Page 1

UNCERTAINTIES IN RISK ASSESSMENT OF DIOXIN-LIKE COMPOUNDS A focus on systemic relative potencies and species differences

Karin van Ede


UNCERTAINTIES IN RISK ASSESSMENT OF DIOXIN-LIKE COMPOUNDS A focus on systemic relative potencies and species differences

Karin van Ede


ISBN: 978-90-393-6213-6 Cover Art: Anneke van Ede

Cover design: Multimedia, Faculteit Diergeneeskunde, Universiteit Utrecht Thesis design: Karin van Ede

Print: Gildeprint Drukkerij, Enschede, The Netherlands

The research described in this thesis was performed at the Institute for Risk Assessment Sciences (IRAS), Faculty of Veterany medicine, Utrecht University.

The research was financially supported by the European Commission Seventh Framework Programme FP7, SYSTEQ under grant agreement n°226694. Copyright Š 2014 K.I. van Ede

All rights reserved. No parts of this book may be reproduced in any form or by any means without permission of the author.


UNCERTAINTIES IN RISK ASSESSMENT OF DIOXIN-LIKE COMPOUNDS A focus on systemic relative potencies and species differences

ONZEKERHEDEN IN RISICOBEOORDELING VAN DIOXINE-ACHTIGE STOFFEN Een focus op systemische relatieve potenties en diersoort verschillen (met een samenvatting in het Nederlands)

Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof. dr. G.J. van der Zwaan, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op woensdag 28 mei 2014 des middags te 2.30 uur door

Karin Irene van Ede geboren op 13 augustus 1979 te Amsterdam


Promotor:

Prof. dr. M. van den Berg

Co-promotor: Dr. M.B.M. van Duursen


Alles is relatief

Albert Einstein


Contents

Abbreviations

PART I

GENERAL INTRODUCTION

Chapter 1

Introduction and outline of the thesis

PART II

INTAKE versus SYSTEMIC REPs

Chapter 2

Comparison of Intake and Systemic Relative Effect Potencies of Dioxin-like Compounds in Female Mice after a Single Oral Dose

31

Chapter 4

Tissue Distribution of Dioxin-like Compounds: Potential Impacts on Systemic Relative Potency Estimates

89

Chapter 3

Comparison of Intake and Systemic Relative Effect Potencies of Dioxin-like Compounds in Female Rats after a Single Oral Dose

PART III

HUMAN versus RODENT REPs

Chapter 5

Differential relative effect potencies of some dioxin-like compounds in human peripheral blood lymphocytes and murine splenic cells

Chapter 6

In vitro and in silico derived relative effect potencies of Ah-receptor mediated effects by PCDD/Fs and PCBs in human, rat, mouse and guinea pig CALUX cell lines

PART IV

DISCUSSION, CONCLUSION and ANNEX

Chapter 7

Summary, general discussion, conclusions and recommendations

9 15

57

111 131

173


Annex

References Nederlandse samenvatting Dankwoord About the author

199 215 224 228


Abbreviations

abbreviations AhR Aryl hydrocarbon receptor AhRR Aryl hydrocarbon receptor repressor ANOVA Analysis of variance BMR Benchmark response CALUX Chemical-activated luciferase gene expression assays CYP Cytochrome P450 EROD Ethoxyresorufin-O-deethylase Dioxin-like compound DLC HepG2 Human hepatoblastoma cell line HOMO Highest occupied molecular orbital 1234678-HpCDD 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin 123678-HxCDD 1,2,3,6,7,8-hexachlorodibenzo-p-dioxin 123789-HxCDD 1,2,3,7,8,9-hexachlorodibenzo-p-dioxin 1234678-HpCDF 1,2,3,4,6,7,8-heptachlorodibenzofuran 1234789-HpCDF 1,2,3,4,7,8,9-heptachlorodibenzofuran 123478-HxCDF 1,2,3,4,7,8-hexachlorodibenzofuran 123678-HxCDF 1,2,3,6,7,8-hexachlorodibenzofuran 123789-HxCDF 1,2,3,7,8,9-hexachlorodibenzofuran 234678-HxCDF 2,3,4,6,7,8-hexachlorodibenzofuran LUMO Lowest unoccupied molecular orbital NDL Non-dioxin-like OCDD Octachlorodibenzo-p-dioxin OPLS Orthogonal projection to latent structures PAH Polycyclic aromatic hydrocarbon PBL Peripheral blood lymphocytes Principal component analysis PCA PCB Polychlorinated biphenyls PCB 74 2,4,4’,5-tetrachlorobiphenyl PCB 77 3,3’,4,4’-tetrachlorobiphenyl PCB 81 3,4,4’,5-tetrachlorobiphenyl PCB 105 2,3,3’,4,4’-pentachlorobiphenyl PCB 114 2,3,4,4’,5-pentachlorobiphenyl PCB 118 2,3’,4,4’,5-pentachlorobiphenyl PCB 123 2’,3,4,4’,5-pentachlorobiphenyl PCB 126 3,3’,4,4’,5-pentachlorobiphenyl PCB 153 2,2’,4,4’,5,5’-hexachlorobiphenyl PCB 156 2,3,3’,4,4’,5-hexachlorobiphenyl

9


PCB 157 PCB 167 PCB 169 PCB 189 PCDD PCDF PeCDD 12378-PeCDF 4-PeCDF PLS Q2 QSAR R2 REP RfD RMSEcv RMSEE RMSEP SD TCDD 2378-TCDF TDI TEF TEQ

10

2,3,3’,4,4’,5’-hexachlorobiphenyl 2,3’,4,4’,5,5’-hexachlorobiphenyl 3,3’,4,4’,5,5’-hexachlorobiphenyl 2,3,3’,4,4’,5,5’-heptachlorobiphenyl Polychlorinated dioxin Polychlorinated furan 1,2,3,7,8-pentachlorodibenzodioxin 1,2,3,7,8-pentachlorodibenzofuran 2,3,4,7,8,-pentachlorodibenzofuran Partial least squares Cross-validated R2 Quantitative structure−activity relationship Determination coefficient Relative effect potency Reference dose Root mean square error of cross validation Root mean square error of the estimation Root mean square error of the prediction Sprague Dawley 2,3,7,8-tetrachlorodibenzodioxin 2,3,7,8-tetrachlorodibenzofuran Tolerable daily intake Toxic equivalency factor Toxic equivalency


Abbreviations

11


Part

I

General Introduction

Te weten dat je onwetend bent, is het begin van alle wijsheid.

Viviane van Avalon


Chapter

1

Introduction and outline of the thesis


Introduction

Dioxins and dioxin-like compounds

D

ioxins and dioxin-like compounds belong to the group of persistent organic pollutants (POPs). They are highly lipophilic and resistant to metabolism. Because of these characteristics, they bioaccumulate and biomagnify in the food chain and humans (Van den Berg et al., 1994). The term “dioxins” is commonly used to refer to the family of structurally and chemically related polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzo-p-dibenzofurans (PCDFs) and some dioxin-like polychlorinated biphenyls (PCBs). The structure of PCDDs and PCDFs comprises of a dioxin or furan ring, respectively, stabilized by two flanking benzene rings. PCBs consist of two connected phenyl rings. In total 75 PCDDs, 135 PCDFs and 209 PCBs exist based on the number of chlorine atoms and their positions on the aromatic rings. However, only 7 PCDDs, 10 PCDFs and 12 PCBs are classified to cause toxic effects. These 29 congeners are referred to as “dioxin-like” compounds (DLCs) of which 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) is the most toxic and wellstudied congener. Figure 1 shows the structural formula of PCDDs, PCDFs and PCBs and the numbering of the carbon atoms where chlorine substitution may occur.

Figure 1. Chemical structure of PCDDs (left), PCDFs (middle) and PCBs (right) and numbering of carbons where chlorine substitution may occur.

Source and exposure Source PCDDs and PCDFs have no commercial applications. They are mainly unwanted byproducts of industrial activities such as production of herbicides and fungicides, paper bleaching, and combustion processes including incineration(Fiedler, 1996; IARC, 2012). Furthermore, PCDDs and PCDFs can also be formed during natural combustion processes such as forest fires and volcanic eruptions (Czuczwa and Hites, 1984; Czuczwa and Hites, 1986). In contrast, PCBs have been intentionally produced and used for example, in electric fluids in transformers and capacitors, as pesticide extenders, as flame-retardants, dedusting agents, and as ingredients in paint. Their manufacture was banned in the 1980s. PCBs can also be produced as accidental byproducts of various 17

1


combustion processes (Breivik et al., 2007; Safe, 1990; Van Caneghem et al., 2014).

Human exposure PCDDs, PCDFs and PCBs have been detected in almost every component of the global ecosystem including air, water, fish, wildlife, food, human adipose tissue, serum and milk (Safe, 1990). Human background exposure is primarily through the diet, with food of animal origin being the most important source. Strict regulatory controls on major industrial sources and regulatory national monitoring programs that screen and quantify the presence of PCDDs, PCDFs and PCBs in feed and food have contributed to reduce human exposure by approximately 90% since the late 1960s (EFSA, 2012; Hays and Aylward, 2003). As a result, for the general population, a significant decrease in plasma levels of PCDDs and PCDFs has been seen over the time frame from 1970 to 2010 (Consonni et al., 2012; Hays and Aylward, 2003). Exposure to mixtures of dioxins, furans, and PCBs is often expressed in terms of TCDD toxic equivalencies (TEQs; discussed further below). Today, the major contributors in dietary exposure to DLCs, on the basis of TEQs, are 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), 1,2,3,7,8-pentachlorodibenzop-dioxin (PeCDD), 1,2,3,6,7,8-hexachlorodibenzo-p-dioxin (123678-HxCDD), 2,3,4,7,8-pentachlorodibenzofuran (4-PeCDF), 3,3’,4,4’,5-pentachlorobiphenyl (PCB 126) for the PCDDs, PCDFs and non-ortho-PCBs, respectively, and 2,3’,4,4’,5-pentachlorobiphenyl (PCB 118) and 2,3,3’,4,4’,5-hexachlorobiphenyl (PCB 156) for the mono-ortho-PCBs (Liem et al., 2000; Parvez et al., 2013). Several regulatory authorities and scientific organizations have concluded that a tolerable daily intake (TDI) of 1-4 pg TEQ/kg body weight (BW) is likely to be without adverse effects (ECSCF, 2001; JECFA, 2001; Van Leeuwen et al., 2000). Nevertheless, for parts of the population in some countries, human exposure is still above the current TDI (Bilau et al., 2008; De Mul et al., 2008). In particular breast-fed infants are a sensitive group that can exceed the TDI up to two orders of magnitude (Li et al., 2009). Human exposure to DLCs is often assessed through measurement of DLC concentrations in blood. In a general population, the average total TEQs as measured in blood is approximately 10 pg TEQ/g lipid (Rawn et al., 2012). Gender does not affect the blood concentration, however, special consumption habits (low or high consumption of animal products), living area (industrialized or not), age and lactation can have an effect. Human exposure incidents Elevated exposure to dioxins can also occur during accidental events, intentional poisoning or occupational scenarios. Probably one of the most notable incidents that happened is the assassination attempt on the Ukrainian President Viktor Yushchenko, in 2004 with TCDD (Sorg et al., 2009). However, several accidental exposures to DLCs have

18


Introduction

also taken place in the last decades. Well-known is the chemical plant explosion near Seveso in 1976, which resulted in the highest known exposure to TCDD in a residential population (Mocarelli et al., 1988). Other accidental exposures include the Yusho and Yu Cheng poisonings in Japan and Taiwan, when rice oil contaminated with PCBs was used for cooking in 1968 and 1979, respectively (Chen et al., 1985; Olafsson et al., 1988).

Also chlorophenoxy herbicides, one of them being Agent Orange, which was used as a defoliant during the Vietnam War and contaminated with TCDD, led to exposure of military personnel as well as populations in areas it was used (Michalek et al., 1990; Schecter et al., 1989; Wolfe et al., 1990). Furthermore, improper disposal of residues from the manufacture of chlorophenoxy herbicides resulted in contaminated residential soil at the town Times Beach, Missouri, and led to a complete evacuation of the town in 1983 due to concerns for exposure and possible health effects (Kimbrough et al., 1984). There have been several incidents of food contamination after animal feeds have been accidentally mixed with dioxin-containing substances. After the so-called “dioxin crisis� in Belgium in 1999 (Bernard and Fierens, 2002, Bernard et al., 1999), where high levels of PCBs were found in poultry and eggs, regulatory national monitoring programs began consistently screening food and feed samples in the European Union. After that, several smaller incidents with contaminated animal feed or food were reported (Hoogenboom et al., 2007; Hoogenboom et al., 2004). Currently, occupational exposure may still occur during industrial activities in which DLCs are unintentionally produced, such as in waste incinerators or during the production of certain pesticides or chemicals. Mechanism of toxicity Most, if not all, toxic effects associated with dioxin exposure are mediated through the aryl hydrocarbon receptor (AhR). The AhR is a ligand-activated transcription factor present in many cells. Although a clear endogenous ligand is still not known, the AhR appears to play an important role in many biological functions (Nguyen and Bradfield, 2008). In absence of a ligand, the receptor is present in the cytosolic compartment of the cell as a multiprotein complex containing a heat shock protein 90 (hsp90), HBV X-associated protein (XAP2), and the co-chaperone protein p23 (Beischlag et al., 2008; Hankinson, 1995a). Following ligand binding, multiple signaling pathways and cellular regulatory factors 19

1


are induced that involves a combination of both classical and non-classical AhRdependent mechanisms. For the classical mechanism, AhR undergoes a conformation change that results in translocation of the complex into the nucleus (Hord and Perdew, 1994; Pollenz et al., 1994), where it binds to the aryl hydrocarbon receptor nuclear translocator (ARNT). Subsequently, the AhR:ARNT dimer binds to and activates the dioxin responsive elements (DREs), after which transcription and translation of a battery of genes occur, such as drug-metabolizing enzymes cytochrome P450 (CYP)1A1, 1A2, 1B1, glutathione-S-transferase, and UDP-glucuronosyltransferase. In addition, many other cellular pathways such as, aryl hydrocarbon receptor repressor (AhRR), TCDD-inducible poly(ADP-ribose) polymerase (TiPARP) or son of sevenless (SOS1), the primary mediator of Ras activation (Denison and Nagy, 2003; Denison et al., 2011).

Figure 2. The classical mechanism of AhR-dependent gene activation (Denison et al., 2011).

This classical mechanism of AhR-dependent gene activation has long been considered the pathway by which DLCs produce their biological and toxicological effects. However, ongoing research reveals many newly characterized AhR-dependent alterations in diverse cell signaling pathways and protein regulatory factors that are induced via non-classical mechanisms. Although currently only very few of these non-classical mechanisms have been elucidated, they certainly contribute to AhR-ligand induced toxic and biological responses (Denison et al., 2011). The biochemical and toxic responses upon exposure to DLCs in experimental animals 20


Introduction

are characterized by enzyme induction, retinoid changes, severe weight loss, thymic atrophy, hepatoxicity, immunotoxicity, endocrine disruption and tumorogenesis (Birnbaum, 1994; Birnbaum and Tuomisto, 2000; Safe, 1990). In humans, short-term exposure to high levels of dioxins may result in skin lesions, such as chloracne and patchy darkening of the skin as well as altered liver function. Long term exposure is linked to impairment of the immune system, developing nervous system, the endocrine system, reproductive functions, and formation of extra-hepatic carcinogenic responses (ECSCF, 2001; IARC, 2012; JECFA, 2001; UKCOT, 2001; USEPA, 2012). Risk assessment Toxic equivalency factor Assessing the potential risk associated with exposure to dioxins and dioxin-like compounds is challenging, as humans and wildlife are exposed to a complex mixture of these structurally related compounds (Safe, 1994a). Based on the assumption that they share the same mechanism of action, it is assumed that their individual potencies are additive. This has led to the development of the toxic equivalency concept (Safe, 1990; Safe, 1994b), in which each congener is assigned a specific toxic equivalency factor (TEF) that reflects its potency to produce an AhR-mediated biological or toxicological effect compared with the most potent AhR agonist known, TCDD. For inclusion in the TEF concept, a compound must • show a structural relationship to the PCDDs and PCDFs; • bind to the AhR; • elicit AhR-mediated biochemical and toxic responses; • be persistent and accumulate in the food chain.

To characterize the total toxicity in a matrix, such as food, total TEQs can be calculated by multiplying the concentration of each congener with its TEF value, after which it is summed up to calculate total TEQs. This approach is now used world-wide for risk characterization in food, feed and human populations.

From the early 1990s, the World Health Organization (WHO) started organizing international expert meetings with the objective of harmonizing TEFs for dioxin and dioxin-like compounds. In 1993, the first evaluation was done that resulted in human and mammalian WHO-TEFs (Ahlborg et al., 1994). Since 1998 these TEF values have also been differentiated between mammals, birds, and fish, with mammalian TEFs being used for human risk assessment (Van den Berg et al., 1998). In June 2005, a third 21

1


Table 1: Congeners assigned with a WHO-TEF Congener Clorinated dibenzo-p-dioxins 2378-TCDD* 12378-PeCDD* 123478-HxCDD 123678-HxCDD** 123789-HxCDD 1234678-HpCDD** OCDD Chlorinated dibenzofurans 2378-TCDF** 12378-PeCDF 23478-PeCDF* 123478-HxCDF** 123678-HxCDF 123789-HxCDF 234678-HxCDF** 1234678-HpCDF** 1234789-HpCDF** OCDF Non-ortho-substituted PCBs 3,3’,4,4’-tetraCB (PCB 77)** 3,4,4’,5-tetraCB (PCB 81) 3,3’,4,4’,5-pentaCB (PCB 126)* 3,3’,4,4’,5,5’-hexaCB (PCB 169)** Mono-ortho-substituted PCBs 2,3,3’,4,4’-pentaCB (PCB 105)** 2,3,4,4’,5-pentaCB (PCB 114) 2,3’,4,4’,5-pentaCB (PCB 118)* 2,3,3’,4,4’,5-hexaCB (PCB 156)* 2,3,3’,4,4’,5’-hexaCB (PCB 157) 2,3’,4,4’,5,5’-hexaCB (PCB 167)** 2,3,3’,4,4’,5,5’-heptaCB (PCB 189)**

WHO-TEFa

1 1 0,1 0,1 0,1 0,01 0,0003 0,1 0,03 0,3 0,1 0,1 0,1 0,1 0,01 0,01 0,0003

0,0001 0,0003 0,1 0,03

0,00003 0,00003 0,00003 0,00003 0,00003 0,00003 0,00003

a Current WHO-TEF (Van den Berg et al., 2006) * Congeners used in the in vivo + in vitro studies (including the non-dioxin like PCB 153) ** Congeners used in the in vitro studies (including the non-dioxin like PCB 74)

WHO expert meeting to reevaluate the mammalian 1998 WHO-TEF values was held in Geneva, Switzerland. For this, a database with all relative effect potencies (REPs) from known endpoints of DLCs (e.g. CYP1A1 activity) was compiled containing in vivo and in vitro data (Haws et al., 2006). During the 2005 expert meeting, the expert panel typically 22


Introduction

assigned TEF factors based on a point estimate between the 50th and 75th percentiles of the REP range, which was generally closer to the 75th percentile in order to be health protective. As a default, all TEF values are assumed to vary in uncertainty by at least +/- an order of magnitude around the median value, depending on the congener and its REP distribution (Van den Berg et al., 2006). Currently there are 7 PCDDs, 10 PCDFs, 4 non-ortho-PCBs and 8 mono-ortho-PCBs that have been assigned with a TEF-value (See Table 1).

Uncertainties for TEFs Despite the many scientific expert consultations and huge amount of scientific data that has been published since the development of these TEFs, some crucial gaps still exist in the TEF methodology. One of these concerns is related to the question whether the current TEFs, which are primarily based on in vivo studies with oral dosage as the principal route of exposure, can be also be used for risk assessment based on a systemic concentration, e.g. human blood. Another important uncertainty in the current TEF concept comes from the fact that TEFs are generally based on rodent studies, but are ubiquitously applied for human risk assessment, without scientific validation for this applicability domain. Are intake- and systemic-REPs similar? At present, the TEF concept for human risk assessment is mainly based on in vivo animal experiments with oral dosage as the principal route of exposure. Consequently, the present human TEFs may only be valid for estimating the risk in a population upon dietary exposure to DLCs (Van den Berg et al., 2006). Using these intake-based TEFs to assess the risk of humans based on systemic concentrations, for example blood, may therefore be scientifically incorrect and unsound.

Differences in toxicokinetics between the congeners can also influence the potency of a congener when calculated on either administered dose or systemic concentrations. Basically, each step in toxicokinetics (absorption, distribution, metabolism, and excretion) may contribute to the relative potency of a congener, if it behaves significantly differently from TCDD. For example, if congeners are more poorly absorbed or faster metabolized than TCDD, higher administered doses are needed to reach or maintain an effect concentration at the target tissue. As a consequence, REPs that are determined based on systemic concentrations (e.g. blood, adipose tissue or liver) will most likely be higher, compared to those based on an administered dose. In other words, absorption, metabolism and elimination rates can really make a difference between systemic versus 23

1


intake concentrations relative to TCDD.

