intake. The reasons for the differences in maternal exposure (12 hr/day vs 24 hr/day) were not explained by the authors. Thus, the radioiodide, dietary iodide, and perchlorate models were operated independently with interactions between the three models occurring at the level of the NIS and, although not explicitly stated by the authors, presum- ably via competitive plasma-protein binding. The resulting models were able to simulate maternal and neonatal perchlorate and iodide kinetics whether only the dams were exposed or the neonates were exposed directly. Validation of the model simulations was based on several datasets from the literature that were not used to estimate model parameter values and on the thyroid iodide uptake inhibition studies. The lactation-transfer model was able to simulate perchlorate and iodide kinetics and inhibition of NIS transport in the thyroid reliably over a nearly complete period of lactation (up to postnatal day 20).

252 Health Implications of Perchlorate Ingestion Lactating Female Neonate IV RBCs RBCs ClO4-, I- Free Free Plasma ClO4-, I- Plasma Bound Plasma Bound Plasma Iodide Bound Bound Iodide Colloid Colloid Thyroid Bound Thyroid Bound Follicle Iodide Follicle Iodide Capillary Bed Capillary Bed Milk GI Contents Mammary Tissue GI Tissue Capillary Bed Capillary Bed Oral GI Contents Liver GI Tissue Pup Urine Richly Capillary Bed Perfused Liver Kidney Skin Richly Perfused Capillary Bed Kidney Fat Urine Skin Poorly Capillary Perfused Bed Fat Poorly Perfused FIGURE E-6 Diagram of PBPK model for perchlorate and iodide in lactating rat and neonate (adapted from Clewell et al. [2003b]). Bold arrows indicate active transport of perchlorate and iodide NIS in thyroid, skin, gastric mucosa, mammary gland, and placenta and by apical iodide channels in thyroid and mammary gland. Model compartments that were altered or added to pregnancy model of Clewell et al. (2003a) are designated by shading. See Figure E-7 for more detailed diagram

Appendix E 253 RBCs Free Plasma De-Iodination (Alb-T3 Alb-I- Alb-T4 Bound fT3, fT4) Organic Inorganic Incorporated Thyroid Hormone Colloid (Alb-T3, Alb-T4, fT3, fT4) Thyroid Follicle Capillary Bed FIGURE E-7 Diagram of thyroid and blood subcompartments from iodide PBPK model of lactating and neonatal rat (adapted from Clewell et al. 2003b). Bold arrows indicate active transport of iodide. Model subcompartments that were altered or added to pregnancy model of Clewell et al. (2003a) are designated by shading. Thyroid hormones (free and bound) were lumped together and secreted by thyroid follicle into plasma where first-order deiodination reaction (a surrogate of tissue metabolism) releases inorganic iodide, which can compete with perchlo- rate for binding to plasma proteins. Abbreviations: Alb, albumin; I!, iodide; fT3, free triiodothyronine; fT4, free thyroxine; T3, triiodothyronine; T4, thyroxine. of thyroid-blood compartments of iodide model outlined in this figure. Abbrevi- ations: ClO4-, perchlorate ion; GI, gastrointestinal; I-, iodide; IV, intravenous; NIS, sodium-iodide symporter; PBPK, physiologically based pharmacokinetic; RBCs, red blood cells.