Another important aspect that can affect the systemic target concentration of a DLC is the degree of hepatic sequestration due to CYP1A2 protein binding in the liver. Several studies have shown that DLCs can bind strongly to the CYP1A2 protein and, as a result, sequester in the liver (Devito et al., 1998; Diliberto et al., 1995; Diliberto et al., 1997; Diliberto et al., 1999). The variation in distribution due to sequestration between a congener and TCDD can cause differences between hepatic and extra-hepatic systemic REPs. Limited data support the idea that differences in absorption, distribution, metabolism, and excretion may contribute to differences in REP of a congener when either based on an administered dose or systemic concentration (Budinsky et al., 2006; Devito and Birnbaum, 1995; DeVito et al., 1997; DeVito et al., 2000). As a result, the 2005 WHO expert meeting, concluded that there was insufficient data available to develop “systemic” TEFs based on the available knowledge at that time. Are human and rodent REPs similar? It is generally assumed that TEFs based on rodent studies are appropriate for human risk assessment. Yet, it is well known that upon AhR activation a wide variety of species-specific toxic and biological effects can occur (Denison et al., 2011). Generally, the human AhR is considered to be somewhat less responsive to DLCs than the AhR in many rat and some mouse strains (Connor and Aylward, 2006; Ema et al., 1994). This has been shown by in vitro studies that showed human cells to be 10-1000 times less sensitive for TCDD-induced effects than those of the rat and monkey cells (Silkworth et al., 2005). In addition, congener-specific REPs have been suggested to vary across species, which is of special importance for the major research themes of this thesis (Nagayama et al., 1985; Silkworth et al., 2005; Sutter et al., 2010; Van Duursen et al., 2005; Zeiger et al., 2001). Especially the species-differences in REPs of the non-ortho substituted 3,3’,4,4’5-pentachlorobiphenyl (PCB 126) has been subject of much scientific debate. In addition, some PCDDs and PCDFs also show species-specific differences in REPs (Nagayama et al., 1985; Sutter et al., 2010). Even though these species-specific differences in potency, in particular for PCB 126, were acknowledged by the expert panel during the WHO-TEF re-evaluation in 2005, it was concluded that more information regarding the difference between rodents and humans is needed (Van den Berg et al., 2006).

24


Introduction

EU-project SYSTEQ The work presented in this thesis is part of the European Seventh Framework Programme SYSTEQ (www.systeqproject.eu). The aim of this project was to develop, validate and implement human systemic Toxic Equivalencies (TEQs) as biomarkers for dioxin-like compounds. The project was a collaboration between different universities and research institutes from The Netherlands (IRAS, Utrecht University), Sweden (Umeå University, Karolinska Institute), Germany (Technical University of Kaiserslautern), Czech Republic (Veterinary Research Institute Brno) and Slovakia (Slovak Medical University) and was coordinated by IRAS, Utrecht University. Major objectives within SYSTEQ were to establish possible differences between “intake” and “systemic” REPs, to identify novel quantifiable biomarkers for exposure in human and rodent models and to establish possible differences between humans and experimental animal species. Scope of this thesis The availability of systemic-TEFs as well as human-TEFs to determine systemic-TEQs might be essential for accurate human risk assessment, because concentrations in human blood or tissues are often used and proven most suitable to determine abovebackground exposure situations, e.g. by accidental food poisoning or suspected differences in environmental exposure between populations. Measured blood concentrations are also widely used to track changes in population exposure levels, and to assess potential relationships between health outcomes and DLC exposure levels. The major question is: does the use of current rodent derived intake-TEFs only give a minimal error in the risk assessment process compared to the many other uncertainties inherent to the TEF concept? This question is addressed in this thesis, which is divided into four parts. Following this introduction, which includes a description of background information on the topic (Part I, Chapter 1), Part II provides an assessment of the comparability of intakeREPs and systemicREPs. Part III provides a comparison of REPs in human versus rodent systems. Finally, Part IV presents a discussion and integration of the research in the thesis. Part II Intake- versus systemic-REPs In this part of the thesis, intakeREPs and systemicREPs were compared in female C57BL/6 mice (Chapter 2) and Sprague-Dawley rats (Chapter 3) based on the administered dose and liver, adipose, or plasma concentrations. C57BL/6 mice and Sprague Dawley rats were chosen as both species have been commonly used for studies on dioxin-like 25

1


compounds during the last decades (Haws et al., 2006).

As can be seen in Table 1, the congeners TCDD, PeCDD, 4-PeCDF, PCB 126, 118 and 156 were used for these in vivo studies, representing approximately 90% of the dioxinlike activity (TEQs) in the human food chain (Liem et al., 2000). The non-dioxin-like 2,2’,4,4’,5,5’-hexachlorobiphenyl (PCB 153) was also included as a negative control. In the past, it has been shown that low levels of contamination of the mono-ortho-PCBs with more potent dioxin-like compounds can significantly impact the potency of a congener (Peters et al., 2006). For this reason, PCB 118, 156 and 153 were specially purified to a very high degree before the start of the in vivo studies. After purification, the remaining TEQ contributions that were present in these three PCBs were calculated to have no influence on the final outcome of the in vivo studies.

Three days after oral exposure, intakeREPs and systemicREPs were calculated based on hepatic cytochrome P450 (CYP)1A1-associated ethoxyresorufin-O-deethylase (EROD) activity and Cyp1a1, 1a2, 1b1 and AhRR gene expression in the livers and peripheral blood lymphocytes (PBLs). Although induction of CYP1A1, 1B1, 1A2 enzymes and AhRR are not necessarily a measure of toxicity, it is considered to be the most sensitive biomarker for AhR activation (Abel and Haarmann-Stemmann, 2010; Denison and Heath-Pagliuso, 1998). Moreover, studies have shown a high correlation between the degree of induction of these P450 enzymes and toxic responses caused by DLCs, such as wasting syndrome, thymic atrophy, or hepatic porphyrin accumulation (Safe, 1990; van Birgelen et al., 1996). Within the SYSTEQ project, special attention was also given to investigate sensitive novel biomarkers that would have a more direct link with a toxic effect such as ALDH3A1 (Muzio et al., 2012). However, for the work described in this thesis, it was decided to use only the classical biomarkers as those are predominantly driving the current TEFs. For calculating relative potencies it was decided to use a benchmark response (BMR) approach instead of using an effect concentration of 50% (EC50), which generally forms the basis of REP determination and the TEF concept. Many experimental studies done in the past have reported significant differences in maximal induction between dioxinlike congeners. This phenomenon was also observed in the studies presented in this thesis. Many of the dose–response curves did not attain a similar maximum efficacy or parallel Hill slope as the reference compound TCDD. As a result, significant uncertainties in calculating EC50 values can occur. Therefore, REPs in this thesis were calculated using concentrations at which a congener reached 20% of the TCDD response (BMR20TCDD concentrations). The advance of this benchmark approach using the lower part of the dose–response curve is that dissimilarities in efficacy and Hill slope are less pronounced 26


Introduction

than when using the EC50-based approach.

A relevant question related to the 3-day single dose in vivo studies described in chapters 2 and 3, is their relevance to subchronic animal studies. The latter have frequently formed a major contribution to the derivation and selection of a TEF, because chronic human exposures are nowadays the most common situation for application of TEFs (Haws et al., 2006; Van den Berg et al., 2006). Therefore, we compared in chapter 4 the tissue distribution data across the tested compounds from our 3-day, single dose studies with those of previous studies in rodents using single as well as subchronic dosing regimens. In addition, we also evaluated tissue distribution data from human studies and compared these with results from rodent studies. Concentration-response data of hepatic EC50 concentrations for CYP1A1 activity or gene expression following TCDD exposure were also compared between both types of studies. Part III Human- versus rodent-REPs In part III of this thesis, species-specific differences in REPs between human and rodents were investigated for 20 congeners (See Table 1). In chapter 5, REPs were determined based on CYP1A1, 1B1 and aryl hydrocarbon receptor repressor (AhRR) gene expression as well as CYP1A1 activity in human peripheral blood lymphocytes (PBL) and Cyp1a1 gene expression in murine splenic cells. Human PBLs are relatively easy to collect, which makes this an interesting target for monitoring human health among other for DLCs. Changes in AhR-mediated gene expression in PBLs have been used as biomarkers of human exposure by AhR agonists such as DLCs. This in spite of the fact that significant inter-individual variability in responses have been observed (Van Duursen et al., 2005). In chapter 6, REPs of 20 selected DLCs were determined in chemical-activated luciferase expression (DR-CALUX速) cell lines from rat, mouse and human hepatoma cells, and guinea pig intestinal adenocarcinoma cells. Furthermore, quantitative structure-activity relationship (QSAR) analysis were performed to provide a prediction of the biological activity of structurally similar but untested compounds, as well as discovering structural analogies that might influence the activity of a group of compounds. Finally, a summary and general discussion of the results described in this thesis are given in Part IV / chapter 7. This discussion provides an overall view of the pattern of results and the potential implications for human risk assessment. 27

1


Part

II

Intake- versus systemic-REPs

Part of the secret of success in life is to eat what you like and let the food fight it out inside.

Mark Twain


Chapter

2

Comparison of Intake and Systemic Relative Effect Potencies of Dioxin-like Compounds in Female Mice after a Single Oral Dose Karin I. van Ede1 Patrik L. Andersson2 Konrad P.J. Gaisch1 Martin van den Berg1 Majorie B.M. van Duursen1 1

Endocrine toxicology group, Institute for Risk Assessment Sciences (IRAS), Utrecht University, the Netherlands 2 Department of Chemistry, Umeå University, Umeå, Sweden

Environmental Health Perspectives 121 (7): 847 – 853 (2013)


Abstract Background: Risk assessment for mixtures of polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and polychlorinated biphenyls (PCBs) is performed using the toxic equivalency factor (TEF) approach. These TEF values are derived mainly from relative effect potencies (REPs) linking an administered dose to an in vivo toxic or biological effect, resulting in “intake” TEFs. At present, there is insufficient data available to conclude that intake TEFs are also applicable for systemic concentrations (e.g., blood and tissues). Objective: We compared intake and systemic REPs of 1,2,3,7,8-pentachlorodibenzodioxin (PeCDD), 2,3,4,7,8-pentachlorodibenzofuran (4-PeCDF), 3,3´,4,4´,5-pentachlorobiphenyl (PCB 126), 2,3´,4,4´,5-pentachlorobiphenyl (PCB 118), and 2,3,3´,4,4´,5-hexachlorobiphenyl (PCB 156) in female C57BL/6 mice 3 days after a single oral dose. Methods: We calculated intake REPs and systemic REPs based on administered dose and liver, adipose, or plasma concentrations relative to TCDD. Hepatic cytochrome P450 1A1–associated ethoxyresorufin-O-deethylase (EROD) activity and gene expression of Cyp1a1, 1a2 and 1b1 in the liver and peripheral blood lymphocytes (PBLs) were used as biological end points. Results: We observed up to one order of magnitude difference between intake REPs and systemic REPs. Two different patterns were discerned. Compared with intake REPs, systemic REPs based on plasma or adipose levels were higher for PeCDD, 4-PeCDF, and PCB 126 but lower for the mono-ortho PCBs 118 and 156. Conclusions: Based on these mouse data, the comparison between intake REPs and systemic REPs reveals significant congener-specific differences that might warrants the development of systemic TEFs to calculate toxic equivalents (TEQs) in blood and body tissues.

32


Intake and systemic REPs of DLCs in C57BL/6 mice

Introduction

P

olychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and polychlorinated biphenyls (PCBs) are persistent and widespread contaminants. Of the 419 possible congeners that exist, 7 PCDDs, 10 PCDFs, and 12 non-ortho and mono-ortho PCBs are classified as having dioxin-like effects. Most, if not all, toxic effects of dioxin-like compounds (DLCs) are mediated through the aryl hydrocarbon receptor (AHR); the toxic effects of these DLCs include endocrine, developmental, immune, and carcinogenic effects, among others (Birnbaum, 1994; Birnbaum and Tuomisto, 2000; Safe, 1990; White and Birnbaum, 2009). Humans are exposed to a complex mixture of these DLCs mainly through the diet, with food of animal origin being the most important source. Although exposure has significantly decreased during the past decades (De Mul et al., 2008; Fürst, 2006), current human exposure is still above the tolerable daily intake (TDI) or reference dose (RfD) levels for parts of the population in some countries (Bilau et al., 2008; De Mul et al., 2008; Llobet et al., 2008; Loutfy et al., 2006; Tard et al., 2007). Therefore, improving the risk assessment process for this class of compounds remains important and societally relevant.

Currently, risk assessment of DLCs is based on the toxic equivalency factor (TEF) approach (Safe, 1990; 1994a) endorsed by the World Health Organization (WHO) (Van Den Berg et al., 1998; 2006). Each congener-specific TEF is derived from multiple relative effect potencies (REPs) determined from a range of AhR-specific end points [e.g., cytochrome P450 1A1 (CYP1A1) activity]. The toxic or biological potency of a congener is compared to that of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). A shortcoming of the TEF concept originates from the fact that the TEFs were established primarily from in vivo end points linking administered dose levels (via oral exposure) to toxic or biological effects, resulting in “intake” TEFs (intakeTEFs) (Haws et al., 2006; Van den Berg et al., 2006). Consequently, these intakeTEFs are applicable only for situations in which ingestion (e.g., food intake, consumption of breast milk) is known. However, because ingestion data for humans is often lacking or difficult to establish, blood or adipose tissue levels are frequently used to quantify the relative exposure to humans. Subsequently, regulatory authorities commonly calculate risks based on blood or adipose tissue (systemic) levels using these intakeTEFs. Unfortunately, even for the most relevant DLCs, experimental validation is in sufficient to either reject or accept this application of intakeTEFs for blood or tissue levels. There is limited evidence suggesting that the use of intakeTEFs instead of systemicTEFs may lead to inaccurate interpretation of the risk because of congener-specific toxicokinetic differences (Chen et al., 2001; Devito et al., 1998; Hamm et al., 2003). Properties such as absorption, distribution, metabolism, and excretion can clearly contribute to the potency of a congener (Budinsky et al., 2006; 33

2


Devito and Birnbaum, 1995; DeVito et al., 1997; 2000) and may be misinterpreted when relying solely on intakeTEFs. At the most recent WHO expert meeting (in 2005) where the TEFs were (re)evaluated, it was concluded that insufficient data were available to develop systemicTEFs, leaving a major gap in the risk assessment process for DLCs (Van den Berg et al., 2006). To fill this data gap, the European Union (EU) project SYSTEQ was initiated, with the main objectives of establishing in vivo systemicREPs in the mouse and rat, with special focus on effects in peripheral blood lymphocytes (PBLs) as potential biomarkers of exposure.

In the present study we compared intakeREPs and systemicREPs in female C57BL/6 mice based on the administered dose and liver, adipose, or plasma concentrations. We used 2,3,7,8-tetrachlorodibenzodioxin (TCDD), 1,2,3,7,8-pentachlorodibenzodioxin (PeCDD), 2,3,4,7,8-pentachlorodibenzofuran (4-PeCDF), 3,3´,4,4´,5-pentachlorobiphenyl (PCB 126), 2,3´,4,4´,5-pentachlorobiphenyl (PCB 118), and 2,3,3´,4,4´,5-hexachlorobiphenyl (PCB 156), which represent approximately 90% of the dioxin-like activity in the human food chain (Liem et al., 2000); we also included the non-dioxin-like 2,2´,4,4´,5,5´-hexachlorobiphenyl (PCB 153). Three days after exposure, we calculated intake REPs and systemicREPs for hepatic CYP1A1-associated ethoxyresorufin-O-deethylase (EROD) activity and Cyp1a1, 1a2, and 1b1 gene expression in the mouse liver and PBLs. Materials and Methods Chemicals TCDD, PeCDD, 4-PeCDF, and PCB 126 were purchased from Wellington Laboratories Inc. (Guelph, Ontario, Canada) and dissolved in corn oil (ACH Food Companies Inc., Oakbrook, IL, USA); concentrations were then checked and confirmed by Wellington Laboratories Inc. We purchased PCB 118, PCB 156, and PCB 153 from Cerilliant Corp. (Round Rock, TX, USA). These three PCBs and corn oil (Sigma- Aldrich, Stockholm, Sweden) were purity checked; and PCB 118 and PCB 156 were purified at the Department of Chemistry, Umeå University. Before purification, PCB 118 contained 85 ng toxic equivalents (TEQ)/g and PCB 156 contained 201 ng TEQ/g. The final toxic equivalent (TEQ) contributions of impurities were 6.6 ng TEQ/g (PCB 118), 36 ng TEQ/g (PCB 156), and 0.41 ng TEQ/g (PCB 153), levels we considered to have no influence on the final outcome of our results. PCBs were dissolved in corn oil after purification. All tested congeners were further diluted in corn oil (Sigma-Aldrich) at the Institute for Risk Assessment Sciences, Utrecht University). 34


Intake and systemic REPs of DLCs in C57BL/6 mice

Animals Eight-week-old female C57BL/6 mice (Harlan laboratories, Venray, the Netherlands) were randomly assigned to treatment groups (six animals per group) and allowed to acclimate for 1.5 weeks. The animals were housed in groups in standard cages and conditions (23 ± 2°C, 50–60% relative humidity, 12 hr dark/light cycle) with free access to food and water. Mice received a single dose of test compound by oral gavage at a dosing volume of 10 mL/kg body weight (BW). Mice treated with corn oil vehicle (10 mL/kg BW) served as controls. For each congener, five different doses were administered, ranging from 0.5–100 mg/kg BW (TCDD) to 5,000–500,000 μg/kg BW (PCB 153). Detailed information on doses is provided in Supplemental Material, Table S1. On day 3 after dosing, animals were euthanized by CO2/O2 asphyxiation, and blood was immediately collected from the abdominal aorta. The liver, thymus, spleen, and adipose tissue were removed, weighed (liver, thymus, and spleen), snap frozen, and stored at –80°C. All animal treatments were performed with permission of the Animal Ethical Committee (DEC Utrecht) and performed according to the Law on Animal Experiments (1977). Animals were treated humanely and with regard for alleviation of suffering.

Compound analysis Adipose and liver tissues samples were homogenized in Na2SO4, followed by extraction and clean-up in one step, and then eluted with 200 mL 1:1 hexane:dichloromethane on an open column packed with 40% wt/wt H2SO4-impregnated silica and KOH-silica. Blood plasma samples were extracted on an open column using Chem-Elut (Agilent Technologies, Santa Clara, CA, USA) and then NaCl eluted with 75 mL 3:2 hexane:2propanol. Clean-up was performed using a miniaturized silica column (as described above), and samples were eluted using 30 mL hexane. Because the samples typically contained high levels of the analytes, only a small fraction was evaporated and analyzed. Prior to evaporation, we spiked a fraction of the samples with 13C-labeled standards. We checked potential loss of analytes during extraction and clean-up by reextracting the samples using the identical protocol used for the samples. This procedure indicated that the losses that occurred during this first step were minor, and thus most likely do not significantly contribute to the measured outcomes. Tetradecane was added prior to evaporation. Sample analysis followed the U.S. Environmental Protection Agency Method 1613 (U.S. Environmental Protection Agency 1994) using single ion monitoring mode on an Agilent 6809N gas chromatograph (Agilent Technologies) coupled with a Micromass Ultima Autospec Ultra high-resolution mass spectrometer (HRMS; Waters Corp., Milford, MA, USA). Compounds were separated on a 60 m x 0.25 mm DB5-MS column (0.25 mM; J&W Scientific, Folsom, CA, USA). The HRMS was operated with electron impact ionization with electron energy of 35 eV and an ion source temperature of 250°C. To reduce the number of analyses, samples were pooled before clean-up. To 35

2


retain unique individual results, liver, adipose, and plasma samples were not pooled within the same treatment group of one congener, but between similar exposure levels of TCDD, PeCDD, 4-PeCDF, and PCB 126 or PCB 118, PCB 156, and PCB 153. This method was used because full congener–specific separation could be achieved on the highresolution GC–HRMS. For lipid determination, samples were evaporated to dryness after the extraction step, and the amount of lipids was determined gravimetrically. Concentrations were calculated based on lipid weight and wet weight. The analysis of samples for the PCB 118 5,000 mg/kg BW dose failed during the procedure; thus analysis for this group could not be completed.

Plasma and PBL isolation Blood from two mice was pooled (total volume of approximately 1.4 mL); plasma and PBLs were then isolated using Ficoll Paque gradient (GE Healthcare Europe, Diegem, Belgium). Plasma samples were stored at –80°C until compound analysis. Isolated lymphocytes were lysed with RLT buffer (QIAGEN, Venlo, the Netherlands) as described in the QIAGEN RNAeasy kit protocol and stored at –80°C until use. EROD activity We determined hepatic CYP1A1 activity using ethoxyresorufin-O-deethylase (EROD) activity in hepatic microsomal fractions as described by Schulz et al. (2012).

RNA isolation and quantitative real-time polymerase chain reaction (PCR) Total RNA was isolated from liver and PBLs using a QIAGEN RNeasy kit (QIAGEN). Purity and concentration of the isolated RNA was determined by measuring the absorbance ratio at 260/280 nm and 230/260 nm with a Nanodrop 2000 spectrophotometer (Thermo Scientific, Asheville, NC, USA). RNA was reverse transcribed to complementary DNA (cDNA) using the iScript cDNA synthesis Kit (Bio-Rad, Veenendaal, the Netherlands). Quantitative real-time PCR analyses were performed using the iQ Real-Time PCR Detection System with SYBR green (Bio-Rad). Amplification reactions were set up with 15 mL mastermix containing 12.5 mL iQ SYBR Green Supermix (Bio-rad), 0.5 mL distilled H2O, 1 mL (10 mM) forward primer, 1 mL (10 mM) reverse primer, and 10 mL first strand cDNA (10X diluted). Primer sequences were as follows: Cyp1a1: forward5´-GGTT AACC ATGA CCGG GAAC T-3´ and reverse- 5´-TGCC CAAA CCAA AGAG AGTG A-3´ (Schulz et al., 2012); Cyp1a2: forward-5´- ACATT CCCA AGGA GCGC TGTA TCT-3´ and reverse-5´-GTCG ATGG CCGA GTTG TTAT TGGT-3´ (Flaveny et al., 2010); Cyp1b1: forward-5´-GTGG CTGC TCAT CCTC TTTA CC-3´ and reverse-5´-CCCA CAAC CTGG TCCA ACTC-3´ (Berge et al., 2004); β-actin: forward-5´-ATGC TCCC CGGG CTGT AT-3´ and reverse-5´-CATA GGAG TCCT TCTG ACCC ATTC-3´ (Schulz et al., 2012). All primers were run through the National Center for Biotechnology Information Primer-BLAST database 36


Intake and systemic REPs of DLCs in C57BL/6 mice

(http://www.ncbi.nlm.nih.gov/tools/primer-blast/) to confirm specificity and validate for optimal annealing temperature (60°C for all primers) and efficiency. The efficiency of all primer pairs was 98–102% (tested at 60°C). The following program was used for denaturation and amplification of cDNA: 3 min at 95°C, followed by 40 cycles of 15 sec at 95°C and 45 sec at 60°C. Gene expression for each sample was expressed as threshold cycle (Ct), normalized to the reference gene β-actin (ΔCt). We calculated fold induction relative to the control group. Data analysis We obtained concentration–response curves using a sigmoidal dose–response nonlinear regression curve fit with variable slope (GraphPad Prism 6.01; GraphPad Software Inc., San Diego, CA, USA): [1]

In this Hill equation, y is the dependent variable (EROD activity or fold induction of mRNA levels), x the independent variable (administered or systemic dose), E0 is the estimated background response level, Emax is the maximum response, b is the estimated median effective concentration (EC50), and n is the shaping parameter of the Hill curve. We calculated the potency of a congener relative to TCDD using the dose or concentration [benchmark response (BMR)] needed for a congener to reach 20% of the TCDD response (BMR20TCDD). Using the congener-specific BMR20TCDD concentration, REPs were calculated relatively to TCDD:

[2]

Statistical analysis Statistical significant differences of the means and variances were determined using one-way analysis of variance (ANOVA) followed by Tukey-Kramer multiple comparisons. Differences were considered statistically significant at p < 0.05. Statistical calculations were performed using GraphPad 6.01 (GraphPad Software Inc.). Results Effect on body and organ weight To evaluate the possible toxic effects of the congeners tested, we examined body and 37

2


organ weights. We observed no changes in body weight in congener-treated mice compared with vehicle controls. Relative thymus weights showed a decreasing trend for all compounds except PCB 126; however, this decrease was statistically significantly different from the vehicle controls only in mice treated with TCDD (≥ 2.5 μg/kg BW), PeCDD (0.5, 10, and 100 μg/kg BW), and PCB 153 (500,000 μg/kg BW). We also observed a dose-dependent increasing trend in liver weight for all compounds, but this increase was significantly different from vehicle controls only at doses of ≥ 10 μg/kg BW (TCDD), ≥ 100 μg/kg BW (PeCDD), ≥ 100 μg/kg BW (4-PeCDF), ≥ 1,000 μg/kg BW (PCB 126), ≥ 150,000 μg/kg BW (PCB 118), ≥ 50,000 μg/kg BW (PCB 156), and ≥ 500,000 μg/kg BW (PCB 153). In addition, we observed a dose-dependent increasing trend in hepatic lipid content of mice treated with all compounds except PCB 153, compared with vehicle controls. No statistically significant changes in spleen weight were observed for any of the compounds tested. Additional information is provided in Supplemental Material, Table S2.

Distribution of the compounds To calculate systemicREPs, we analyzed liver, adipose, and plasma concentrations of the test compounds (see Supplemental Material, Table S3). Within the 3-day period between dosing and sacrifice, concentrations of all congeners increased linearly with the administered dose (Figure 1), which indicates an absence of autoinduction of metabolism for the different dose levels within this time period.

On a wet weight basis (nanograms per gram of tissue), concentrations of TCDD, PeCDD, 4-PeCDF, and PCB 126 were higher in the liver than in adipose tissue (see Supplemental Material, Table S3). In contrast, concentrations of the mono-ortho PCBs 156 and 118, and the non-dioxin-like PCB 153 were lower in liver than in adipose tissue. These differences were even more pronounced when concentrations were expressed as percent of dose per gram of tissue. Thus, the more potent DLCs had a higher liver affinity than the less potent PCBs 118 and 156. Therefore, we determined the ratio between liver and adipose tissue concentrations to study congener-specific hepatic sequestration. Diliberto et al. (1997) previously suggested that a liver:adipose ratio > 0.3 reflects congener-specific hepatic sequestration. In our study, we observed liver:adipose ratios > 0.3 for TCDD, PeCDD, 4-PeCDF, and PCB 126 but liver:adipose ratios < 0.3 for PCBs 118, 156, and 153 (Table 1). Hepatic sequestration was dose dependent for TCDD and PCB 126 (as shown by increasing liver:adipose ratios at higher dose levels) but not for PeCDD and 4-PeCDF.

38


Intake and systemic REPs of DLCs in C57BL/6 mice

Systemic concentration (ng/g tissue)

1000000

TCDD PeCDD

100000 10000

4-PeCDF PCB-126

1000

PCB-156

PCB-118 PCB-153

100 10 1 0.1 10 -1

10 0

10 1

10 2

10 3

10 4

10 5

10 6

Oral dose (g/kg bw) Figure 1. Relation between oral dose and mean systemic concentration in liver (—) or adipose tissue (---) of female C57BL/6 mice 3 days after administration of a single dose of TCDD, PeCDD, 4-PeCDF, PCB-126, PCB-118, PCB-156 and PCB-153. Data represent mean ± SD of 6 mice.

Dose-response curves We used tissue and plasma concentrations to determine dose-response relationships of hepatic EROD activity and gene expression of Cyp1a1, 1b1, and 1a2 in liver and PBLs (see Supplemental Material, Figure S1). All compounds except PCB 153 caused a statistically significant, dose-dependent increase in hepatic EROD activity and in Cyp1a1 and 1a2 mRNA levels. Hepatic Cyp1b1 mRNA expression was dose-dependently increased by TCDD, PCDD, 4-PeCDF, and PCB 156. We observed a dose-dependent trend for PCB 118; however, the maximum induction for PCB 118 was < 0.3% of the maximal response of TCDD. PCB 126 did not induce Cyp1b1 mRNA levels in the liver. In PBLs, Cyp1a1 mRNA levels were dose-dependently induced by all compounds except PCB 118 and PCB 153. Cyp1b1 mRNA was statically significantly and dose-dependently induced by TCDD, PeCDD, and 4-PeCDF. PCB 126 induced Cyp1b1 mRNA only at the highest dose tested, with 3.5% of the maximal induction of TCDD. PCB 118, PCB 156, and PCB 153 did not induce Cyp1b1 mRNA levels in PBLs, and Cyp1a2 mRNA was not expressed in PBLs. For all DLCs, a maximum induction (Ymax) was reached only for hepatic EROD activity but not for Cyp1a1, 1b1, or 1a2 mRNA in the liver and PBLs, even at the highest doses tested. Furthermore, we observed differences in curve Hill slopes between congeners for all end points tested (see Supplemental Material, Figure S1). Dose–response curves of Cyp1a1 mRNA in liver and PBLs based on administered dose or on liver or plasma 39

2


concentration are provided in Supplemental Material, Figure S2. Congener-specific differences in Ymax and Hill slopes can add a significant uncertainty in calculating EC50 values that generally form the basis of REP determination. To reduce this uncertainty, we focused on the lower part of the dose–response curves (BMR20TCDD) as a comparative end point (see Supplemental Material, Figures S1 and S2). Table 1: Liver:adipose concentration ratios. Congener TCDD

Dose

µg/kg BW 0.5

liver:adipose ratio

1.8 ± 0.2

2.5

2.9 ± 0.5*

2.5

7.0 ± 1.1*

10

4.2 ± 0.7*

PeCDD

0.5

4-PeCDF

5

11.5 ± 1.7

PCB-126

5

3.2 ± 0.3

PCB-118

10

25

100

25

100

15000

50000

4.4 ± 0.9

6.7 ± 1.4

13.2 ± 1.5 13.3 ± 2.6 5.9 ± 0.9*

9.1 ± 0.9*

0.08 ± 0.01

0.07 ± 0.02

PCB-156

5000

0.09 ± 0.02

PCB-153

5000

0.08 ± 0.02

15000 50000

15000 50000

0.11 ± 0.03 0.12 ± 0.02

0.11 ± 0.02 0.08 ± 0.03

Data represents the mean ± SD (based on ng/g tissue) of 6 mice. * p < 0.05 compared with the next lower dose, determined by one-way ANOVA followed by Tukey’s multiple comparisons test.

BMR20TCDD concentrations and REPs BMR20TCDD values for hepatic end points were calculated based on administered dose and on hepatic, adipose, or plasma concentration, whereas BMR20TCDD for PBL end points were calculated using only the administered dose or plasma concentration. The administered dose or systemic levels needed for a congener to reach the BMR20TCDD varied 40


Intake and systemic REPs of DLCs in C57BL/6 mice

strongly between end points, but also between the liver and PBLs (Table 2). Compared with liver, a higher concentration was usually needed in PBLs to reach a BMR20TCDD for the same end point. In the liver, EROD activity was the most sensitive biomarker for TCDD, PeCDD, 4-PeCDF, and PCB 126 exposure, followed by Cyp1a1 and Cyp1a2 mRNA induction. In contrast, hepatic Cyp1a2 mRNA induction appeared to be the most sensitive biomarker for PCB 118 and PCB 156, followed by EROD activity and Cyp1a1 gene expression. In PBLs in the TCDD group, the BMR20TCDD for Cyp1a1 and Cyp1b1 were similar. In contrast, for PeCDD and 4-PeCDF, the BMR20TCDD of Cyp1b1 expression was at least twice that of Cyp1a1 gene expression. In Figure 2, we present an overview of the REP differences based on liver, adipose, and plasma concentrations. A BMR20TCDD was not reached for all congeners or end points studied; thus, these data were excluded from the REP calculations. For comparison of congener-specific REPs across exposure matrices (intake, liver, adipose, or plasma), the intakeREP was set to 1 and deviations were calculated for various systemic REPs with the same end point (Figure 2). We observed two different types of deviations between systemicREPs and intakeREPs. Based on liver concentrations (wet weight or lipid weight), systemicREPs of PeCDD, 4-PCDF, and PCB 126 were at most one-third of the intakeREPs. In contrast, systemicREPs of PCB 118 and PCB 156 are up to one order of magnitude higher than their intakeREPs. When systemicREPs for hepatic effects of PeCDD, 4-PeCDF, and PCB 126 were calculated using adipose tissue and plasma concentrations, systemic REPs were up to one order of magnitude higher than intakeREPs, depending on the end point studied. We found the opposite for the systemicREPs of PCB 118 and PCB 156, which were at most one-third of the intakeREPs. In PBLs, systemicREPs based on plasma concentrations also deviated from intakeREPs, in a manner similar to that of systemicREPs of hepatic end points based on plasma concentration. These two different types of deviations from intakeREPs that we found for systemicREPs differentiate the more potent AhR agonists (PeCDD, 4-PeCDF, and PCB 126) from the less potent mono-ortho PCBs (PCB 118 and PCB 156). In both groups, systemicREPs can differ as much as one order of magnitude from the intakeREPs (Figure 2).

41

2


Table 2: Mean BMR20TCDD concentrations for TCDD, PeCDD, 4-PeCDF, PCB-126, PCB-118 and PCB-156 and corresponding REPs for various endpoints in liver and PBLs Biomarker

Dose metric

TCDD

Liver

Adm. dose (µg/kg bw)

0.29

EROD activity

Liver

Cyp1a1 mRNA

Liver

Cyp1b1 mRNA

Liver

Cyp1a2 mRNA

PBLs

Sys. liver (ng/g liver)

Sys. liver (ng/g lipid)

Sys. adipose (ng/g lipid)

Sys. plasma (ng/g lipid)

Adm. dose (µg/kg bw) Sys. liver (ng/g liver)

Sys. liver (ng/g lipid)

Sys. adipose (ng/g lipid)

BMR20TCDD

1.61

34.6

1.23

1.38 0.64 4.35

77.5

2.50

1

1

1 1 1

1

1

0.54

4.85

99.6

1.25

2.31 1.25 12.0

216

2.36

REP

0.5 0.3

0.3

1

0.6 0.5 0.4

0.4

1

BMR20TCDD

4.11 32.9

913

3.47

3.50 81.3 725

13768

59.8

1

3.94

0.7

37.6

Sys. liver (ng/g lipid)

391

1

1655

0.2

32921

Adm. dose (µg/kg bw) Sys. liver (ng/g liver)

Sys. adipose (ng/g lipid)

Sys. plasma (ng/g lipid)

Adm. dose (µg/kg bw)

Sys. liver (ng/g liver)

Sys. liver (ng/g lipid)

Sys. adipose (ng/g lipid)

Sys. plasma (ng/g lipid) Adm. dose (µg/kg bw)

3.55

29.1

10.3

19.4

51.1

1

95.3

0.41

2.59

1.73

1.85 22.4

ND

Sys. plasma (ng/g lipid)

1

105

18.5

Adm. dose (µg/kg bw)

Sys. plasma (ng/g lipid)

1

10.1

1

PBLs

Adm. dose (µg/kg bw)

1

11.6

50.6

Cyp1a2 RNA

1

BMR20TCDD

2.66

Sys. plasma (ng/g lipid)

Cyp1b1 mRNA

1

4-PeCDF

Sys. plasma (ng/g lipid)

Cyp1a1 mRNA PBLs

REP

PeCDD

20.9

53.5 ND

1 1

1

1 1

1

1

1

0.56 4.59

1.20

0.4

0.3

0.5

0.6 0.7

0.6

0.5

1

150

1577

461

44.6 8.83

68.1

1712

6.53

2.36

0.8

8.01

51.8

0.4

514

33.7

34.0 63.8 ND

ND

0.7 1.5 0.8

117

40.8 212 ND

ND

Abbreviations: Adm, administered; ND, not determined because BMR20TCDD was not reached; Syst, systemic. Data are expressed as mean BMR20TCDD derived from dose-response curves of 6 mice. REPs were calculated as described in “Materials and Methods”.

42


Intake and systemic REPs of DLCs in C57BL/6 mice

Table 2: Mean BMRconcentrations for TCDD, PeCDD, 4-PeCDF, PCB-126, PCB-118 and PB-1and corresponding REPs for various endpoints in liver and PBLs REP

0.07

0.05

0.04

0.4

0.4

0.008

0.006

0.006

0.04

0.07

0.02

0.02

0.01

0.02

0.3

0.05

0.04

0.03

0.3

0.2 0.2 1

0.04 0.3

PCB-126

BMR20TCDD 29.3

373

9938

72.7

72.3 558

4299

70368

315

0.32 ND

ND

REP

0.01

0.004

0.003

0.02

0.02

0.001

0.001

0.001

0.008

0.008

ND

ND

ND

87.4

912

21240

120

135 603

847 ND ND

ND ND

PCB-118

BMR20TCDD 55259

25441

720241

359114

311118 139631

62418

1693882

ND

803766 ND

ND

REP

0.000005

0.00006

0.00006

0.000003

0.000004 0.000005

0.00007

0.00005

0.000003

ND

0.005

0.003

0.002

0.01

0.01 0.04

0.06

ND

ND

15522

8833

267405

117517

103230 ND

ND

ND ND ND ND

PCB-156

BMR20TCDD 15664

7501

217711

82483

98188 44305

35669

634215

180515

303586 95664

72446

1158251

0.00003

0.0003

0.0002

0.00001

0.00002

745126

553459

12085

4239

166060

22134

60702

747734

2359081 ND ND ND ND

REP

0.00002

0.0002

0.0002

0.00001

0.00001

0.00001

0.0001

0.0001

0.00001

0.000009

0.00004

0.0004

0.0003

0.00001

0.00002

0.00003

0.0006

0.0003

0.00008

0.00003 0.00003 0.00002

       

43

2




 

    

   . .  . .  . /

 

   

    

    

    

    

    

!



!

!

  



!



  

    





 

   

  



!





 

    

    



!

!



!

   

    

!





    



  

   

   

    

!

  



 

   

!

 

!

  

!



!

    



!



!

  



 

   





    





    

    

!

44

, (*()#&$"    $')!%#   +0 + -./2+ -..3  -.12,              . .+  . .  . /   ,4,



   





!



!

  



!


Discussion

Intake and systemic REPs of DLCs in C57BL/6 mice

The TEF approach is the most commonly used method of assessing the risk of complex mixtures of dioxins and DLCs. Current TEF values are derived mainly from a range of intakeREPs, preferably from (sub) chronic in vivo studies. These intakeREPs link the administered dose to a toxic or biological effect, subsequently leading to the derivation of intakeTEFs (Van Den Berg et al., 1998; 2006).

At present, available data are insufficient to establish whether or not intakeTEFs are valid for risk assessment based on plasma or adipose tissue concentrations. Thus far, the limited experimental evidence available suggests that systemicREPs of DLCs may differ from intakeREPs (Budinsky et al., 2006; DeVito et al., 1997; 2000). This discrepancy originates most likely from toxicokinetic differences between various DLCs. Several studies have shown that many DLCs bind strongly to CYP1A2 protein and, as a result, strongly sequester in the rodent liver (Devito et al., 1998; Diliberto et al., 1995; 1997; 1999). This binding affinity toward CYP1A2 influences the hepatic, plasma, and adipose tissue disposition of DLCs. This was confirmed using CYP1A2 knockout mice in which the liver:adipose ratio decreased to < 0.3 for TCDD and 4-PeCDF, which is indicative of no hepatic sequestration (Diliberto et al., 1997). These ratios are significantly lower than those we observed in the present study for both congeners (Table 1). It is worth nothing that the dose dependency and hepatic sequestration we observed in our single dose, 3-day study are similar for all tested compounds, except for 4-PeCDF at the two highest concentrations tested, to those observed by DeVito et al. (1998) in a multiple dose, subchronic 13-week study of female B6C3F1 mice. In addition, the TCDD EC50 systemic liver concentrations for hepatic EROD activity were similar. Comparable findings can also be expected for the other DLCs tested because metabolism and elimination of these compounds are very similar. In light of the similarities between results of the two studies, we assume that intakeREPs and systemicREPs do not deviate over time, even when they have not reached a steady state. In the present study, intakeREPs and systemic REPs for Cyp1a1, 1a2, and 1b1 induction were determined 3 days after a single oral dose. Previous studies have shown that hepatic CYP1A1, 1A2, and 1B1 protein levels are already maximal in rats 3 days after a single dose of TCDD (Santostefano et al., 1997). Although induction of CYP1A1, 1A2, and 1B1 enzymes is not a measure of toxicity, this is considered to be the most sensitive biomarker for AHR activation (Abel and HaarmannStemmann, 2010; Denison and Heath-Pagliuso, 1998). Moreover, studies have shown a high correlation in REPs between induction of these enzymes and toxic responses inflicted by DLCs, such as wasting syndrome, thymic atrophy, or hepatic porphyrin accumulation (Budinsky et al., 2006; van Birgelen et al., 1996). Similar to earlier studies, we observed distinct deviations between intakeREPs and systemicREPs based on liver, plasma, or adipose tissue concentrations (Budinsky et al., 2006; Devito and Birnbaum, 45

2


1995; DeVito et al., 1997; 2000). We observed congener-specific differences between the potent PeCDD, 4-PeCDF, and PCB 126 versus the less potent mono-ortho PCBs, PCB 118 and PCB 156 (Figure 2). On the basis of the liver:adipose ratios established in our study (Table 1), it appears that these congener-specific differences have a toxicokinetic basis, in which hepatic sequestration due to CYP1A2 binding plays a significant role. It is unclear whether a CYP1A2-sequestered compound is bioavailable to activate the AhR and cause dioxin-like responses. For this reason, REPs calculated on total hepatic tissue concentration, instead of the “free” available concentrations, may lead to either an overor underestimation of the potency of a congener, depending on the relative degree of hepatic sequestration compared with that of TCDD. The systemicREPs based on plasma concentrations for Cyp1a1 and 1b1 gene expression in PBLs and liver show similar deviations from intakeREPs for all DLCs tested. The systemicREPs are sometimes more than half a log unit different from the intakeREPs, which is more than the assumed uncertainty range applied to the WHO-TEF values (Van den Berg et al., 2006). To further address this issue, we compared intakeREPs and systemicREPs from the present study with existing WHO-TEF values and the half log uncertainty around that value (Figure 3). On the basis of this comparison, we observed that: •

• •

REPs of PeCDD fall mostly within the uncertainty range of the WHO-TEF of 1, with no large difference between systemicREPs and intakeREPs. Based on the intake dose and hepatic concentrations, deviations from the half log unit uncertainty are observed for 4-PeCDF, but systemicREPs based on plasma concentrations are close to the WHO-TEF of 0.3. For PCB 126, intakeREPs and systemicREPs are up to two orders of magnitude below the WHO-TEF value of 0.1. Of all end points studied, only Cyp1a1 mRNA expression in PBLs falls within the half log unit uncertainty. REPs based on intake dose and plasma concentrations for mono-ortho PCBs 118 and 156 are consistently lower than the WHO-TEFs of 0.00003. In contrast, REPs based on liver effects and concentrations are significantly higher than the WHOTEFs for both PCBs. However, because of differences in Cyp1a2 sequestration between the mono-ortho PCBs and the reference compound TCDD, caution should be taken not to over interpret these liver-based systemicREPs.