254 Health Implications of Perchlorate Ingestion A rudimentary dietary and drinking-water exposure iodide PBPK model was developed to simulate "endogenous" iodide from the radioidide studies in the lactating rat and nursing pups, so the authors conducted additional simulations to determine the effect of changing dietary iodide on predicted acute radioiodide kinetics and perchlorate-induced inhibition of dietary uptake. Their simulations were consistent with the committee's calculations for humans discussed in Chapter 3 and Appendix D. Changes in dietary iodide, even at concentrations over at least 100 times those used in standard laboratory diets, should have no effect on the sensitivity of the thyroid to the inhibition of iodide uptake by perchlorate in the lactating rat. As with the other PBPK models, Clewell et al. (2003b) also conducted a sensitivity analysis to determine which model parameters had the greatest effects on the predictions of serum perchlorate and thyroid iodide uptake inhibition. The results were similar to those of previous model sensitivity analyses and included parameters associated with plasma-protein binding, renal clearance, and rates of transfer between dams and nursing pups. COMPARISONS OF INTERNAL-DOSE SURROGATES ACROSS SPECIES AND LIFE STAGE Clewell et al. (2003b) compared the results of simulating the serum concentrations of perchlorate measured as area under the curve after drinking-water exposure with results of simulating the inhibition of thyroid iodide uptake after acute exposure to perchlorate in each of the PBPK models described above for the rat. The results of area under the curve simulations for serum perchlorate are shown in Figure E-8, and those of simulated inhibition of thyroid iodide uptake in Figure E-9. Clewell et al. (2003b) did not report whether up-regulated NIS parame- ter values were used in the simulations. Regardless, on the basis solely of the two internal-dose metrics, the lactating dam may be at greatest risk from perchlorate exposure if serum perchlorate area under the curve is used as the internal-dose surrogate, whereas the fetus appears to be the most sensitive if inhibition of thyroid iodide uptake is used as the surrogate. Clewell et al. (2003b) attributed the increased sensitivity of the lactating dam to the increase in protein binding of perchlorate in plasma (see Table E-1); fetal sensitivity to iodide uptake inhibition is multifactorial and is influenced by maternal transfer of perchlorate, inhibition of iodide transfer at the placenta, and other factors. EPA (2002b) used such simulations to calculate human equivalent exposures that corresponded to the internal-dose surrogates.

Appendix E 255 5 Serum Perchlorate AUC 4 Adult male rat Pregnant rat 3 (mg/L) Fetal rat 2 Lactating rat 1 Neonatal rat 0 0.01 0.1 1 10 Drinking-Water Perchlorate Dose (m g/kg per day) FIGURE E-8 Dose-response simulations of area under the curve for concentra- tions of perchlorate in serum of adult male, pregnant, fetal, lactating, and neonatal rats after drinking-water exposure as described in Clewell et al. (2003b). Abbrevia- tions: AUC, area under the curve; kg, kilogram; mg, milligram; mg/L, milligrams per liter; mg/kg/day, milligrams per kilogram of body weight per day. Inhibition of Iodide Uptake (%) 100 75 Adult male rat Pregnant rat 50 Fetal rat Lactating rat 25 Neonatal rat 0 0.01 0.1 1 10 Perchlorate Dose (m g/kg per day) FIGURE E-9 Dose-response simulations of inhibition of thyroid iodide uptake in adult male, pregnant, fetal, lactating, and neonatal rats after acute perchlorate exposure as described in Clewell et al. (2003b). Abbreviation: mg/kg/day, milli- gram per kilogram of body weight per day.

256 Health Implications of Perchlorate Ingestion SUMMARY AND CONCLUSIONS Although the PBPK models simplified the interactions between per- chlorate and iodide to a series of rate-limiting steps, their underlying bio- logic basis effectively reduced the uncertainties associated with extrapolat- ing blood concentrations of perchlorate and inhibition of thyroid iodide uptake across species, dose, and routes of exposure for adult animals. Furthermore, the parallelogram approach discussed in Chapter 4, applied where models have been validated in the adult, pregnant, and lactating rat, and adult humans, serves as a useful tool for constraining extrapolations to adult human females during pregnancy or lactation. However, given the important species differences in developmental biology and the current inability to validate extrapolation to human fetuses and neonates, such an approach should be used with caution for these potentially sensitive popula- tions. The committee concludes that it was appropriate for both DOD and EPA to limit animal-to-human extrapolations for pregnancy and lactation to maternal serum perchlorate concentrations or interactions with iodide in maternal thyroid NIS. The suite of PBPK models developed by DOD for the adult rat, adult human, pregnant rat and fetus, and lactating rat and neonate represent the current state-of-the-science approach for integrating available animal (rat) and human data on the disposition of perchlorate and iodide and their interactions at the level of the thyroid NIS. Although many of the PBPK model parameter values had to be estimated from a limited set of in vivo pharmacokinetic studies rather than independently measured, enough studies were available for validation of model simulations over a broad range of perchlorate doses and iodide concentrations to lend confidence to the application of the models for extrapolating internal-dose surrogates from animals to adult humans. If future studies are conducted to elucidate further the toxicity or mode of action of perchlorate in animal models, consider- ation should be given to updating the PBPK models because they provide a convenient framework to assemble current knowledge on the disposition of perchlorate in the body and on how it may interact with iodide at various stages of development. REFERENCES Brown, R.P., M.D. Delp, S.L. Lindstedt, L.R. Rhomberg, and R.P. Beliles. 1997. Physiological parameter values for physiologically based pharmacokinetic models. Toxicol. Ind. Health 13(4):407-484.