Most REPs determined in the present study are significantly lower than those established by the WHO (Van den Berg et al., 2006). However, the WHO-TEFs were derived from a range of intakeREPs often involving (semi)chronic studies and different species, whereas our study involves a single-dose exposure with relatively acute effects after 3 days only in mice. In the present study, we did not aim to recalculate or debate the current WHO-TEFs or methodology. However, the current WHO-TEF concept is based on the 46


0.0001

0.001

0.01

0.1

1

10

Intake

Liver

PeCDD

Plasma Intake

Liver

4-PeCDF Plasma Intake

Liver

PCB-126 Plasma

Relative Effect Potencies (REPs) 0.000001

0.00001

0.0001

0.001

Intake

Liver

PCB-118 Plasma Intake

Liver

PCB-156 Plasma

Figure 3. Relative effect potencies (REPs) determined in this study in relation to the WHO-TEF ± half log uncertainty range. REPs were determined for hepatic EROD activity ( T ), hepatic gene expression of Cyp1a1 ( ¢ ), Cyp1b1 ( ◧ ), Cyp1a2 ( o ), and gene expression of Cyp1a1 ( ) and Cyp1b1 ( ◐ ) in PBLs of PeCDD, 4-PeCDF, PCB-126 (left graph) and PCB-118 and PCB-156 (right graph). REPs for hepatic endpoints were calculated based on administered dose (Intake), lipid-based liver concentration (Liver) or lipid-based plasma concentration (Plasma), whereas for PBL, REPs were calculated using the administered dose or plasma concentration. The black line represents the mean of the REPs. The black dotted line together with its grey area represents the WHO-TEF ± half log uncertainty range.

Relative Effect Potencies (REPs)

Figure 3:

Intake and systemic REPs of DLCs in C57BL/6 mice

2

47


assumption that intakeREPs represent systemicREPs, but a full data set to reject or accept this assumption is lacking. In our study, we compared intakeREPs with systemicREPs obtained from a mouse model to provide more knowledge about possible deviations between both types of REPs. More data, for example, additional in vivo rat data and human in vitro data from our EU-SYSTEQ project studies, may provide additional information with respect to deviation of the intakeREPs and systemicREPs from our studies with current WHOTEF values. With these additional data, it can then be discussed whether systemicREPs would better reflect a risk than intake WHO-TEFs. Conclusions There are significant differences between intakeREPs and systemicREPs for hepatic EROD activity and Cyp1a1, 1a2, and 1b1 gene expression in the liver and PBLs. To avoid flawed calculations due to, for example, congener-specific hepatic sequestration, it may be more appropriate to use blood or adipose tissue as a matrix to calculate systemicREPs. The systemicREPs based on plasma/adipose concentration in our study are sometimes more than half a log unit different from the intakeREPs. This suggests that using intakeREPs or intakeTEFs to calculate TEQs in blood for PeCDD, 4-PeCDF, and PCB 126 result in an underestimation of the risk. In contrast, using intakeREPs or intakeTEFs for the mono-ortho PCBs 118 and 156 to calculate blood TEQs in blood may lead to an overestimation of the risk. Overall, our comparison of intakeREPs and systemicREPs in mice reveals significant congener-specific differences that warrants the development of systemicTEFs to calculate TEQs in blood and body tissues.

48


Intake and systemic REPs of DLCs in C57BL/6 mice - Supplemental Material

supplemental material Table S1: Congeners, TEF-values and dose ranges Congener

TEF

Single oral dose (µg/kg bw)

4-PeCDF

0.3

5

250

PCB-156

0.00003

5000

TCDD

PeCDD

PCB-126 PCB-118 PCB-153

1 1

0.1

0.00003 ND

0.5 0.5

5

2.5

10

25

100

25

100

250

1000

2.5

25

10

100

25

100

5000

15000

50000

150000

500000

5000

15000

50000

150000

500000

15000

50000

150000

2

1000

500000

49


50

PCB-126

4-PeCDF

PeCDD

TCDD

Congener

19.4 ±

18.8 ±

100

18.9 ±

19.3 ±

20.0 ±

25

5

0

20.2 ±

19.8 ±

19.2 ±

19.0 ±

19.3 ±

20.0 ±

19.4 ±

18.9 ±

1000

250

100

25

5

0

25

10

18.6 ±

19.3 ±

2.5

0.5

0

19.4 ±

19.6 ±

18.8 ±

19.2 ±

19.3 ±

19.3 ±

1.4

0.9

0.5

1.1

1.0

1.0

1.1

1.2

0.5

1.1

1.5

1.2

1.1

1.3

1.0

0.6

0.8

1.1

1.3

0.8

1.0

(gram)a

Body weight

100

25

10

2.5

0.5

0

µg/kg bw

Oral dose

0.29 ±

0.27 ±

0.28 ±

0.22 ±

0.23 ±

0.22 ±

0.24 ±

0.26 ±

0.28 ±

0.24 ±

0.26 ±

0.23 ±

0.29 ±

0.21 ±

0.32 ±

0.19 ±

0.23 ±

0.18 ±

0.23 ±

0.28 ±

0.32 ±

6.09 ± 0.44

c

0.05

0.06

0.07

0.04

0.03

0.04

0.04

0.07

0.07

c

c

4.96 ± 0.35

4.97 ± 0.38

5.01 ± 0.25

6.35 ± 0.30

6.05 ± 0.42

5.94 ± 0.27cd

5.32 ± 0.19

5.37 ± 0.32

5.01 ± 0.25

5.83 ± 0.50c

5.26 ± 0.27

5.32 ± 0.43

5.02 ± 0.26

4.56 ± 0.34

4.88 ± 0.18

c

5.86 ± 0.22 c

c

c

0.05c

0.05

0.06

0.04

0.30 ±

0.33 ±

0.33 ±

0.32 ±

0.32 ±

0.34 ±

0.32 ±

0.33 ±

0.29 ±

0.32 ±

0.33 ±

0.32 ±

0.32 ±

0.31 ±

0.30 ±

0.34 ±

0.34 ±

0.34 ±

0.32 ±

0.34 ±

5.49 ± 0.37c

5.08 ± 0.27

0.34 ±

0.06

0.04

0.03

0.02

0.04

0.03

0.04

0.05

0.03

0.04

0.04

0.04

0.03

0.02

0.03

0.02

0.02

0.04

0.05

0.05

0.03

Spleen

% of bwa

4.88 ± 0.18

5.37 ± 0.20

0.03cd

0.03

0.04

0.02

Liver

% of bwa

c

0.06c

0.05

0.06

0.03

Thymus

% of bwa

Table S2: Body weight, relative thymus, liver and spleen weights and % lipid/g liver.

3.7

3.4

3.1

NA

NA

4.5

4.5

3.6

3.1

NA

NA

6.2

6.0

4.7

4.2

NA

NA

7.4

7.2

5.3

4.2

g liverb

% lipid /


0

19.4 ±

18.7 ±

18.8 ±

500000

150000

50000

15000

0

18.8 ±

19.9 ±

19.6 ±

19.1 ±

19.7 ±

18.7 ±

5000

500000

19.2 ±

150000

18.0 ±

17.7 ±

19.4 ±

15000

50000

5000

0

19.1 ±

19.2 ±

18.0 ±

18.3 ±

18.3 ±

500000

150000

50000

15000

5000

18.8 ±

19.3 ±

19.5 ±

0.7

1.1

0.8

0.5

1.1

0.8

0.6

1.1

0.9

1.3

0.7

0.9

1.0

0.8

1.8

0.8

0.9

0.9

0.9

1.3

0.7

0.24 ±

0.25 ±

0.29 ±

0.26 ±

0.25 ±

0.31 ±

0.18 ±

0.21 ±

0.25 ±

0.23 ±

0.21 ±

0.26 ±

0.23 ±

0.21 ±

0.20 ±

0.23 ±

0.27 ±

0.26 ±

0.27 ±

0.26 ±

0.25 ±

0.04c

0.04

0.03

0.04

0.05

0.05

0.05

0.04

0.02

0.04

0.05

0.07

0.04

0.05

0.05

0.04

0.03

0.07

0.04

0.04

0.06

cd

c

cd

cd

5.80 ± 0.30c

5.47 ± 0.37

5.15 ± 0.38

4.96 ± 0.10

4.71 ± 0.34

5.11 ± 0.27

8.10 ± 0.17

6.79 ± 0.51

5.96 ± 0.62cd

5.14 ± 0.41

4.71 ± 0.43

5.15 ± 0.30

7.23 ± 0.29cd

6.14 ± 0.45

4.87 ± 0.41

5.06 ± 0.22

5.04 ± 0.22

5.15 ± 0.30

5.64 ± 0.20

5.43 ± 0.19

5.07 ± 0.19

0.33 ±

0.34 ±

0.33 ±

0.36 ±

0.30 ±

0.33 ±

0.33 ±

0.29 ±

0.31 ±

0.30 ±

0.27 ±

0.31 ±

0.31 ±

0.34 ±

0.30 ±

0.33 ±

0.32 ±

0.31 ±

0.39 ±

0.32 ±

0.31 ±

0.05

0.05

0.04

0.04

0.04

0.04

0.03

0.02

0.04

0.03

0.03

0.05

0.03

0.04

0.03

0.05

0.03

0.05

0.16

0.02

0.04

NA

NA

3.4

3.9

3.3

3.7

NA

6.8

5.7

3.3

3.0

2.7

NA

4.6

3.5

3.3

2.7

2.7

NA

NA

4.4

b

a

Data represents the mean ± SD of 6 mice Data represents the % lipid per gram of pooled liver samples from 6 mice Statistically significant changes were determined by one-way ANOVA analysis followed by a Tukey’s multiple comparisons test, cSignificantly different from control group (p<0.05). d Significantly different from previous concentration (p<0.05) NA = not analysed

PCB-153

PCB-156

PCB-118

1000

250

100

Intake and systemic REPs of DLCs in C57BL/6 mice - Supplemental Material

51

2


Table S3: PCDD / PCDF / PCBs concentration in liver, adipose tissue and plasma 3 days after a single oral dose Congener

TCDD

PeCDD

4-PeCDF

Oral dose

µg/kg bw 0.5 2.5

PCB-118

3.3 ± 0.6 20 ± 6

10

85 ± 13

100

NA

25

NA

0.5

4.0 ± 0.7

10

103 ± 23

100

NA

2.5 25

5

25

100 PCB-126

ng/g tissue

250

1000

5

25

a

27 ± 6 NA

PCB-156

5000

15000

15000

0.9 ± 0.2

1669 ± 378

51

16 ± 5

54

3.8 ± 0.7 NA

NA

12 ± 3

23203 ± 4986

51

323 ± 49 NA -

21667 ± 7554

64500 ± 10055 NA

2233 ± 880

6800 ± 2656

89500 ± 1786 1783 ± 571

7500 ± 2707

50000

19833 ± 9218

500000

NA

150000

40

38

NA

38 ± 11

NA

5000

85 ± 14

446 ± 96

1119 ± 334

NA

500000

PCB-153

NA

NA

71 ± 24

41000 ± 9960

150000

21 ± 3

45

50000

6.9 ± 1.5

43

1.8 ± 0.4

20037 ± 4531

195 ± 27

7950 ± 2870

500000

40

33

902 ± 204

15000

ng/g tissuea

3.4 ± 0.7

NA

150000

277 ± 83

1152 ± 176

% dose / g tissue

39

1000

50000

62 ± 12

Adipose

1067 ± 178

1012 ± 217

5000d

ng/g lipid

a

39 ± 6

100 250

Liver

NA

4295 ± 603

8858 ± 1350

-

39

65

-

242378 ± 87490

2.6

1853448 ± 288932

2.2

622605 ± 217076 75450 ± 29735

2.2 2.2

203593 ± 79530

2.3

1567426 ± 314993

3.0

718039 ± 174429 54872 ± 17559

190840 ± 68881

585054 ± 271910

4.1 1.8

2.5 2.0

15 ± 3 NA

NA

56 ± 11

110 ± 19 NA

NA -

95790 ± 26799

333750 ± 137042 NA

NA

25217 ± 11490

65017 ± 29435

339300 ± 102498 NA

NA

23250 ± 8546

74167 ± 33252

249400 ± 100810 NA

NA

Data represents the mean ± sd of 6 mice Data represents the mean ± sd of 3 pooled blood plasma samples (1 plasma sample = plasma of 2 mice) c Data represents the outcome of a pooled blood plasma sample from 6 mice a

b

52


Intake and systemic REPs of DLCs in C57BL/6 mice - Supplemental Material

Table S3: PCDD / PCDF / PwCBs concentration in liver, adipose tissue and plasma 3 days after a single oral dose ng/g lipid 2.1 ± 0.5

a

Adipose

7.6 ± 1.7 24 ± 4

1.1 ± 0.2

4.3 ± 0.8 19 ± 6

4.0 ± 0.8 17 ± 3

85 ± 29 14 ± 4

67 ± 13

% dose / g tissue

ng/g tissue

14

0,0134

18

10

9.3

7.7 8.0

3.4

3.0

3.5 12

11

127 ± 22

5.5

-

-

103000 ± 28817

375000 ± 153981

32

33

28333 ± 12910

25

390000 ± 117813

34

69167 ± 31314

25833 ± 9496

83333 ± 37361

286667 ± 115873

22

23

25

25

Analysis failed in the sampling procedure NA = not analysed

d

0,0040c

b

c

Plasma

ng/g lipidb 2,2c 8,4

c

% dose / g tissue 0.040 0.027

0.053 ± 0,008

35 ± 6

0.027

0.432 ± 0,054

240 ± 30

0.022

0,0120c

6,3c

0.108 ± 0,005

60 ± 3 2,2

0.037

0.032 ± 0,011

19 ± 7

0.016

0.298 ± 0,039

157 ± 21

0,0192c

20c

0,0037

c

0.052 ± 0,004

c

30 ± 2

0.024

0.010

0.015

4,1c

0.008

0.074 ± 0,018

41 ± 10

0.004

1.108 ± 0,067

583 ± 35

0,0924c

66c

0,0082c

0.173 ± 0,023

87 ± 11

0.004

0.003

0.006

27c

0.035

0.249 ± 0,026

147 ± 15

0.012

2.773 ± 0,721

1733 ± 451

0.014

161 ± 15

100333 ± 9504

0.054

890000 ± 182483

0.062

0,0351c

0.654 ± 0,091 -

672 ± 157

1869 ± 383

363 ± 50 -

280000 ± 65574

5000 ± 721

2500000 ± 360555

182 ± 68

91000 ± 34395

79 ± 15

659 ± 147

1338 ± 279

0.013 -

0.067 0.05

0.079

346667 ± 77675

0.066

743333 ± 155027

2200000 ± 458258

155 ± 44

91000 ± 26211

75 ± 7

0.018

46667 ± 9074

3740 ± 779

0.061

0.045 0.037

37333 ± 3512

0.075

595 ± 273

350000 ± 160935

0.060

4180 ± 503

2200000 ± 264575

1487 ± 291

743333 ± 145717

2

0.022

0.052

0.050 0.042

53


 



  

 





    



 

  







    

 











 

    

  







    



     











 











 





 









"!&%"!" #$%""!%$"







"!&%"!" #$%""!%$"



 

"!&%"!" #$%""!%$"

 



# " ! #$"%!

 

"!&%"!" #$%""!%$"



"!&%"!" #$%""!%$"

 

  







    



     

  









    

Figure S1. Dose-response curves for EROD activity (A) and gene expression of Cyp1a1 (B), Cyp1b1 (C), Cyp1a2 (D) in mouse liver and gene expression of Cyp1a1 (E) and Cyp1b1 (F) in peripheral blood lymphocytes (PBL) three days after a single oral dose of TCDD ( ¢ ), PeCDD ( o ), 4-PeCDF ( ), PCB126 ( ), PCB-118 ( ¿ ) and PCB-156 ( ¯ ). Dose response curves are expressed using administered dose. BMR20TCDD is indicated with a black dotted line. Data are represented as mean ¹ SD (n=6).

54


Intake and systemic REPs of DLCs in C57BL/6 mice - Supplemental Material



 

 









         

   

 







 

    

  

 $#  !"#  #" 



 









 



  

  

  





 $#  !"#  #" 





 $#  !"#  #" 





 $#  !"#  #" 

  

 $#  !"#  #" 





 















         



   

 

 

 

    



    

 





  

    

Figure S2. Dose-response curves of Cyp1a1 gene expression in mouse liver and peripheral blood lymphocytes (PBL) three days after a single oral dose of TCDD ( ¢ ), PeCDD ( o ), 4-PeCDF ( ), PCB126 ( ), PCB-118 ( ¿ ) and PCB-156 ( ¯ ). Dose response curves are expressed using administered dose (A and D), plasma concentration (B and E) or liver concentration (C). BMR20TCDD is indicated with a black dotted line. Data are represented as mean ¹ SD (n=6).

55

2


Chapter

3

Comparison of Intake and Systemic Relative Effect Potencies of Dioxin-like Compounds in Female Rats after a Single Oral Dose Karin I. van Ede1 Patrik L. Andersson2 Konrad P.J. Gaisch1 Martin van den Berg1 Majorie B.M. van Duursen1 1

Endocrine toxicology group, Institute for Risk Assessment Sciences (IRAS), Utrecht University, the Netherlands 2 Department of Chemistry, Umeå University, Umeå, Sweden

Archives of Toxicology 88 (3): 637 – 646 (2014)


Abstract Risk assessment for mixtures of dioxin-like compounds uses the toxic equivalency factor (TEF) approach. Although current WHO-TEFs are mostly based on oral administration, they are commonly used to determine toxicity equivalencies (TEQs) in human blood or tissues. However, the use of “intake” TEFs to calculate systemic TEQs in for example human blood, has never been validated. In this study, intake and systemic relative effect potencies (REPs) for 1,2,3,7,8-pentachlorodibenzo-p-dioxin (PeCDD), 2,3,4,7,8-pentachlorodibenzofuran (4-PeCDF), 3,3’,4,4’,5- pentachlorobiphenyl (PCB126), 2,3’,4,4’,5-pentachlorobiphenyl (PCB-118) and 2,3,3’,4,4’,5-hexachlorobiphenyl (PCB-156) were compared in rats. The effect potencies were calculated based on administered dose and liver, adipose or plasma concentrations in female Sprague– Dawley rats 3 days after a single oral dose, relative to 2,3,7,8-tetrachlorodibenzopdioxin (TCDD). Hepatic ethoxyresorufin-O-deethylase activity and gene expression of Cyp1a1, 1a2, 1b1 and aryl hydrocarbon receptor repressor in liver and peripheral blood lymphocytes were used as endpoints. Results show that plasma-based systemic REPs were generally within a half log range around the intake REPs for all congeners tested, except for 4-PeCDF. Together with our previously reported systemic REPs from a mouse study, these data do not warrant the use of systemic REPs as systemic TEFs for human risk assessment. However, further investigation for plasma-based systemic REPs for 4-PeCDF is desirable.

58


Intake and systemic REPs of DLCs in SD rats

Introduction

P

olychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs) and polychlorinated biphenyls (PCBs) are persistent organic pollutants and commonly occur in the environment and human food chain. Human risk assessment for dioxin-like compounds (DLCs) is challenging because these compounds are present in the environment in complex mixtures. The common approach used by risk assessors is based on the toxic equivalency factor (TEF) concept (Safe, 1990; 1994a) endorsed by the World Health Organization (WHO) (Van den Berg et al., 1998; 2006). Each congener-specific TEF is an estimate of its relative potency compared to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). In total, 29 PCDDs, PCDFs and dioxin-like PCBs have been assigned with a TEF value. These TEFs are mainly derived from relative effect potencies (REPs) determined in (sub)chronic in vivo studies with the administered dose as exposure metric, resulting in “intake” TEFs (intakeTEFs) (Haws et al., 2006). However, these intakeTEFs are widely used to assess the risk of humans based on concentrations in for example blood. Thereby, it is assumed that an intakeTEF can also be applied for risk assessment when using systemic concentrations in human blood and tissues. However, differences in toxicokinetics may influence the potency of a congener when calculated on either administered dose or systemic concentrations(Budinsky et al., 2006; Devito and Birnbaum, 1995; DeVito et al., 1997; 2000). For this reason, the use of intakeTEFs to assess a possible risk based on blood or serum concentrations may potentially lead to a misinterpretation of the actual risk. Currently, there are insufficient data available to either accept or reject the use of intakeTEFs for risk assessment when based on, e.g. blood concentrations (Van den Berg et al., 2006).

Previously we described up to one order of magnitude difference between intakeREPs and systemicREPs in C57bl/6 mice for 1,2,3,7,8-pentachlorodibenzo-p-dioxin (PeCDD), 2,3,4,7,8-pentachlorodibenzofuran (4-PeCDF), 3,3’,4,4’,5-pentachlorobiphenyl (PCB126), 2,3’,4,4’,5-pentachlorobiphenyl (PCB-118) and 2,3,3’,4,4’,5-hexachlorobiphenyl (PCB-156) compared to TCDD, 3 days after a single oral dose (van Ede et al., 2013a). Based on plasma or adipose levels, higher systemicREPs were calculated for PeCDD, 4-PeCDF and PCB-126, and lower systemicREPs for the mono-ortho PCBs 118 and 156 when compared to intakeREPs. In the present study, we describe and compare intakeREPs and systemicREPs for the same congeners in female Sprague–Dawley rats, based on the administered dose or the systemic liver, adipose or plasma concentrations. Similar to our earlier study with mice, intakeREPs and systemicREPs were calculated 3 days after exposure, using sensitive biomarkers for AhR activation, e.g. Cyp1a1, 1a2, 1b1 and aryl hydrocarbon receptor repressor (Ahrr) expression and/or activity in the liver and peripheral blood lymphocytes (PBLs). The results from this study are compared with 59

3


those from our mouse study and other data from the literature that allow calculations of both intake and systemic REPs. Materials and Methods Chemicals TCDD, PeCDD, 4-PeCDF and PCB-126 were purchased from Wellington Laboratories Inc. (Guelph, Ontario, Canada). After dissolving in corn oil (ACH Food Companies Inc., Oakbrook, IL, USA), concentrations were then checked and confirmed by Wellington Laboratories Inc. PCB-118, PCB-156 and 2,2’,4,4’,5,5’-hexachlorobiphenyl (PCB-153) were purchased from Cerilliant Corp. (Round Rock, TX, USA). These PCBs and corn oil (Sigma-Aldrich, Stockholm, Sweden) were purity checked and, when necessary, purified at the Department of Chemistry, Umeå University, Umeå, Sweden. Final toxicity equivalency (TEQ) contributions of impurities were 6.6 (PCB-118), 36 (PCB-156) and 0.41 (PCB-153) ng TEQ/g. These levels were considered to have no influence on the final outcome of our results. Further dilutions of the congeners in corn oil (Sigma-Aldrich, Stockholm, Sweden) were prepared at the Institute for Risk Assessment Sciences (IRAS, Utrecht University), the Netherlands.

Animals Eight-week-old female Sprague–Dawley rats (Harlan laboratories, Venray, the Netherlands) were randomly assigned to treatment groups (6 animals/group) and allowed to acclimate for 1.5 weeks. The animals were housed in groups in standard cages and conditions (temperature 23 ± 2 °C, 50–60 % relative humidity, 12-h dark and light cycle) with free access to food and water. Rats received a single dose by oral gavage at a dosing volume of 10 ml/kg bw. Depending on the congener used, five different dose levels were administered in the range from 0.5 μg/kg bw (TCDD) up to 500 mg/kg bw (PCB-153), spanning a similar range of administered TEQ doses across congeners based on the 2006 WHO-TEF values. Detailed information regarding the administered dose levels can be found in Supplementary Material; Table S1. Animals were killed 3 days after dosing using CO2/O2. Blood was obtained from the abdominal aorta directly after killing, and liver, thymus, spleen and adipose tissue were removed, weighed (liver and thymus), snap frozen and stored until use at −80 °C. All animal treatments were performed with permission of the Animal Ethical Committee and performed according to Dutch law on Animal Experiments (http://wetten.overheid.nl/BWBR0003081). Animals were treated humanely and with regard for alleviation of suffering. 60


Intake and systemic REPs of DLCs in SD rats

Compound analysis Analysis of the compounds in blood plasma, adipose and liver tissue samples was performed as described earlier by Van Ede et al. (2013a). In short, adipose and liver tissue samples were cleaned using a combined solid-phase extraction using Na2SO4 and KOH-silica. Blood plasma samples were extracted on an open column using Chem Elut and NaCl. Clean-up was performed using a miniaturized silica column. Samples were spiked after evaporation with 13C-labelled standards. Sample analysis followed the US EPA Method 1613 (http://water.epa.gov/scitech/methods/cwa/organics/dioxins/ index.cfm) using single ion monitoring mode on an Agilent 6809 N (Agilent technologies, Santa Clara, CA, USA) gas chromatograph coupled with a Micromass Ultima Autospec Ultra high resolution mass spectrometer (HRMS, Waters Corp., Milford, MA, USA). To retain unique individual results of each animal, tissue samples (liver, plasma or adipose fat) were not pooled from various animals within the same treatment group but tissues from individual animals that were exposed to different congeners at the same dose level were pooled (TCDD + PeCDD + 4-PeCDF + PCB-126 or PCB-118 + PCB-156 + PCB-153) (See Supplementary Material; Table S1). For example, to determine liver concentrations, a liver from a rat treated with TCDD at the lowest dose was pooled with a liver from another rat treated with PeCDD, one liver from a 4-PeCDF-exposed rat and a liver from a rat dosed with PCB-126 at the lowest dose. This time and cost-effective approach was chosen because a full separation and quantification of individual congeners could be obtained in a single HRGC-HRMS run. The concentrations were calculated on lipid and wet weight basis. Plasma and peripheral blood lymphocyte (PBL) isolation From blood (approximately 7 ml) of each individual rat, plasma and PBLs were isolated using Ficoll-Paque gradient (GE Healthcare Europe, Diegem, Belgium) according to manufacturer’s instructions. Plasma samples were stored directly at −80 °C until compound analysis. Isolated lymphocytes were lysed with RLT buffer (Qiagen, Venlo, the Netherlands) as described in the Qiagen RNAeasy kit protocol and stored until use at −80 °C. EROD activity Hepatic CYP1A1 activity was determined by means of ethoxyresorufin-O-deethylase (EROD) activity in microsomal fractions of liver tissue according to Schulz et al. (2012).

RNA isolation and quantitative real‑time polymerase chain reaction (PCR) RNA isolation and quantitative real-time PCR were performed as described earlier by Van Ede et al. (2013a). Primer sequences were as follows: Cyp1a1, forward-5’-ATGTCCA GCTCTCAGATGATAAGGTC-3’ and reverse-5’-ATCCCTGCCAATCACTGTGTCTAAC-3’ 61

3


(Vondracek et al., 2006), Cyp1a2, forward-5’-GTGAGAACTACAAAGACAACGGTG-3’ and reverse-5’-GTGACTGTTTCAAATCCAGCTC C-3’ (Vondracek et al., 2006), Cyp1b1, forward-5’- CT CATCCTCTTTACCAGATACCCG-3’ and reverse-5’- GA CGTATGGTAAGTTGGGTTGGTC-3’ (Vondracek et al., 2006), Ahrr, forward-5’CCCCAAGGGGACTTCAGGG GAC-3’ and reverse-5’TGCTCCAGTCCAGGTGCC TCA-3’ [designed with the Primer designing tool (NCBI)] Arbp, forward-5’CCTAGAGGGTGTCCGCAATGTG-3’ and reverse-5’- CAGTGGGAAGGTGTAGTCAGTCTC-3’ [designed with the Primer designing tool (NCBI)]. All primers were run through National Center for Biotechnology Information (NCBI) Primer-BLAST database to confirm specificity and validated for optimal annealing temperature (60 °C for all primers) and efficiency. For Cyp1a1, Cyp1b1, Cyp1a2 and Arbp, the efficiency of the primer pairs was 98–102 % (tested at 60 °C). The Ahrr primer pair efficiency was 120 %. The following programme was used for denaturation and amplification of the cDNA: 3 min at 95 °C, followed by 40 cycles of 15 s at 95 °C and 45 s at 60 °C. Gene expression for each sample was expressed in terms of the threshold cycle (Ct), normalized to the reference gene Arbp (ΔCt). Fold induction was calculated between the treated and vehicle control groups.

Data analysis Dose–response curves, effect concentrations and REP calculations for the tested congeners were determined as described previously by Van Ede et al. (2013a). Briefly, dose–response curves were obtained using a sigmoidal dose–response nonlinear regression curve fit with variable slope (GraphPad Prism 6.01, GraphPad Software Inc., San Diego, CA). Next, REPs were calculated using a benchmark response (BMR) approach. To determine REPs, the dose or concentration needed for a congener to reach 20 % of the TCDD response (BMR20TCDD) was calculated. Using the congener specific BMR20TCDD concentration, REPs were calculated relatively to TCDD. The selection of the BMR20TCDD concentration instead of effect concentration 50 % (EC50), that generally form the basis of REP determination, was based on several arguments. Many of the obtained dose– response curves in our study did not attain a maximum efficacy or similar Hill slope. Both differences in maximum efficacy and Hill slope could add a significant uncertainty in calculating EC50 values. For this situation, it has been suggested that, e.g. LO(A)ELs or benchmark, dose levels could be used to determine REPs (Van den Berg et al., 2006). In the case of incomplete dose–response curves, also several other studies have suggested the use of other than EC50 values for calculation of REPs (DeVito et al., 2000; Toyoshiba et al., 2004; Villeneuve et al., 2000). The advance of a benchmark approach at the lower part of the dose–response curve is that the lack of agreement in curve shape is less pronounced compared to EC50. Furthermore, in many cases, the BMR20TCDD also present an exposure situation that is more relevant and closer to the actual human exposure. Though the BMR20TCDD value was preferred above a lower BMR value, e.g. BMR10TCDD or 62


Intake and systemic REPs of DLCs in SD rats

BMR05TCDD, as these BMRs would usually fell within the background noise. Thus, REPs were calculated by dividing the concentration of BMR20 of TCDD by the BMR20TCDD concentration of another congener.

Statistical analysis Statistically significant differences of the means and variances were determined using analysis of variance (oneway ANOVA) test followed by a Tukey–Kramer multiple comparisons test. Differences were considered statistically significant if P < 0.05. Statistical calculations were performed using GraphPad 6.01 (GraphPad Software Inc., San Diego, CA). Results

Body and organ weights and tissue concentrations To evaluate the possible toxic effects of the tested congeners, body and organ weights were examined. A dose dependent decrease in relative thymus weight was observed for all compounds, but was only statistically significant for 4-PeCDF and PCB-126. In addition, a significant increase in liver weight was observed for all DLCs tested. Furthermore, the analysis of the hepatic lipid fraction (% lipid/g liver) of the pooled samples suggests a dose-dependent increase compared to the vehicle control-treated rats for all congeners. In this case, no statistical test could be performed due to the use of pooled samples. More detailed information is provided in Supplementary Material Table S2. Generally, tissue concentrations of all congeners increased linearly with the administered dose (See Supplementary Material; Figure S1 and Table S3). Furthermore, liver sequestration was seen for TCDD, PeCDD, 4-PeCDF and PCB-126 with liver–adipose concentration ratios >0.3, a suggested cut-off for liver sequestration (Diliberto et al., 1997), while the mono-ortho PCBs 118, 156, and the nondioxin- like PCB-153 did not show this liver sequestration with liver–adipose concentration ratios of 0.07, 0.13 and 0.06, respectively (Table 1). More details on tissue distribution of these congeners have been described elsewhere by van Ede et al. (2013c).

Dose–response curves Dose–response relationships for hepatic EROD activity and gene expression of Cyp1a1, 1b1, 1a2 and Ahrr in liver and PBLs were determined using intake or administered dose levels and liver, adipose tissue or plasma concentrations (See Supplementary Material; Figure S2 and Figure S3). In the liver, all compounds, except the non-dioxin-like PCB153, significantly induced hepatic EROD activity as well as Cyp1a1, 1b1, 1a2 and Ahrr gene expression. For hepatic EROD activity, TCDD caused already a maximum 63

3


Table 1: Liver:adipose concentration ratios Congener TCDD

0,5

4-PeCDF

PCB-156 PCB-153  

3,9 ± 0,3

5,0 ± 1,0* 4,4 ± 0,5

11,0 ± 1,2

25

40,8 ± 6,7*

25

9,0 ± 0,8

100

ratio

18,7 ± 3,2*

5

liver:adipose

2,5 10

0,5

10

PeCDD

PCB-118

µg/kg bw 2,5

PCB-126

Dose

5

100

5000

16,8 ± 1,4 21,0 ± 6,7 30,7 ± 3,1 7,2 ± 1,1

10,7 ± 1,5

0,05 ± 0,01

15000

0,06 ± 0,01

15000

0,14 ± 0,02

50000 5000

50000 5000

15000 50000

0,08 ± 0,02 0,12 ± 0,03

0,13 ± 0,02 0,04 ± 0,01

0,06 ± 0,01 0,06 ± 0,01

Data represents the mean ± SD (based on ng/g tissue) of 6 rats. * p < 0.05 compared with the next lower dose, determined by one-way ANOVA followed by Tukey’s multiple comparisons test.

induction at the lowest dose tested (0.5 μg/kg bw) and it was not possible to define a dose–response curve. Also for PeCDD, 4-PeCDF, PCB-126 and PCB-156, EROD activity was already at 60–75 % of their maximal responses at the lowest doses tested (0.5, 5, 5 and 5,000 μg/kg bw, respectively). A clear distinction in hepatic Cyp1a1, 1b1, 1a2 and Ahrr gene expression was observed between more potent AhR agonists TCDD, PeCDD, 4-PeCDF and PCB-126 with induction of 60–100 % of the maximal TCDD response and less potent AhR agonists PCB-118 and PCB-156 with a significant induction below 10 % of maximal TCDD response. In PBLs, induction of Cyp1a1, 1b1 and Ahrr genes was up to three orders of magnitude lower compared to those in the liver. Gene expression of Cyp1a2 could not be determined in PBLs. Dose–response curves of Cyp1a1 gene 64


Intake and systemic REPs of DLCs in SD rats

expression in PBLs could be determined for all compounds tested, except for the non-dioxin-like PCB- 153. Furthermore, all compounds except PCB-118 and PCB-153 statistically significantly induced gene expression of Cyp1b1 and Ahrr in PBLs (See Supplementary Material; Figure S2).

BMR20TCDD concentrations and relative effect potencies (REPs) With the obtained doseâ&#x20AC;&#x201C;response curves, comparative BMR20TCDD concentrations for different congeners were calculated. Each congener-specific BMR20TCDD concentration was calculated based on the intake dose or systemic concentration needed for a congener to reach 20 % effect caused by TCDD for that particular endpoint. It was not possible to calculate a BMR20TCDD concentration for each congener studied and every endpoint measured in the liver or PBLs. For example, no BMR20TCDD concentrations could be calculated for hepatic EROD activity of the different congeners, as no full doseâ&#x20AC;&#x201C; response curve could be defined for TCDD. Furthermore, some congeners did not reach a BMR20TCDD for all endpoints studied; these data were excluded for REP calculations (See Table 2).

With the BMR20TCDD concentrations based on administered dose, tissue and plasma concentrations, intakeREPs and systemicREPs for the different congeners could be calculated (Table 2). To compare changes between systemicREPs and intakeREPs, each congener-specific intake REP was set to 1 and deviations from the intakeREP are given for each systemicREP with the same biological endpoint (see Fig. 1). For PCB 118, a REP could only be calculated for Cyp1a1 mRNA expression in PBLs and this congener has therefore not been included in Fig. 1.

When looking at Fig. 1, it is evident that hepatic systemicREPs for PeCDD, 4-PeCDF and PCB-126 based on liver concentrations (wet or lipid weight) were similar or up to onethird of the corresponding intakeREPs. systemicREPs for hepatic effects using adipose tissue or plasma concentrations were up to threefold higher compared to intakeREPs for PeCDD and up to one order of magnitude higher for 4-PeCDF. For PCB-126, up to twofold higher systemic REPs were calculated based on adipose concentrations compared to intakeREPs. In contrast, based on plasma concentrations, systemicREPs for this congener were at most one-fifth of the intakeREP. systemicREPs were also calculated for endpoints in PBLs based on plasma concentrations. The deviations observed from intakeREPs for PeCDD and 4-PeCDF showed striking similarities with those observed for the same endpoints in the liver if calculated using adipose or plasma concentrations (See Fig. 1). For PCB-126, systemicREPs in PBLs were about twofold higher compared to intakeREPs, which is similar to hepatic systemic REPs based on adipose tissue. PBLbased systemicREPs could also be calculated for PCB-156, and these were at most one-third of the intakeREPs. When we compare intakeREPs 65

3


Table 2: Mean BMR20TCDD concentrations for TCDD, PeCDD, 4-PeCDF, PCB-126, PCB-118 and PCB-156 and corresponding relative effect potencies (REPs) for various endpoints in liver and peripheral blood lymphocytes. Biomarker

Dose metric

Liver

Adm. dose (µg/kg bw)

mRNA Cyp1a1

Liver

mRNA Cyp1b1

Liver

mRNA Cyp1a2

PBLs

mRNA Cyp1a1 PBLs

mRNA Cyp1b1 PBLs

mRNA Cyp1a2 PBLs

mRNA Ahrr

Sys. liver (ng/g liver)

Sys. liver (ng/g lipid)

Sys. adipose (ng/g lipid) Sys. plasma (ng/g lipid)

TCDD

BMR20TCDD

REP

BMR20TCDD

2.21

1

10.7

0.2

63.2

0.22

58.3 0.73 0.86

2.34

Sys. adipose (ng/g lipid)

3.48

Sys. liver (ng/g lipid)

Sys. plasma (ng/g lipid)

28.7

630.56 4.23

Adm. dose (µg/kg bw)

0.15

Sys. adipose (ng/g lipid)

0.54

Sys. liver (ng/g liver)

Sys. liver (ng/g lipid)

Sys. plasma (ng/g lipid) Adm. dose (µg/kg bw)

Sys. plasma (ng/g lipid) Adm. dose (µg/kg bw)

Sys. plasma (ng/g lipid) Adm. dose (µg/kg bw)

Sys. plasma (ng/g lipid) Adm. dose (µg/kg bw)

Sys. plasma (ng/g lipid)

4-PeCDF

REP

Adm. dose (µg/kg bw) Sys. liver (ng/g liver)

PeCDD

BMR20TCDD

1.53 40.7 0.59 7.97 21.2 11.1 30.1 ND ND

9.23 26.3

1 1 1 1

0.86 263

0.86 2.88

1

9.17

1

3.79

1 1 1

308

3247 9.02

1

2.01

1

1.56

1 1 1 1 1 1 1 1 1

25.3 610

5.00 99.7 106

67.1 100 ND ND

23.6 29.4

0.3 0.2 0.8 0.3 0.3

0.09 0.2 0.9 0.5

0.08 0.06 0.07 0.3 0.1

0.08 0.2 0.2 0.3 0.4 0.9

5.79

1650 3.34 4.25 35.7 460

9712 7.70 6.02 19.7 203

5068 5.98 5.53 512 134 348

78.7 ND ND

52.9 12.1

Data are expressed as mean BMR20TCDD derived from dose-response curves of 6 rats. REPs are calculated as described in Materials & Methods. ND = not determined, because BMR20TCDD was not reached PBLs = Peripheral blood lymphocytes

66


Intake and systemic REPs of DLCs in SD rats

Table 2: Mean BMRoncentrations for TCDD, PeCDD, 4-PeCDF, PCB-126, PCB-118 and PCB-156 and corresponding relative effect potencies (REPs) for various endpoints in liver and peripheral blood lymphocytes PCB-126

REP

BMR20TCDD

REP

0.04

19.8

0.1

0.04 0.04 0.2 0.2

0.07 0.06 0.06 0.5 0.7

0.008 0.008 0.008 0.09 0.1

0.02 0.2

0.03 0.4 0.2 2.2

2.24 493

3.72 34.6 29.4 393

8409 29.2 57.1 7.06 82.0

2001 10.9 44.9

3591 5900 733

1344 ND ND

139 322

0.1 0.1 0.2

PCB-118

BMR20TCDD ND ND

0.07

ND

0.07 0.1

0.07 0.02 0.02 0.02 0.05 0.01

0.002 0.004 0.02 0.02 0.07 0.08

ND ND

ND

ND ND

ND

ND

ND

ND

ND

ND

ND 182180 730092 ND ND

ND ND

ND ND

3

ND

ND

ND

ND

REP

ND

ND

ND

BMR20TCDD ND

ND

ND

PCB-156 ND

ND

0.02 0.08

REP

ND 0.00004 0.00003

ND ND

227864

0.00003

5138852

0.000006

1143663 723945 ND ND

120195 593618

0.00002

0.00002

0.00008 0.00004

67


 

 



   

     



   

 

 

   

    

   #

  !#

   

   #

  !#

         

    





    



"



"

"

  



"

    

"



"

   

   #

 

 

   

   #



    



 "



"

"

"



"





 

   

  

    

  

    

    

    





    

"

  

"

  



"

    

"

  



"

  

Figure 1. Fold change in systemicREP compared with intakeREP for PeCDD (A), 4-PeCDF (B), PCB126 (C), and PCB-156 (D). Changes in REPs are calculated for Cyp1a1, Cyp1b1, Cyp1a2 and Ahrr gene expression in liver and/or PBLs. Abbreviations: ND, not determined; Syst, systemic

and systemicREPs from this rat study with the current WHO-TEFs, several observations can be made, and these are visualized in Fig. 2.