Appendix E 257 Clewell, R.A. 2001a. Consultative letter AFRL-HE-WP-CL-2001-0006. Phys- iologically Based Pharmacokinetic Model for the Kinetics of Perchlor- ate-Induced Inhibition of Iodide in the Pregnant Rat and Fetus. Memoran- dum to Annie M. Jarabek, NCEA, U.S. Environmental Protection Agency, Research Triangle Park, NC, from Rebecca Clewell, Air Force Research Laboratory, Department of the Air Force, Wright-Patterson Air Force Base, OH. May 10, 2001. Clewell, R.A. 2001b. Consultative letter, AFRL-HE-WP-CL-2001-0007. Physiologically-Based Pharmacokinetic Model for the Kinetics of Perchlor- ate-Induced Inhibition of Iodide in the Lactating and Neonatal Rat. Memo- randum to Annie M. Jarabek, NCEA, U.S. Environmental Protection Agency, Research Triangle Park, NC, from Rebecca Clewell, Air Force Research Laboratory, Department of the Air Force, Wright-Patterson Air Force Base, OH. May 24, 2001. Clewell, R.A., E.A. Merrill, P.J. Robinson. 2001. The use of physiologically- based models to integrate diverse data sets and reduce uncertainty in the prediction of perchlorate and iodide kinetics across life stages and species. Toxicol. Ind. Health 17(5-10):210-222. Clewell, R.A., E.A. Merrill, K.O. Yu, D.A. Mahle, T.R. Sterner, D.R. Mattie, P.J. Robinson, J.W. Fisher, and J.M. Gearhart. 2003a. Predicting fetal per- chlorate dose and inhibition of iodide kinetics during gestation: a physio- logically-based pharmacokinetic analysis of perchlorate and iodide kinetics in the rat. Toxicol. Sci. 73(2):235-255. Clewell, R.A., E.A. Merrill, K.O. Yu, D.A. Mahle, T.R. Sterner, J.W. Fisher, and J.M. Gearhart. 2003b. Predicting neonatal perchlorate dose and inhibition of iodide uptake in the rat during lactation using physiologically based pharmacokinetic modeling. Toxicol. Sci. 74(2):416-436. Clewell, R.A., E.A. Merrill, L. Narayanan, J.M. Gearhart, and P.J. Robinson. 2004. Evidence for competitive inhibition of iodide uptake by perchlorate and translocation of perchlorate into the thyroid. Int. J. Toxicol. 23(1):17-23. Corley, R.A., T.J. Mast, E.W. Carney, J.M. Rogers, and G.P. Daston. 2003. Evaluation of physiologically-based modeling of pregnancy and lactation for their application in children’s health risk assessments. Crit. Rev. Toxicol. 33(2):137-211. Durand, J. 1938. Recherches sur l'elimination des perchlorates, sur leur repartition dans les organs et sur leur toxicite [Research on the elimination of perchlor- ate, its distribution in organs and its toxicity]. Bull. Soc. Chim. Biol. 20:423-433 (as cited in Stanbury and Wyngarrden 1952). Eichler, O. 1929. Zur Pharmakologie der Perchloratwirkung [The pharmacology of the perchlorate effect]. Naunyn-Schmeideberg's Arch. Exp. Path. U. Pharmak. 144:251-260 (as cited in Stanbury and Wyngaarden 1952). EPA (U.S. Environmental Protection Agency). 2002a. A Review of the Reference Dose and Reference Concentration Process. Final Report. EPA/630/P-02/ 002F. Risk Assessment Forum, U.S. Environmental Protection Agency. December, 2002.