For PeCDD, 4-PeCDF and PCB-126, the median intakeREPs were with 0.3, 0.04 and 0.05, respectively, all well below the WHO-TEF values of 1, 0.3 and 0.1, respectively. For PeCDD and 4-PeCDF, most systemicREPs based on liver concentrations are with median REPs of 0.2 and 0.04 below the WHO-TEFs and even outside the half log uncertainty

68


Intake and systemic REPs of DLCs in SD rats

range suggested for these TEF values. All systemicREPs of 4-PeCDF based on plasma concentrations are higher than the intakeREPs, but fall mostly within the suggested WHOTEF uncertainty range (median 0.3). For PCB-126, all systemicREPs are below the WHOTEF value of 0.1, but partly overlap with the WHO-TEF uncertainty range (median liver systemicREP 0.08 and plasma systemicREP 0.02). The only intakeREP and plasma-based systemic REP that could be calculated for PCB-118 was that of Cyp1a1 gene induction in PBLs, and this value was similar to the WHO-TEF of 0.00003. For PCB-156, intakeREPs (median 0.00003) and plasma-based systemicREPs could only be determined for endpoints measured in PBLs, and these values are mostly within the uncertainty range around the WHO-TEF of 0.00003 (See Fig. 2). Discussion During the latest WHO-TEF re-evaluation in 2005, it was concluded that more data are needed to confirm that intakeTEFs can reliably be used for risk assessment based on systemic biological matrices, such as blood. Here, we compare for the first time intakeREPs and systemicREPs based on liver, adipose and plasma concentration for female SD rats 3 days after a single oral dose of different DLCs. Toxicokinetics In our study, we used a 3-day experimental protocol which may raise questions to which extent these results are relevant for (sub)chronic exposure situations. Based on the known toxicokinetic properties of the selected compounds (Van den Berg et al., 1994), we estimated that the initial body distribution 3 days after dosage would be mostly completed. In addition, the free tissue concentration must be considered as direct cause of an induced effect, if such an effect is rapidly expressed in time, like CYP1A1, 1A2 and 1B1 expression and activity. A previous study in rats showed a maximum protein level of CYP1A1, 1A2 and 1B1 3 days after exposure to TCDD (Santostefano et al., 1997). In this respect, the attained tissue concentration, derived either from a single or (sub)chronic exposure, may produce similar effects if metabolism of a DLC does not play a role of importance during the experimental time period (Van den Berg et al., 1994). To examine whether the above supposition is indeed true, the tissue distributions observed in our 3-day rat and mouse studies were compared with results from experiments using a (sub) chronic dosage regime. Results from this comparison showed that congener distributions for the seven in vivo congeners used here and responding effect concentrations for at least TCDD are approximately similar (van Ede et al., 2013c). Distribution of these congeners is very much dependent on the amount of sequestration in the liver, due to CYP1A2 binding (Diliberto et al., 1997; 1999). This hepatic sequestration is likely to 69

3


play an important role in differences between intakeREPs and systemicREPs. Because TCDD, the reference compound for most studies, strongly sequesters in the liver with a large hepatic fraction bound to CYP1A2, it can be argued that a significant part of the systemic hepatic concentration is unavailable for AhR activation. If total hepatic concentrations are used as metric and another DLC sequesters differently from TCDD, this may lead to either an underestimation or overestimation of the hepatic systemicREP. As a consequence, systemic REPs based on blood or adipose concentrations and an (extra)hepatic response may better predict the congener-specific potency. Here, the induction of CYP1A2 and subsequent binding to this enzyme is less likely to play a role of concern (van Ede et al., 2013a; 2013c). Intake REP versus WHO-TEF In this study, REPs were calculated based on Cyp1a1, 1b1, 1a2 and AhRR gene expression. These endpoints are considered to be among the most sensitive biological responses for AhR-mediated effects upon exposure to DLCs. Although these biomarkers are not a measure of toxicity, several studies have shown a high correlation between REPs calculated based on these genes and toxic responses associated with exposure to DLCs such as, wasting syndrome, thymus atrophy or hepatic porphyrin accumulation (Safe, 1990; van Birgelen et al., 1996). In general, this rat study shows lower median intakeREPs than the WHO-TEF values for PeCDD, 4-PeCDF and PCB-126 (See Fig. 2). This might be partly because the WHO-TEFs are based on a range of REPs derived from many different experiments. Furthermore, we have selected to use the BMR20TCDD concentrations for comparison, which for some congeners, do deviate from those based on EC50 concentrations. The BMR20TCDD as reference point on the lower part of the doseâ&#x20AC;&#x201C;response curve was selected, because for many compounds and endpoints, no similar Ymax or Hill slope as TCDD could be observed. As a result, using EC50 values to determine REPs would most like have provided a larger error than based on the BMR20TCDD concentrations (See also Materials and Methods/Data analysis). Intake REP versus systemic REP When comparing intakeREPs with systemicREPs within this 3-day rat study, it is clear that systemic REPs based on liver concentrations and biological endpoints in the liver showed minimal deviations from intakeREPs. In contrast, systemicREPs of PeCDD, 4-PeCDF and PCB-126 based on either plasma or adipose tissue concentrations and these hepatic biological endpoints were up to one order of magnitude higher than their intakeREPs (see Fig. 1). Comparable effects were seen if systemicREPs for plasma concentrations and biological effects in PBLs were compared with intakeREPs. For PCB-126, large differences 70


Intake and systemic REPs of DLCs in SD rats

in systemicREPs based on either adipose or plasma concentrations were observed. These differences may be due to incomplete (re)distribution between plasma and adipose tissue for PCB-126 after 3 days (See Supplementary Material Table S2). Systemic REPs There are only limited data in the literature on systemicREPs for DLCs. Recently, we have described a similar study with mice where differences between intakeREPs and systemicREPs were determined (van Ede et al., 2013a). To our knowledge, only two subchronic mouse studies and four NTP (sub) chronic rat studies have been published that report liver concentrations combined with doseâ&#x20AC;&#x201C;response curves for several congeners. This allows comparison of our liver-based systemicREPs with those reported in the literature. Generally, intake REPs and systemicREPs for mice based on liver concentrations from the literature are in good agreement with the results from our mouse study (see Supplementary Material; Figure S4) (DeVito et al., 1997; van Birgelen et al., 1996). Data on EROD and ACOH activity from the NTP studies with TCDD, 4-PeCDF, PCB-126 and PCB-118 were used to calculate intakeREPs and systemicREPs based on liver concentrations after 14, 31 and 53 weeks (National Toxicology Program, 2006b; 2006c; 2006d; 2010). For that, we used the same benchmark approach (BMR20TCDD) as in our present rat study and previously reported mouse data (See Supplementary Material; Figure S5 and Table S4).

For 4-PeCDF, intakeREPs from the NTP studies were generally above the REP ranges from our studies. However, systemicREPs based on liver concentration were in good agreement. For PCB-126 and 118, NTP intake and systemic REPs were mostly within the same range as the REPs determined in our studies. Exceptions were the intakeREPs and systemicREPs for ACOH activity for 14 and 31 weeks, which were generally above the REP ranges from our studies (See Supplementary Material Figure S4). To our knowledge, no studies are available that would allow a comparison with our plasma-based systemicREPs. Deviation systemic REPs from WHOâ&#x20AC;&#x2018;TEF The same experimental design of the present study and our previously reported mouse data allows us to combine the results from both studies and compare these with the present WHO-TEF values that are generally used for human risk assessment (See Fig. 3 and Supplementary Material Figure S6). Plasma-based systemicREPs for PeCDD and PCB-126 are in the same range as our intakeREPs, thus lower than the WHO-TEF values. In contrast, the median plasma systemicREP of 4-PeCDF is an order of magnitude higher compared to intake and systemic liver-based REPs and similar to the WHO-TEF. Plasmabased systemicREPs distribute mostly within the TEF uncertainty range. However, it must 71

3


72

0.0001

0.001

0.01

0.1

1

10

Intake

Liver Plasma Intake

4-PeCDF

Liver Plasma Intake

PeCDD Liver Plasma

PCB-126

0.000001

0.00001

0.0001

0.001

0.01

Intake

Liver Plasma Intake

PCB-118

Liver Plasma

PCB-156

Figure 2. Intake and systemic relative effect potencies determined in this study in relation to the 2005 WHO-TEF ± half log uncertainty range. REPs were determined for hepatic gene expression of Cyp1a1 ( ¢ ), Cyp1b1 ( ◧ ), Cyp1a2 ( o ), and gene expression of Cyp1a1 ( ), Cyp1b1 ( ◐ ) and Ahrr ( ¤ ) in PBLs of PeCDD, 4-PeCDF, PCB-126 (left graph) and PCB-118 and PCB-156 (right graph). REPs for hepatic endpoints were calculated based on administered dose (Intake), lipid-based liver concentration (Liver) or lipid-based plasma concentration (Plasma), whereas for PBL, REPs were calculated using the administered dose or plasma concentration. The black lines represents the median of the REPs. The black dotted line represents the WHO-TEF ± half log uncertainty range (grey area).

Relative Effect Potencies (REPs)


0.0001

0.001

0.01

0.1

1

Intake

Liver Plasma Intake

4-PeCDF

Liver Plasma Intake

PeCDD Liver Plasma

PCB-126

0.000001

0.00001

0.0001

0.001

Intake

Liver Plasma Intake

PCB-118

Liver Plasma

PCB-156

Figure 3. Box plot of intake and systemic relative effect potencies from this rat study and our mouse study (Van Ede et al. 2013a) in relation to the WHO-TEF Âą half log uncertainty range (black dotted line and grey area).

Relative Effect Potencies (REPs)

10

Intake and systemic REPs of DLCs in SD rats

3

73


be noted that the WHO-TEF of 0.3 is somewhat below the 75th percentile of plasmabased systemicREP distribution. In the 2005 WHO-TEF reassessment, the 75th percentile of the in vivo REP distribution for an individual congener was used as initial decision point to reassess the WHO 1998 TEF (Van den Berg et al., 2006). Based on that reassessment, the TEF for 4-PeCDF was then adjusted from 0.5 (1998 WHO-TEF) to 0.3. The two mono-ortho PCBs 118 and 156 show a large variation depending on the matrix of choice, but plasma-based systemicREPs are mostly below their WHO-TEF of 0.00003. Liverbased systemicREPs of PCB-118 and PCB-156 are much higher than intakeREPs and plasmabased systemicREPs and the WHO-TEF. However, this can be explained due to the hepatic sequestration of TCDD when compared to PCB-118 and PCB-156 (van Ede et al., 2013c).

In addition, it should be recognized that a comparison of our REPs with the WHO-TEFs were based on a limited number of endpoints, whereas the WHO-TEFs consist of REPs that cover a much broader range of endpoints. This obviously offers an uncertainty, which warrants caution for a very absolute comparison between the REPs of our study and the WHO-TEFs. Nevertheless, the endpoints used in these studies, e.g. Cyp1a1, have also been frequently used in studies that were included in the WHO-TEF derivation. Conclusion

The combined results of the previously reported mouse and present rat study indicate that within this experimental model, plasma-based systemicREPs for all congeners, except 4-PeCDF, are within a half log range around the intakeREP. These data suggest that the use of systemicREPs as systemicTEFs would not contribute to better human risk assessment. However, further investigation for plasma-based systemicREPs for 4-PeCDF is desirable.

74


Intake and systemic REPs of DLCs in SD rats - Supplemental Material

Supplemental material Table S1: Congeners, TEF-values and dose ranges Congener

TEF

Single oral dose (µg/kg bw)

PeCDD

1

0.5

2.5

10

25

PCB-118

0.00003

5000

15000

50000

150000

500000

ND

5000

15000

50000

150000

500000

TCDD

4-PeCDF

PCB-126 PCB-156 PCB-153

1

0.3

0.1

0.00003

1

0.5

5

5

5000

2

2.5

25

25

15000

3

10

100

100

50000

4

5

250

1000

25

250

150000

100 100

1000

3

500000

75


76

PCB-126

4-PeCDF

PeCDD

TCDD

Congener

(gram)

227,9 ± 11,4

231,0 ± 10,0

2,5

228,8 ± 14,5

224,8 ± 12,9

221,0 ± 11,1

222,9 ± 12,0

25

5

0

208,6 ± 6,6

224,2 ± 6,7

231,5 ± 5,2

228,5 ± 7,2

1000

250

100

5

222,9 ± 12,0

25

0

243,7 ± 9,3

239,2 ± 6,2

233,5 ± 14,7

232,1 ± 8,0

239,4 ± 10,1

100

25

10

0,5

0

225,0 ± 7,6

235,6 ± 11,0

100

25

10

225,0 ± 14,3

212,2 ± 5,6

239,4 ± 10,1

a

Body weight

2,5

0,5

0

µg/kg bw

Oral dose

0,17 ±

0,22 ±

0,24 ±

0,15 ±

0,16 ±

0,18 ±

0,19 ±

0,19 ±

0,24 ±

0,15 ±

0,16 ±

0,17 ±

0,18 ±

0,18 ±

0,19 ±

0,15 ±

0,15 ±

0,16 ±

0,17 ±

0,18 ±

0,19 ±

a

d

5,24 ± 0,32 0,02

d

0,03

0,03

0,02

4,11 ± 0,29

3,86 ± 0,19

3,72 ± 0,22

d

d

4,58 ± 0,27

d

4,25 ± 0,23 d

0,02

d

0,04 d

4,01 ± 0,21

3,89 ± 0,21

3,72 ± 0,22

5,08 ± 0,40

d

4,49 ± 0,10d

4,17 ± 0,36 d

4,00 ± 0,20

3,83 ± 0,24

3,60 ± 0,21

5,61 ± 0,18

d

5,31 ± 0,32 d

4,99 ± 0,12

de

4,55 ± 0,24d

4,20 ± 0,26

3,60 ± 0,21

% of bw

Liver

0,03d

0,02

0,03

0,02

0,03

0,02

0,01

0,02

0,03

0,01

0,03

0,03

0,03

0,04

0,03

% of bw

a

Thymus

Table S2: Body weight, relative thymus, liver and spleen weights and % lipid/g liver.

0,03

0,32 ±

0,29 ±

0,29 ±

0,30 ±

0,32 ±

0,30 ±

0,32 ±

0,29 ±

0,31 ±

0,30 ±

0,31 ±

0,30 ±

0,29 ±

0,29 ±

0,31 ±

0,28 ±

0,01

0,04

0,02

0,04

0,03

0,04

0,02

0,03

0,02

0,02

0,02

0,02

0,03

0,03

0,02

0,01

0,03

0,02

0,30 ±

0,30 ±

0,31 ±

0,01

0,02

a

0,29 ±

0,28 ±

% of bw

Spleen

3,77

3,98

3,43

4,23

4,51

4,10

4,09

3,81

3,43

4,46

4,85

4,56

4,28

3,98

3,38

NA

NA

3,81

3,94

3,85

3,38

g liverb

% lipid /


233,0 ± 7,3

257,5 ± 40,5

225,1 ± 13,6

500000

150000

50000

15000

213,6 ± 6,5

217,9 ± 5,6

233,2 ± 10,7

219,2 ± 9,6

225,7 ± 7,8

211,6 ± 6,0

5000

0

500000 c

219,1 ± 14,5

230,7 ± 6,5

150000

50000

233,4 ± 14,9

222,4 ± 5,4

15000

5000

0

219,0 ± 2,8

225,3 ± 8,5

231,2 ± 14,1

500000

150000

50000

232,2 ± 7,7

222,4 ± 5,4

15000

5000

0

220,1 ± 21,5

1000

225,6 ± 12,3

192,7 ± 6,0c

0,17 ±

0,18 ±

0,20 ±

0,19 ±

0,19 ±

0,20 ±

0,15 ±

0,13 ±

0,16 ±

0,17 ±

0,19 ±

0,19 ±

0,16 ±

0,18 ±

0,19 ±

0,19 ±

0,19 ±

0,19 ±

0,15 ±

0,14 ±

0,21 ±

0,02

0,02

0,02

0,03

0,04

0,03

0,03

0,02

0,01

0,04

0,03

0,05

0,01

0,02

0,03

0,02

0,01

0,05

0,02

d

0,02

d

0,07d

4,08 ± 0,35

4,00 ± 0,29

4,15 ± 0,31

3,62 ± 0,31

3,77 ± 0,46

3,83 ± 0,25

6,33 ± 0,35

b

a

d

5,36 ± 0,59

d

4,46 ± 0,22

3,98 ± 0,65

3,67 ± 0,15

3,88 ± 0,18

4,90 ± 0,30d

4,76 ± 0,43 de

4,05 ± 0,15

3,70 ± 0,20

3,65 ± 0,12

3,88 ± 0,18

5,12 ± 0,47

d

4,86 ± 0,46

de

4,03 ± 0,32

Data represents the mean ± SD of 6 rats Data represents the % lipid per gram of pooled liver samples from 6 rats Statistically significant changes were determined by one-way ANOVA analysis followed by a Tukey’s multiple comparisons test, cSignificantly different from day 0 and control group (p<0.05) d Significantly different from control group (p<0.05) e Significantly different from previous concentration (p<0.05) NA = not analysed

PCB-153

PCB-156

PCB-118

250

100

0,01

0,29 ±

0,29 ±

0,32 ±

0,28 ±

0,30 ±

0,32 ±

0,28 ±

0,28 ±

0,02

0,04

0,02

0,02

0,04

0,02

0,03

0,01

0,07

0,02 0,28 ±

0,29 ±

0,29 ±

0,03

0,02d

0,04

0,02

0,03

0,32 ±

0,26 ±

0,30 ±

0,30 ±

0,31 ±

0,02

0,03

0,29 ±

0,02

0,04

0,32 ±

0,30 ±

0,31 ±

0,04

0,29 ±

NA

NA

3,32

3,16

3,10

2,86

NA

5,33

4,09

4,63

3,97

3,73

NA

NA

4,27

3,63

3,54

3,73

5,14

4,14

6,09

Intake and systemic REPs of DLCs in SD rats - Supplemental Material

3

77


Table S3: PCDD / PCDF / PCB concentrations in liver, adipose tissue and plasma 3 days after a single oral dose Congener

TCDD

PeCDD

Oral dose µg/kg bw

0,5

2,5

100

NA

25

5,9 ± 1,6

10

575 ± 224

4,0

4,5 ± 1,1

125 ± 21

1986 ± 541

4,3

3,4

1,2 ± 0,2 17 ± 3 NA

NA

117 ± 14

2558 ± 300

5,2

7,0 ± 1,1

100

822 ± 148

18423 ± 3309

3,7

25

258 ± 66

6316 ± 1625

4,6

6,5 ± 1,9

63562 ± 10556

5,1

NA

1491 ± 122

5,3

8,4 ± 1,3

2,5

PCB-126

5

1000 25

33 ± 5

240 ± 38 55 ± 5

878 ± 106

2867 ± 476

771 ± 116

4948 ± 782

1439 ± 142

21423 ± 2597

9900 ± 938

234043 ± 22177

247 ± 29

6543 ± 781

59 ± 5

5,9

4,3

4,9

3,9

4,4

4,4

1,8 ± 0,4 NA

NA

2,8 ± 0,9 29 ± 5 NA

28 ± 4

100

1533 ± 288

25178 ± 4721

6,8

147 ± 38

1000

7783 ± 1858

151427 ± 36156

3,5

NA

15000

4400 ± 1026

121212 ± 28255

250

5000

2200 ± 276 782 ± 131

50000

16833 ± 4262

500000

NA

150000

NA

5000

2250 ± 1313

50000

19833 ± 3764

15000

150000

PCB-153

5000

500000

15000

6850 ± 1626

22081 ± 3694

394223 ± 99818

3,9

0,1

0,1

0,1

56675 ± 33064

0,2

484923 ± 92026

0,2

147948 ± 35113

0,2

1175735 ± 293228

945 ± 654

30484 ± 21099

0,1

309237 ± 94228

0,1

NA

3567 ± 1308

50000

10267 ± 3128

500000

NA

150000

53140 ± 6659

62667 ± 15629

NA

Data represents the mean ± sd of 6 rats NA = not analysed

78

ng/g tissuea

0,5 ± 0,1

100

Adipose

% dose / g tissue

5,2

250

a

Liver

ng/g lipida

148 ± 40

5

PCB-156

NA

0,5

23 ± 9

76 ± 21

4-PeCDF

PCB-118

4,8 ± 0,8

10

25

ng/g tissuea

112869 ± 41390

0,2

0,1

NA

15470 ± 2373

71100 ± 22243

226500 ± 90434 NA

NA

18000 ± 5577

51567 ± 14545

148800 ± 20375 NA

NA

20093 ± 9050

61050 ± 16482

168667 ± 30186 NA

NA


Intake and systemic REPs of DLCs in SD rats - Supplemental Material

Table S3: PCDD / PCDF / PCB concentrations in liver, adipose tissue and plasma 3 days after a single oral dose ng/g lipida 1,4 ± 0,2

Adipose

4,9 ± 1,2 19 ± 4

% dose / g tissue

ng/g tissuea

1,1

0,0053 ± 0,0004

0,8

0,077 ± 0,015

0,8

0,019 ± 0,005

0,197 ± 0,047

0,3

0,042 ± 0,004

31 ± 6

0,1

7,1 ± 2,1

0,1

9,2 ± 1,4

0,7

160 ± 41

0,7

31 ± 5

0,5

17000 ± 2608

1,4

251667 ± 100482

2,0

79000 ± 24714

2,1

20000 ± 6197

1,6

160000 ± 21909

1,3

56667 ± 15983

1,5

22833 ± 10284

1,8

191667 ± 34303

1,5

67833 ± 18313

1,8

0,0035

29 ± 6

0,089 ± 0,007

2,0 ± 0,4

0,0045

16 ± 2

0,0018

31 ± 2

167 ± 49

0,018 ± 0,004

6,0 ± 1,2

0,061 ± 0,013

0,161 ± 0,027

4,1 ± 1,3 23 ± 5

57 ± 10

1,231 ± 0,324

352 ± 93

0,168 ± 0,026

62 ± 10

0,044 ± 0,004

0,599 ± 0,195

1,323 ± 0,335 5,927 ± 1,538 47 ± 7

245 ± 62

801 ± 305

1890 ± 457

44 ± 4

285 ± 93

490 ± 124

2117 ± 549

0,0021

0,0009

0,0003

0,0003 0,0003

0,0005

0,0039 0,0030

0,0027 0,0024

0,0026

0,0071

700000 ± 169234 104667 ± 22447

29167 ± 9579

255000 ± 38341

825000 ± 148963

0,0073 0,0056 0,0056

0,0067

0,0078 0,0073 0,0064

6625 ± 2696

2650000 ± 1078425

0,0059

275 ± 67

161667 ± 39200

0,0081

2233 ± 530

893333 ± 211912

84 ± 53

758 ± 153

7560 ± 2866

32167 ± 20488

303333 ± 61210

2700000 ± 1023719

3

0,0016

258333 ± 98268

262 ± 56

2145 ± 387

0,0029

0,0042

2016667 ± 614546

816 ± 123

0,0031

19667 ± 2733

122667 ± 30943

6252 ± 1905 76 ± 25

0,0034

0,0035

0,467 ± 0,138 0,011 ± 0,003

0,0047

68 ± 16

6,3 ± 0,8

7,6 ± 1,2

0,3

8,8 ± 2,3

0,0164 ± 0,002

0,0051 ± 0,021

3,0 ± 0,9

% dose / g tissue

1,9 ± 0,2

262 ± 60

0,5 0,3

ng/g lipida

0,707 ± 0,161

0,6 ± 0,1 2,0 ± 0,5

Plasma

0,0074 0,0067 0,0066 0,0067

79




   

           



 



 

   

     















     Figure S1. Relation between oral dose and systemic concentration in rat liver (â&#x20AC;&#x201D;) or adipose tissue (---) for TCDD, PeCDD, 4-PeCDF, PCB-126, PCB-118, PCB-156 and PCB-153. Systemic concentrations were determined in female SD rats, 3 days after administration of a single oral dose. Data represents the mean Âą SD of 6 rats.

80


Intake and systemic REPs of DLCs in SD rats - Supplemental Material







  



  









 









   

  

























3

 















  

 



 



 









   



%"!$)(!%$%#&'(%%$('%"

 















 



%"!$)(!%$%#&'(%%$('%"

















 

 

  

%"!$)(!%$%#&'(%%$('%"









   







%"!$)(!%$%#&'(%%$('%"

 

 





   

%"!$)(!%$%#&'(%%$('%"





%"!$)(!%$%#&'(%%$('%"





%"!$)(!%$%#&'(%%$('%"

&#%"#!$# &'%(!$

 

 !

 



















  











Figure S2. Dose-response curves for hepatic EROD activity (A) and gene expression of Cyp1a1 (B), Cyp1b1 (C), Cyp1a2 (D), Ahrr (E) in rat liver and gene expression of Cyp1a1 (F), Cyp1b1 (G) and Ahrr (H) in peripheral blood lymphocytes (PBL) three days after a single oral dose of TCDD ( ¢ ), PeCDD ( o ), 4-PeCDF ( ), PCB-126 ( ), PCB-118 ( ¿ ) and PCB-156 ( ¯ ). Dose response curves are expressed using administered dose. BMR20TCDD is indicated with a black dotted line. Data are represented as mean ¹ SD (N=6). For TCDD curves were Ymax was not reached, GraphPad Prism 6.01 (GraphPad Software Inc., San Diego, CA) extrapolated the curve. The Cyp1a1 gene expression curve of PCB-126 in PBLs and the Cyp1b1 gene expression curve of PCB-156 in PBLs have been manually extrapolated until the BMR20TCDD. 81


Liver

20000

BMR20TCDD -2

-1

0

1

2

3

4

Administered dose (log µg/kg bw)

5

Cyp1a1 expression

60000 40000 20000

BMR20TCDD -1

0

1

2

3

4

5

6

7

Plasma concentration (log ng/g lipid)

Cyp1a1 expression

(fold induction compared to control)

20

10

40

B.

D.

30

0

6

80000

100000

PBL Cyp1a1 expression

40000

(fold induction compared to control)

60000

100000

Cyp1a1 expression

A.

80000

0

(fold induction compared to control)

PBL 40

(fold induction compared to control)

Cyp1a1 expression

(fold induction compared to control)

100000

BMR20TCDD

-2

-1

0

1

2

3

4

Administered dose (log µg/kg bw)

5

6

E.

30

20

10

0

BMR20TCDD

-1

0

1

2

3

4

5

6

Plasma concentration (log ng/g lipid)

7

C.

80000 60000 40000 20000 0

BMR20TCDD -1

0

1

2

3

4

5

6

7

Liver concentration (log ng/g lipid)

Figure S3. Cyp1a1 gene expression in rat liver and peripheral blood lymphocytes (PBL) three days after a single oral dose of TCDD ( ¢ ), PeCDD ( o ), 4-PeCDF ( ), PCB-126 ( ), PCB-118 ( ¿ ) and PCB-156 ( ¯ ). Dose response curves are expressed using administered dose (A and D), plasma concentration (B and E) or liver concentration (C). BMR20TCDD is indicated with a black dotted line. Data are represented as mean ± SD (N=6). For TCDD curves were Ymax was not reached, GraphPad Prism 6.01 (GraphPad Software Inc., San Diego, CA) extrapolated the curve. The Cyp1a1 gene expression curve of PCB-126 in PBLs has been manually extrapolated until the BMR20TCDD.

82


(R E P s )

0 .0 0 0 1

0 .0 0 1

0 .0 1

0 .1

1

In ta k e

L iv e r

PeC D D

P la s m a

In ta k e

L iv e r

4 -P e C D F P la s m a

In ta k e

L iv e r

P C B -1 2 6 P la s m a

0 .0 0 0 0 0 1

0 .0 0 0 0 1

0 .0 0 0 1

0 .0 0 1

0 .0 1

In ta k e

L iv e r

In ta k e

L iv e r

P la s m a

P C B -1 5 6

P C B -1 1 8

P la s m a

Figure S4. Intake and systemic relative effect potencies from this single dose rat study in combination with various other studies employing a single or multiple dosing regimen in relation to the WHO-TEF ± half log uncertainty range. Symbol colours; SYSTEQ rat study, this manuscript (red), SYSTEQ mouse study, Van Ede et al. 2013 (black), DeVito et al. 1997 (dark green), Van Birgelen et al. 1996 (purple), NTP study; 14 weeks (light blue), NTP study; 31 weeks (light green), NTP study; 53 weeks (dark blue), Budinsky et al. 2006 (orange). Symbol legend: hepatic EROD activity ( S ), hepatic ACOH activity ( ), hepatic porphyrin accumulation ( ), hepatic tumour incidence ( ¿ ), hepatic gene expression of Cyp1a1 ( ¢ ), Cyp1b1 ( ◧ ), Cyp1a2 ( o ), and gene expression of Cyp1a1 ( ), Cyp1b1 ( ◐ ) and Ahrr ( ¤ ) in PBLs of PeCDD, 4-PeCDF, PCB-126 (left graph) and PCB-118 and PCB-156 (right graph). Presented plasma-based systemicREPs are from EU-SYSTEQ mouse and rat studies only. The black line represents the median of the REPs. The black dotted line together with its grey area represents the WHO-TEF ± half log uncertainty range.

R e la tiv e E ffe c t P o te n c ie s

10

Intake and systemic REPs of DLCs in SD rats - Supplemental Material

3

83


EROD activity

(pmol/min/mg)

(pmol/min/mg)

0

1000

2000

3000

4000

5000

-2 -1 0

I.

4

4

3

4

(log ng/kg bw)

2

5

5

Administered dose

1

3

(log ng/kg bw)

2

5

Administered dose

1

4000 E. 3500 3000 2500 2000 1500 1000 500 0 -2 -1 0

3

(log ng/kg bw)

2

Administered dose

1

-2 -1 0

0

500

1000

1500

2000

2500 A.

6

6

6

7

7

7

8

8

8

9

9

9

0

1000

2000

3000

4000

5000

2

1

2

1

2

3

3

(log ng/g tissue)

Liver concentration

0

(log ng/g tissue)

0

3

Liver concentration

-4 -3 -2 -1

J.

1

(log ng/g tissue)

0

Liver concentration

-4 -3 -2 -1

4000 F. 3500 3000 2500 2000 1500 1000 500 0 -4 -3 -2 -1

0

500

1000

1500

2000

2500 B.

4

4

4

5

5

5

6

6

6

-2 -1 0

0.5

1.0

1.5

2.0

2.5

3.0

0

1

-2 -1 0

4.0 K. 3.5

0.5

1.0

1.5

2.0

2.5

3.0

4.0 G. 3.5

0.5

1.0

1.5

2.0

2.5

3.0

3.5 C.

4

4

4

5

(log ng/kg bw)

3

5

6

Administered dose

2

3

(log ng/kg bw)

2

5

Administered dose

1

3

(log ng/kg bw)

2

Administered dose

1

6

6

7

7

7

8

8

8

9

9

9

0.5

1.0

1.5

2.0

2.5

3.0

2

1

2

1

2

(log ng/g tissue)

-4 -3 -2 -1

0

3

3

Liver concentration

(log ng/kg bw)

0

3

Liver concentration

-4 -3 -2 -1

4.0 L. 3.5

0.5

1.0

1.5

2.0

2.5

3.0

1

(log ng/g tissue)

0

Liver concentration

-4 -3 -2 -1

4.0 H. 3.5

0.5

1.0

1.5

2.0

2.5

3.0

3.5 D.

4

4

4

5

5

5

6

6

6

Figure S5. Dose-response curves for hepatic EROD activity (A, B, E, F, I, J) and hepatic ACOH activity (C, D, G, H, K, L), for TCDD ( ¢ ), 4-PeCDF ( ), PCB-126 ( ) and PCB-118 ( ¿ ) derived from the NTP (sub)chronic rat studies (National Toxicology Program, 2006a; National Toxicology Program, 2006b; National Toxicology Program, 2006c; National Toxicology Program, 2010). Graphs represents data derived from the NTP; 14 weeks (upper graphs), NTP; 31 weeks (middle graphs) and NTP; 53 weeks (lower graphs) studies. Dose response curves are expressed using administered dose or liver wet weight concentrations. Relative effect potencies were calculated using the BMR20TCDD approach as described in Materials and Methods. BMR20TCDD is indicated with a black dotted line. Data are represented as mean ± SD (n=10).

(pmol/min/mg)

(pmol/min/mg)

(pmol/min/mg)

(pmol/min/mg)

EROD activity

EROD activity

EROD activity

EROD activity EROD activity

(nmol/min/mg)

(nmol/min/mg)

(nmol/min/mg)

(nmol/min/mg)

(nmol/min/mg) (nmol/min/mg)

ACOH activity ACOH activity

ACOH activity

ACOH activity ACOH activity ACOH activity

84


Adm. dose (ng/kg bw)

Adm. dose (ng/kg bw)

Adm. dose (ng/kg bw)

Adm. dose (ng/kg bw)

Adm. dose (ng/kg bw)

Liver ACOH activity

Liver EROD activity

Liver ACOH activity

Liver EROD activity

Liver ACOH activity

Sys. liver (ng/g liver)

Sys. liver (ng/g liver)

Sys. liver (ng/g liver)

Sys. liver (ng/g liver)

Sys. liver (ng/g liver)

0,74

934,35

0,32

511,34

4,25

3253,38

0,33

254,75

3,04

865,24

0,23

73,66

BMR20TCDD

TCDD

1

1

1

1

1

1

1

1

1

1

1

1

REP

35,86

2863,37

10,97

1068,48

73,74

4190,78

26,31

1643,93

43,07

3835,87

10,66

892,45

BMR20TCDD

4-PeCDF

0,02

0,3

0,03

0,5

0,06

0,8

0,01

0,2

0,1

0,2

0,02

0,08

REP

46,80

14528,73

6,81

2629,39

5,10

1938,73

11,42

4062,18

8,71

2032,43

14,35

3166,60

BMR20TCDD

PCB-126

20299840

26494969

3168

18815731 0,02

0,06

2288

15017876

2466 0,05

0,2

0,8

2816 1,7

0,03

30239849

3981

0,06

0,3

2870

0,4

0,02

14964131

BMR20TCDD

0,02

REP

PCB-118

0,0002

0,00005

0,0001

0,00003

0,002

0,0001

0,0001

0,000008

0,0008

0,00004

0,00008

0,000005

REP

Data are expressed as mean BMR20TCDD derived from dose-response curves from NTP-studies (National Toxicology Program, 2006a; National Toxicology Program, 2006b; National Toxicology Program, 2006c; National Toxicology Program, 2010), see also Figure S5. REPs are calculated as described in Materials & Methods of the main document.

(53 weeks)

(53 weeks)

(31 weeks)

(31 weeks)

(14 weeks)

Sys. liver (ng/g liver)

Adm. dose (ng/kg bw)

Liver EROD activity

(14 weeks)

Dose metric

Biomarker

Table S4: BMR20TCDD concentrations for TCDD, 4-PeCDF, PCB-126 and PCB-118 and corresponding relative effect potencies (REPs) derived from NTP-studies.

Intake and systemic REPs of DLCs in SD rats - Supplemental Material

85

3


86

(R E P s )

0 .0 0 0 1

0 .0 0 1

0 .0 1

0 .1

1

10

In ta k e

L iv e r

PeC D D

P la s m a

In ta k e

L iv e r

4 -P e C D F P la s m a

In ta k e

L iv e r

P C B -1 2 6 P la s m a

0 .0 0 0 0 0 1

0 .0 0 0 0 1

0 .0 0 0 1

0 .0 0 1

In ta k e

L iv e r

P C B -1 1 8 P la s m a

In ta k e

L iv e r

P C B -1 5 6 P la s m a

Figure S6. Intake and systemic relative effect potencies determined in this rat study (red) and our mouse study (black) (Van Ede et al. 2013) in relation to the WHO-TEF ± half log uncertainty range. REPs were determined for hepatic gene expression of Cyp1a1 ( ¢ ), Cyp1b1 ( ◧ ), Cyp1a2 ( o ), and gene expression of Cyp1a1 ( ), Cyp1b1 ( ◐ ) and Ahrr ( ¤ ) in PBLs of PeCDD, 4-PeCDF, PCB-126 (left graph) and PCB-118 and PCB-156 (right graph). REPs for hepatic endpoints were calculated based on administered dose (Intake), lipid-based liver concentration (Liver) or lipid-based plasma concentration (Plasma), whereas for PBL, REPs were calculated using the administered dose or plasma concentration. The black line represents the median of the REPs. The black dotted line together with its grey area represents the WHO-TEF ± half log uncertainty range.

R e l a t iv e E f f e c t P o t e n c i e s


Intake and systemic REPs of DLCs in SD rats - Supplemental Material

3

87


Chapter

4

Tissue Distribution of Dioxin-like Compounds: Potential Impacts on Systemic Relative Potency Estimates Karin I. van Ede1* Lesa L. Aylward2* Patrik L. Andersson3 Martin van den Berg1 Majorie B.M. van Duursen1 1

Endocrine toxicology group, Institute for Risk Assessment Sciences (IRAS), Utrecht University, the Netherlands 2 Summit Toxicology, LLP, 6343 Carolyn Drive, Falls Church, VA 22044, USA 3 Department of Chemistry, Umeå University, Umeå, Sweden * Both authors contributed equally to this study

Toxicology Letters 220: 294 – 302 (2013)


Abstract Relative effect potencies (REPs) for dioxins and dioxin-like compounds based on tissue concentration or internal dose (systemicREPs) can be considered of high relevance for human risk assessment. Within the EU-project SYSTEQ, systemicREPs for 1,2,3,7,8-pentachlorodibenzodioxin (PeCDD), 2,3,4,7,8,-pentachlorodibenzofuran (4-PeCDF), 3,3’,4,4’,5-pentachlorobiphenyl (PCB 126), 2,3’,4,4’,5- pentachlorobiphenyl (PCB 118) and 2,3,3’,4,4’,5-hexachlorobiphenyl (PCB 156) were calculated based on a plasma, adipose tissue or liver concentration in Sprague Dawley rats and C57bl/6 mice three days after a single oral dose. Compound-specific distribution as well as differences in accumulation and elimination can influence the tissue concentration and thereby the relative potency estimate of a congener. Here, we show that distribution patterns are generally similar for the tested congeners between the SYSTEQ dataset and other studies using either a single dose or subchronic dosing. Furthermore, the responding concentration for TCDD in single dose studies is comparable to the responding concentrations reported in subchronic studies. In contrast with data for laboratory rodents, available distribution data for humans in the general population display little or no hepatic sequestration. Because hepatic sequestration due to CYP1A2 protein binding may affect the amount of congener that is bioavailable for the AhR to produce hepatic responses, estimates of relative potencies between congeners with differing degrees of hepatic sequestration based on hepatic responses may be misleading for application to human risk assessment. Therefore, extra-hepatic concentration in blood serum/plasma or adipose tissue together with a biological extra-hepatic response might give a more accurate prediction of the relative potency of a congener for human responses under environmental conditions.

90


Tissue distribution of DLCs: Impacts Systemic REPs

Introduction

R

isk assessment of human exposure to mixtures of dioxin-like compounds (DLCs) relies upon the system of Toxicity Equivalency Factors (TEFs), which are estimates of the relative potency of a given DLC compared to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (Safe, 1990; 1994a; Van den Berg et al., 1998; 2006). Although these TEF values are explicitly based upon administered dose as the exposure metric, they are widely applied to the assessment of human exposures as quantified by the measurement of congener concentrations in human blood serum, generally expressed on a lipid-adjusted basis.

Differences in absorption, distribution, metabolism, and excretion can contribute to the relative potency of a congener when assessed on an administered dose basis (Budinsky et al., 2006; Devito and Birnbaum, 1995; DeVito et al., 1997; 2000; Diliberto et al., 1999). Thus, relying upon current “intake” TEF values for risk assessment based on blood serum concentrations may give a misinterpretation of the risk. At the latest 2005 WHO expert meeting, where the TEFs were (re-)evaluated, it was concluded that there was insufficient data available to develop ‘systemic’ TEFs, i.e. TEFs applicable to measured blood or tissue concentrations. This presents a clear gap in the risk assessment process for DLCs (Van den Berg et al., 2006). In order to fill this data gap, the EUproject SYSTEQ was undertaken. The main objective of SYSTEQ was to establish in vivo systemic relative effect potencies (systemicREPs) in mouse and rat and compare these with intakeREPs in the same animal model. The SYSTEQ project used a single oral dose regimen with measurement of tissue concentrations and responses in both hepatic and extra-hepatic tissues three days after dosing. There was special focus on responses in peripheral blood lymphocytes (PBLs) as potential biomarkers of exposure and response (van Ede et al., 2013a; 2013b). It is assumed that tissue concentrations used as the basis for calculations of systemic REPs across congeners in the SYSTEQ project reflect the toxicokinetic aspects of the administered doses. For some DLCs, tissue distribution into hepatic vs. extrahepatic tissues is strongly influenced by congener-specific affinity and binding to the cytochrome P450 1A2 (CYP1A2) protein in the liver (Devito et al., 1998; Diliberto et al., 1995; 1999; Poland et al., 1989; Voorman and Aust, 1987; Yoshimura et al., 1984). This congener-specific and dose-dependent binding to CYP1A2 can result in hepatic sequestration of specific compounds. Yet, the impact of this hepatic protein binding on the free available concentration of a compound for inducing biological responses is not fully understood. In addition, these compounds accumulate over time due to low, but still different, elimination rates. Thus, relative concentrations in hepatic as well as extra91

4


hepatic tissues are dependent on the congener, dose and dosing regimen (e.g. single vs. subchronic dosing). As a consequence, calculated systemicREPs may differ depending on these variables.

The response metrics mentioned above are also of direct relevance to human risk assessment when determining REP and TEF values. So far, many studies that generated data for REPs have focused on hepatic responses in rodents, e.g. enzyme induction, tumorigenesis, retinoid changes or oxidative stress (Haws et al., 2006). However, present concerns regarding sensitive responses in human populations of DLCs are more and more focused on extra-hepatic responses, including (neuro)developmental endpoints, reproductive functions, immunotoxicity, and extra-hepatic carcinogenic responses (ECSCF, 2001; IARC, 2012; JECFA, 2001; UKCOT, 2001; USEPA, 2012). As a result, REP estimates based on responses measured in tissues outside the liver may be of more direct relevance to current risk assessment for humans. Therefore, it is of interest to examine if disposition in extra-hepatic tissues is primarily governed by their lipid content. The general assumption is that lipid-adjusted concentrations in humans are approximately equivalent between plasma lipids and adipose tissue (Van den Berg et al., 1994). However human studies are limited and there is a need to determine such a relationship in more detail between human and rodent species and on a congenerspecific basis.

Thus, selection of the most appropriate values of systemicREPs for application to potential human responses based on measured serum concentrations requires consideration of factors related to both tissue concentration and response metrics. The goals of this paper are to examine the SYSTEQ data in this context. Specifically, this paper presents tissue distribution data across the tested compounds and dose levels, and we compared these data to previous studies in rodents using both single and subchronic dosing regimens. Furthermore, the concentrationâ&#x20AC;&#x201C;response data were evaluated to ascertain whether the short time frame in the SYSTEQ study (3 days from dosing to sacrifice) was sufficient to allow full development of the selected responses to determine REPs. Based on these evaluations, we provide considerations for evaluating and selecting appropriate systemicREPs from these data in the context of applicability for human risk assessment.