258 Health Implications of Perchlorate Ingestion EPA (U.S. Environmental Protection Agency). 2002b. Perchlorate Environmental Contamination: Toxicological Review and Risk Characterization. External Review Draft. NCEA-1-0503. National Center for Environmental Assess- ment, Office of Research And Development, U.S. Environmental Protection Agency, Washington, DC. [Online]. Available: http://cfpub.epa.gov/ncea/ cfm/recordisplay.cfm?deid=24002 [accessed October 6, 2004]. EPA (U.S. Environmental Protection Agency). 2002c. Report on the Peer Review of the U.S. Environmental Protection Agency's Draft External Review Document "Perchlorate Environmental Contamination: Toxicological Review and Risk Characterization". EPA/635/R02/003. June, 2002. National Cencter for Environmental Assessment, Office of Research and Development, U.S. Environmental Protection Agency, Washington, DC. [Online]. Available: http://www.epa.gov/ncea/pdfs/perchlorate/final_rpt. pdf [accessed August 23, 2004]. Fisher, J., P. Todd, D. Mattie, D. Godfrey, L. Narayanan, and K. Yu. 2000. Preliminary development of a physiological model for perchlorate in the adult male rat: a framework for further studies. Drug Chem. Toxicol. 23(1):243-258. Fisher, J.W., T.A. Whittaker, D.H. Taylor, H.J. Clewell, III, and M.E. Andersen. 1989. Physiologically-based pharmacokinetic modeling of the pregnant rat: A multiroute exposure model for trichloroethylene and its metabolite, trichloroacetic acid. Toxicol. Appl. Pharmacol. 99(3):395-414. Fisher, J.W., T.A. Whittaker, D.H. Taylor, H.J. Clewell, III, and M.E. Andersen. 1990. Physiologically-based pharmacokinetic modeling of the lactating rat and nursing pup: A multiroute exposure model for trichloroethylene and its metabolite, trichloroacetic acid. Toxicol. Appl. Pharmacol. 102(3): 497-513. Greer, M.A., G. Goodman, R.C. Pleus, and S.E. Greer. 2000. Does environmental perchlorate exposures alter human thyroid function? Determination of the dose-response for inhibition of radioiodine uptake. [Abstract]. Endocr. J. 40(Supp 1):146. Greer, M.A., G. Goodman, R.C. Pleus, and S.E. Greer. 2002. Health effects assessment for environmental perchlorate contamination: The dose response for inhibition of thyroidal radioiodine uptake in humans. Environ. Health Perspect. 110(9):927-937. Hays, M.T. and D.H. Solomon. 1965. Influence of the gastrointestinal iodide cycle on the early distribution of radioactive iodide in man. J. Clin. Invest. 44: 117-127. Hollowell, J.G., N.W. Staehling, W.H. Hannon, D.W. Flanders, E.W. Gunter, G.F. Maberly, L.E. Braverman, S. Pino, D.T. Miller, P.L. Garbe, D.M. DeLozier, and R.J. Jackson. 1998. Iodine nutrition in the United States. Trends and public health implications: Iodine excretion data from National Health and Nutrition Examination Surveys I and III (1971-1974 and 1988- 1994). J. Clin. Endocrinol. Metab. 83(10):3401-3408.

Appendix E 259 Kamm, G., and G. Drescher. 1973. Demonstration of perchlorate in urine [in German]. Beitr. Gerichtl. Med. 30:206-210. Merrill, E.A. 2000. Consultative letter, AFRL-HE-WP-CL-2000-0036. Human PBPK Model for Perchlorate Inhibition of Iodide Uptake in the Thyroid. Memorandum with attachments to Annie M. Jarabek, NCEA, U.S. Environ- mental Protection Agency, Research Triangle Park, NC, from Elaine Merrill, Air Force Research Laboratory/HEST, Department of the Air Force, Wright-Patterson Air Force Base, OH. June 28, 2000. Merrill, E.A. 2001a. Consultative letter, AFRL-HE-WP-CL-2001-0005. PBPK Model for Iodide Kinetics and Perchlorate-Induced Inhibition in the Male Rat. Memorandum with attachments to Annie M. Jarabek, NCEA, U.S. Environmental Protection Agency, Research Triangle Park, NC, from Elaine Merrill, Air Force Research Laboratory/HEST, Department of the Air Force, Wright-Patterson Air Force Base, OH. May 8, 2001. Merrill, E.A. 2001b. Consultative letter, AFRL-HE-WP-CL-2001-0008. PBPK Model for Perchlorate-Induced Inhibition of Radioiodide Uptake in Hu- mans. Memorandum with attachments to Annie M. Jarabek, NCEA, U.S. Environmental Protection Agency, Research Triangle Park, NC, from Elaine Merrill, Air Force Research Laboratory/HEST, Department of the Air Force, Wright-Patterson Air Force Base, OH. June 5, 2001. Merrill, E.A., R.A. Clewell, J.M. Gearhart, P.J. Robinson, T.R. Sterner, K.O. Yu, D.R. Mattie, and J.W. Fisher. 2003. PBPK predictions of perchlorate distribution and its effect on thyroid uptake of radioiodide in the male rat. Toxicol. Sci. 73(2):256-269. O'Flaherty, E. J., W. Scott, C. Schreiner, and R. P. Beliles. 1992. A physiologi- cally-based kinetic model of rat and mouse gestation: disposition of a weak acid. Toxicol. Appl. Pharmacol. 112(2):245-256. Reidel, C., O. Dohan, A. De la Vieja, C.S. Ginter, and N. Carrasco. 2001a. Jour- ney of the iodide transporter NIS: From its molecular identification to its clinical role in cancer. Trends Biochem. Sci. 26(8):490-496. Reidel, C., O. Levy, and N. Carrasco. 2001b. Post-translational regulation of the sodium/iodide symporter by thyrotropin. J. Biol. Chem. 276(24):21458- 21463. Shelley, M.L., M.E. Andersen, and J.W. Fisher. 1988. An inhalation distribution model for the lactating mother and nursing child. Toxicol. Lett. 43(1- 3):23-29. Soldin, O.P., S.J. Soldin, and J.C. Pezzullo. 2003. Urinary iodine percentile ranges in the United States. Clin. Chim. Acta 328(1-2):185-190. Stanbury, J.B., and J.B. Wyngaarden. 1952. Effect of perchlorate on the human thyroid gland. Metabolism 1(6):533-539. Williams, L.R., and R.W. Leggett. 1989. Reference values for resting blood flow to organs of man. Clin. Phys. Meas. 10(3):187-217 (as cited in Brown et al. 1997). Wolff, J. 1998. Perchlorate and the thyroid gland. Pharmacol. Rev. 50(1):89-105.

260 Health Implications of Perchlorate Ingestion Yu, K.O., D.A. Mahle, L. Narayanan, R.J. Godfrey, P.N. Todd, P. Parish, J. Mac- Cafferty, T. Ligman, T. Sterner, G. Buttler, P.N. Todd, M.A. Parish, J.D. McCafferty, T.A. Ligman, C.D. Goodyear, T.R. Sterner, T.A. Bausman, D.R. Mattie, and J.W. Fisher. 2001. Tissue distribution and inhibition of iodide uptake by perchlorate in pregnant and lactating rats in drinking water studies. Toxicologist 60(1):291-292. Yu, K.O., L. Narayanan, D.R. Mattie, R.J. Godfrey, P.N. Todd, T.R. Sterner, D.A. Mahle, M.H. Lumpkin, and J.W. Fisher. 2002. The pharmacokinetics of perchlorate and its effect on the hypothalamus-pituitary-thyroid axis in the male rat. Toxicol. Appl. Pharmacol. 182(2):148-159.

Health Implications of Perchlorate Ingestion Get This Book
×
Buy Paperback | $57.00 Buy Ebook | $45.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Perchlorate—a powerful oxidant used in solid rocket fuels by the military and aerospace industry—has been detected in public drinking water supplies of over 11 million people at concentrations of at least 4 parts per billion (ppb). High doses of perchlorate can decrease thyroid hormone production by inhibiting the uptake of iodide by the thyroid. Thyroid hormones are critical for normal growth and development of the central nervous system of fetuses and infants. This report evaluates the potential health effects of perchlorate and the scientific underpinnings of the 2002 draft risk assessment issued by the U.S. Environmental Protection Agency (EPA).

The report finds that the body can compensate for iodide deficiency, and that iodide uptake would likely have to be reduced by at least 75% for months or longer for adverse health effects, such as hypothryroidism, to occur. The report recommends using clinical studies of iodide uptake in humans as the basis for determining a reference dose rather than using studies of adverse health effects in rats that serve as EPA's basis. The report suggests that daily ingestion of 0.0007 milligrams of perchlorate per kilograms of body weight—an amount more than 20 times the reference dose proposed by EPA—should not threaten the health of even the most sensitive populations.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!