92


Tissue distribution of DLCs: Impacts Systemic REPs

Materials and methods SYSTEQ data The methods and data collection for the SYSTEQ project have been described in detail elsewhere (van Ede et al., 2013a; 2013b). Briefly, female C57Bl/6 mice and Sprague-Dawley rats were administered a single oral dose (6 animals/dose) of 2,3,7,8-tetrachlorodibenzodioxin (TCDD), 1,2,3,7,8-pentachlorodibenzodioxin (PeCDD), 2,3,4,7,8, pentachlorodibenzofuran (4-PeCDF), 3,3’,4,4’,5-pentachlorobiphenyl (PCB 126), 2,3’,4,4’,5-pentachlorobiphenyl (PCB 118), 2,3,3’,4,4’,5-hexachlorobiphenyl (PCB 156) and 2,2’,4,4’,5,5’-hexachlorobiphenyl (PCB 153) dissolved in corn oil. The doses varied from 0.5 μg/kg bw (TCDD) up to 500 mg/kg bw (PCB 153), spanning a similar range of administered TEQ across congeners based on the 2006 WHO-TEF values (see Table 1). The PCBs were cleaned prior to analysis to avoid contamination with dioxinlike compounds, as described by van Ede et al. (2013a,b). Animals were sacrificed on day 3 following administration and tissues were collected, snap frozen and stored until use at −80 °C. Tissue concentrations were analyzed using single ion monitoring mode on a Hewlett Packard 5890 GC coupled to a Fisons Instruments VG Autospec HRMS. Lipid content was determined gravimetrically in liver, adipose tissue, and plasma. Response metrics included measurement of hepatic ethoxyresorufin-O-deethylase (EROD) activity, hepatic Cyp1a1, Cyp1a2 and Cyp1b1 mRNA, and Cyp1a1 and Cyp1b1 mRNA in peripheral blood lymphocytes. Table 1: Congeners, TEF-values and dose ranges used within the EU-SYSTEQ project for C57Bl/6 mice and Sprague-Dawley rats Congener TCDD

PeCDD

4-PeCDF

PCB-126

PCB-118

PCB-156

PCB-153

TEF 1

1

0.3

0.1

0.00003

0.00003

ND

Single oral dose (µg/kg bw) 0.5

0.5

5

5

5000

5000

5000

2.5

2.5

25

25

15000

15000

15000

10

10

100

100

50000

50000

50000

25

25

250

250

150000

150000

150000

100

100

1000

1000

500000

500000

500000

93

4


Dose–response curves and effect concentrations at 50% response (EC50) were obtained using a sigmoidal dose–response nonlinear regression curve fit with variable slope (GraphPad Prism 6.01, GraphPad Software Inc., San Diego, CA):

Literature data Literature data was collected from a variety of studies employing a single or subchronic dosing regimens in rats or mice that reported measured tissue concentrations in liver, adipose, and blood or plasma. Also studies providing data on tissue distribution of DLCs in humans were identified. See Table 2 for more details about the studies used. We evaluated tissue distribution behavior in two main ways: as the calculated liver:adipose tissue concentration, and as the ratio of lipid-adjusted concentration in adipose tissue compared to lipid-adjusted concentrations in plasma, serum, or whole blood. For ratios that were calculated based on mean concentrations ± standard deviation (SD), the SDs on the ratios were estimated as follows; Ratio (z) =

SDz =

We also collected available data from the literature on dose–response curves, EC50 values and time course for hepatic CYP1A1 activity and/or gene expression for TCDD in rat or mouse (See Table 3 for more details). These studies were compared with the SYSTEQ dose–response curves and EC50 values to assess whether responses measured 3 days following a single oral dose occurred at comparable systemic concentrations to those measured following subchronic administration protocols.

94


Tissue distribution of DLCs: Impacts Systemic REPs

Results Tissue distribution of PCDD, PCDF and PCBs; single oral dose vs. subchronic administration The patterns of distribution between liver and adipose tissue based on wet weight for rats and mice across congeners from single and subchronic studies are displayed in Figs. 1 and 2. Diliberto et al. have suggested that liver:adipose concentration ratios in excess of approximately 0.3 signal some degree of hepatic sequestration, beyond that expected simply due to lipid content of hepatic tissues (Diliberto et al., 1997).

In rats, notable dose-dependent hepatic sequestration occurs for TCDD, PeCDD, 4-PeCDF, and PCB 126. In general, the highest liver:adipose ratios were seen for 4-PeCDF, which were between 4.7 and 58. Ratios between 1.6-18, 0.4-13 and 0.7-5 were found for PeCDD, PCB 126 and TCDD, respectively depending on the dose studied (Fig. 1). Ratios below 0.3 that indicate no significant sequestration were observed for the mono–ortho PCB 118 and the non dioxin-like PCB 153. In general, the dose-dependency and degree of hepatic sequestration observed in the data from the current SYSTEQ study, employing an oral single-dose protocol, are quite similar to those reported for the 14 weeks (sub) chronic NTP studies for TCDD, 4-PeCDF, PCB 118, and PCB 153 (National Toxicology Program, 2006a; 2006b; 2006c; 2006d; 2010) and for PeCDD compared to a 30-day oral administration study (Budinsky et al., 2008). Greater hepatic sequestration of PCB 126 was observed in the SYSTEQ data compared to the rat NTP study (National Toxicology Program, 2006c). However, the SYSTEQ liver:adipose ratios were more consistent with those observed in another subchronic study with PCB 126 by Chu et al. (1994). In a single dose mixture study with pregnant Long-Evans rats, Chen et al. observed a much higher degree of hepatic sequestration for 4-PeCDF than observed in the SYSTEQ study (Chen et al., 2001). However, the latter study used a mixture of DLCs causing a higher effective total dioxin-like compound hepatic disposition than reflected by the 4-PeCDF dose only. In the Chen et al. (2001) study it can be expected that mixture of DLCs caused a higher CYP1A2 induction followed by a higher degree of 4-PeCDF hepatic sequestration than expected for a single compound study. Similar results were found for mice in studies employing both single dose and subchronic administration protocols (Fig. 2). Notable dose-dependent hepatic sequestration occurred for TCDD, PeCDD, 4-PeCDF, and PCB 126, with liver:adipose concentration ratios between 0.2–4.1, 4.3–6.8, 6.5–47 and 0.4–9.2, respectively. Again, no evidence of dose-dependent hepatic sequestration was observed for the mono–ortho PCB 118 and 156, or for the non dioxin-like PCB 153. The degree of hepatic sequestration observed in the SYSTEQ dataset is similar to that reported following subchronic administration by 95

4


96

Sprague-Dawley Chronic administration by oral gavage, with tissue (female) concentrations measured at weeks 14, 31, 53, and 104

SYSTEQ (current study)

van Birgelen et al. (1994)

Single dose by oral gavage on GD 15, tissue concentrations measured on GD 16, 21, and PND 4.a,b

Sprague-Dawley Single dose by oral gavage, sacrifice on day 3 (female)

Sprague-Dawley Subchronic administration in diet, 13 weeks (female)

Long-Evans (female)

NTP studies (2006a-d; 2010)

Hurst et al. (2000b)

Subchronic administration by oral gavage, 13 weeks prior to and through gestation; tissue concentrations measured on GD 16 and 21.a

Long-Evans (female)

Hurst et al. (2000a)

Sprague-Dawley Subchronic administration in diet, 13 weeks (male & female)

Single dose by oral gavage on GD 15, measured concentrations on GD 16, 21, and PND 4. Compounds administered as a mixture.a,b

Long-Evans (female)

Chen et al. (2001)

Chu et al. (1994)

Dosing regimen

Sprague-Dawley Subchronic adminstration in native soil mixed with feed or in prepared corn oil gavage for 30 days. (female) Compounds administered as one of two mixtures.

Strain

Budinsky et al. (2008)

Rat

Study

Table 2: Studies included in distribution comparisons

TCDD, PeCDD, 4-PeCDF, PCB 126, PCB 118, PCB 156, PCB 153

PCB 126

TCDD, 4-PeCDF, PCB 126, PCB 118, PCB 153

TCDD

TCDD

PCB 126

TCDD, PeCDD, TCDF, 1-PeCDF, 4-PeCDF, OCDF, PCB 77, PCB 126, PCB 169

Mixture 1: TCDD, PeCDD, 123678-HxCDD, 1234678-HpCDD, 4-PeCDF Mixture 2: TCDF, 12378-PeCDF, 4-PeCDF, 123478-HxCDF, 123678-HxCDF

Compound(s)


GD = gestational day PND = postnatal day

b

a

Weistrand and Noren (1998)

Watanabe et al. (2013)

Thoma et al. (1990)

Schecter et al. (1991)

Human Iida et al. (1999)

SYSTEQ (current study)

Diliberto et al. (2001)

Diliberto et al. (1999)

Mouse DeVito et al. (1998)

Study

C57Bl/6 (female)

B6C3F1 (female)

C57Bl/6N (male)

B6C3F1 (female)

Strain

Chronic environmental

Chronic environmental

Chronic environmental

Chronic environmental

Chronic environmental

Multiple PCB congeners

17 PCDD/Fs, 4 non-ortho PCBs, 8 mono-ortho PCBs

17 PCDD/Fs

17 PCDD/Fs

17 PCDD/Fs; PCB 77, PCB 126, PCB 169

TCDD, PeCDD, 4-PeCDF, PCB 126, PCB 118, PCB 156, PCB 153

TCDD

Subchronic administration by oral gavage, 13 weeks Single oral dose, sacrifice on day 3

TCDD, 4-PeCDF, PCB 153

TCDD, PeCDD, TCDF, 1-PeCDF, 4-PeCDF, OCDF, PCB 126, PCB 169, PCB 105, PCB 118, PCB 156

Compound(s)

Single dose by oral gavage, sacrifice on day 4

Subchronic administration by oral gavage, 13 weeks

Dosing regimen

Tissue distribution of DLCs: Impacts Systemic REPs

4

97




$ !"%"%"

$ !"%"%"

  

 









      

       











$ !"%"%"

$ !"%"%"

 

    







!" ""!!#







 





 

!" ""!!#

 



!" ""!!#

$ !"%"%"

$ !"%"%"

!" ""!!#



   !" ""!!#



  





 

   !" ""!!#



Figure 1. Liver to adipose tissue concentration ratios for TCDD, PeCDD, 4-PeCDF, PCB 126, PCB 118 and PCB 153 in rats from various studies employing a single or subchronic oral dosing regimen. See Table 2 for more details on the studies included in figures. Symbol legend: SYSTEQ (u), Chen et al. 2001; GD21 data only (n), Hurst et al. 2000b (p), NTP; 14 week data only (ÂĄ), Hurst et al. 2000a (r), Budinsky et al. 2008 (ÂŁ), Chu et al. 1994 (ÂŻ), van Birgelen et al. 1994 (s). Filled symbols denote a single oral dose regimen; open symbols denote a subchronic dosing regimen. SYSTEQ data represents the mean Âą SD of 6 rats. Literature data represent the mean Âą SD as described in Materials and Methods.

98


     !" ""!!#

$ !"%"%"

 

 

 



    

    !" ""!!#

    











$ !"%"%"

!" ""!!#

$ !"%"%"



$ !"%"%"



 

 

$ !"%"%"

 

$ !"%"%"

$ !"%"%"

Tissue distribution of DLCs: Impacts Systemic REPs

  

   

   !" ""!!#



   

   !" ""!!#



4

   

   !" ""!!#



     

    !" ""!!#

Figure 2. Liver to adipose tissue concentration ratios for TCDD, PeCDD, 4-PeCDF, PCB 126, PCB 118, PCB 156 and PCB 153 in mice from various studies employing a single or subchronic oral dosing regimen. See Table 2 for more details on the studies included in figures. Symbol legend: SYSTEQ (u), Diliberto et al. 1999 (p), DeVito et al. 1998 (ÂĄ), Diliberto et al. 2001 (r). Filled symbols denote a single oral dose regimen; open symbols denote a subchronic dosing regimen. SYSTEQ data represents the mean Âą SD of 6 mice. Literature data represent the mean Âą SD as described in Materials and Methods. Data from DeVito et al.1998 represents the mean. 99


DeVito et al. for all tested compounds except 4-PeCDF (Devito et al., 1998). In this study, 4-PeCDF displayed high hepatic sequestration (liver:adipose ratios greater than 40) at the two highest dose levels following subchronic administration in mice.

To our knowledge, only four datasets allow evaluation of liver:adipose tissue concentrations of DLCs in humans based on autopsy samples from the general population (Iida et al., 1999; Thoma et al., 1990; Watanabe et al., 2013; Weistrand and NorĂŠn, 1998). In Fig. 3 liver:adipose tissue concentrations ratios of these studies are shown based on wet weight. Depending on the congener, these ratios range differ by approximately one order of magnitude with the higher hexa- to octachlorinated congeners having a value near 1 and lower chlorinated congeners closer to 0.1. These data suggest that only modest, if any, hepatic CYP1A2 induction with concomitant protein binding and hepatic sequestration occurs in humans at dioxin and furan exposure levels that are applicable to the general population. In contrast, in five individuals highly exposed to levels of PCDFs and PCBs in the Yusho rice poisoning incident, liver:adipose ratios for PCDFs were higher in those individuals with highly elevated adipose tissue concentrations (data tabulated in Carrier et al. (1995)).

Liver:Adipose ratio (wet wt)

10

Iida (n=8)

Thoma (n=28)

Watanabe (n=8 to 22)

Weistrand and Noren (n=7)

1

0.1

11 8

16 9

12 6

77

F

F

15 3 PC B

PC B

PC B

PC B

PC B

OC D

F

Hp CD

Hx CD

Su m

Su m

4P eC DF

TC DF

OC DD

Hp CD D

CD D

Su m

Hx

D TC D

Pe CD D

0.01

Figure 3. Liver to adipose tissue concentration ratios for several dioxin-like compounds and PCBs from human studies. Data represent the mean Âą SD.

100


Tissue distribution of DLCs: Impacts Systemic REPs

Disposition in extra-hepatic tissues; adipose to blood serum or plasma ratios in humans and rodents Limited datasets allow evaluation of distribution patterns in extra-hepatic tissues. While lipid-adjusted blood concentrations are commonly reported in humans due to the recognition of the potential impact of serum lipid content on wet weight congener concentrations in blood, this approach has not been widely used in studies of distribution in animals. The SYSTEQ project measured lipid content of liver, adipose, and plasma allowing congener concentrations in these matrices to be examined on both a wet weight and lipid-adjusted basis. Chen et al. also assessed tissue lipid content and reported concentrations on both a wet weight and lipid-adjusted basis (Chen et al., 2001). Finally, two human studies report results from analysis of paired human blood and adipose tissues samples across a range of dioxin, furan, and PCB congeners (Iida et al., 1999; Schecter et al., 1991). The congener-by-congener ratios of adipose to blood serum or plasma lipid-adjusted concentrations for those studies are illustrated in Fig. 4.

Adipose:Blood ratio (lipid-adj)

10

Human, Iida (n=8)

Human, Schecter (n=20)

Mouse, SYSTEQ

Rat, SYSTEQ

Rat, Chen

4

1

0.1

16 9 PC B

12 6 PC B

77 B PC

O

CD

F

F D

F D

H pC Su m

F

F

H xC Su m

4P eC D

D TC

D CD O

D D H pC

xC

D D

D H Su m

Pe CD

TC

D D

0.01

Figure 4. Lipid-adjusted adipose to blood plasma or blood serum concentration ratios of several dioxin-like compounds and PCBs from humans, mice and rats. Data represent the mean Âą SD.

In general, ratios cluster around 1 on a lipid adjusted basis and do not seem to differ between rodent and human studies, even though in these rodent studies a single-dose administration protocol was employed with exposure and tissue levels far higher than those for humans in the general population (SYSTEQ; Chen et al., 2001). These data suggest that extra-hepatic tissue distribution is relatively rapid and complete regardless 101


of administration protocol and is not dose-dependent.

Dose–response curves; single oral dose vs. subchronic administration The results observed following use of a single oral dose administration protocol in the SYSTEQ project with measurement of the selected responses after three days can be compared with tissue dose–response relationships observed following subchronic exposure, which is more similar to the human background exposure situation. Data on dose–response relationships (e.g. tissue EC50 concentrations for key responses) and time course of the selected responses from SYSTEQ and earlier subchronic studies can provide clarity in this context. Induction of specific enzyme activity or gene expression as a response to DLCs has been commonly used as a sensitive biomarker in in vivo studies (Budinsky et al., 2006; Chu et al., 1994; DeVito et al., 1997; Diliberto et al., 1999; 2001; Drahushuk et al., 1996; Van Birgelen et al., 1994; 1995b; 1996; VanDen Heuvel et al., 1994). To calculate EC50 values on a tissue concentration basis, tissue concentrations as well as full dose–response curves are needed. Only a few studies have reported these types of datasets, often including TCDD only. Relevant datasets for TCDD are summarized in Table 3.

At present, the most comprehensive dataset available is that of DeVito et al., describing systemic–response relationships in mice for numerous DLCs after subchronic administration (DeVito et al., 1997). In this study, the EC50 for hepatic CYP1A1 activity for TCDD was approximately 5 ng/g liver. Other studies in rats and mice, including SYSTEQ data, have reported EC50 values in the same range for induction of hepatic CYP1A1 by TCDD, irrespective of the administration protocol (DeVito et al., 1997; Diliberto et al., 2001; Drahushuk et al., 1996; Van Birgelen et al., 1995a; VanDen Heuvel et al., 1994). Thus, regardless of the use of a single or multiple oral dosages, it is shown that EC50 values for induction of hepatic CYP1A1 by TCDD are comparable if based on tissue concentrations. In view of the limited role of metabolism and elimination of many other toxic DLCs, such comparability between both types of studies can also be expected for other DLCs.

Limited information is available on the time course of the induction of hepatic P450 enzymes. Santostefano et al. evaluated the time course of hepatic CYP1A1, 1A2, and 1B1 induction in rats following a single dose of TCDD (Santostefano et al., 1997). Based on protein levels, the highest induction was observed three days after oral dosage. Fisher et al. examined the time- and dose-dependent induction of hepatic enzymes in rats after a single oral dose of PCB 126 (Fisher et al., 2006). It was found that CYP1A1 activity reached a maximum response between 1 and 5 days after a single oral dose exposure, which is comparable to TCDD. Similar data examining time course of response for other

102


Sprague Dawley (female)

B6C3F1 (female)

C57bl/6 (female)

B6C3F1 (female) Single dose, 3 days

Subchronic, 13 weeks

Subchronic, 13 weeks

Single dose, 3 days

Single dose, 4 days

Subchronic, 13 weeks

Chronic (response at 14 weeks)

Single dose, 24h

Dosing regimen

6a

4.9

EROD CYP1A1 mRNA

EROD

5.2c 10.7c

1.1-3.4a

<4.8c 4.6c EROD

EROD CYP1A1 mRNA

0.4 â&#x20AC;&#x201C; 7a 7a

EROD CYP1A1 mRNA

1.2b

EROD

EROD

0.7-9a 0.7-9a 0.7-9a

Estimated EC50 ng TCDD/g tissue

EROD CYP1A1 protein CYP1A1 mRNA

Endpoint

b

a

No formal EC50 calculation presented; range estimated based on inspection of reported tissue levels and responses. Based on modeling reported in Toyoshiba et al. (2004): ED50 of 5 ng/kg-d external dose, and interpolated corresponding hepatic wet weight concentration of TCDD reported in NTP (2006a). c EC50 concentration is derived from dose-response curve using the Hill slope equation, see Materials and Methods for more details.

SYSTEQ (current study)

Diliberto et al. (2001)

Sprague Dawley (female)

Mouse DeVito et al. (1997)

SYSTEQ (current study)

Sprague Dawley (female)

vanden Heuvel et al. (1994)

Sprague-Dawley (female)

van Birgelen et al. (1995)

NTP (2006a)

Sprague-Dawley (male)

Strain

Rat Drahushuk et al. (1996)

Study

Table 3: Estimated EC50 values based on hepatic concentrations for hepatic CYP1A1 activity, protein and mRNA induction in rodents in vivo systems.

Tissue distribution of DLCs: Impacts Systemic REPs

103

4


congeners is lacking. If the dynamics of hepatic enzyme responses to the other tested congeners differs substantially from the dynamics of response to TCDD, this could influence the validity of the estimated systemicREPs. Discussion Within the SYSTEQ project, systemicREPs for PeCDD, 4-PeCDF, PCB 126, PCB 118 and PCB 156 were calculated based on plasma, liver and adipose tissue concentration in rats and mice (van Ede et al., 2013a; 2013b). However, tissue concentration and body burden of a congener can be influenced by congener-specific toxicokinetics. This raises the question whether systemicREPs calculated based on a single-dose protocol as applied in the SYSTEQ project can provide an appropriate prediction of REPs following a chronic exposure situation, which is most relevant for human risk assessment. This question is particularly of interest given the highly persistent nature of these compounds.

In the present study, we compared tissue distribution data for different congeners as well as EC50 values of TCDD for hepatic endpoints from the SYSTEQ project, applying a single-dose regimen, with earlier studies employing single and subchronic dosing regimens. This comparison shows that for these compounds, a single dose and subchronic exposure generally resulted in similar body distribution and EC50 values based on liver concentration of hepatic endpoints in rats and mice. This means that the liver to adipose ratios and induction of hepatic CYP1A1 expression and activity were generally comparable no matter which dosing regime was applied within rodent studies among the different DLCs studied. The more potent AhR agonists, TCDD, PeCDD, 4-PeCDF and PCB 126, show higher liver sequestration compared to the mono–ortho PCBs 118, 156 and the non dioxin-like PCB 153. This phenomenon is generally attributed to a higher hepatic CYP1A2 induction of the more potent AhR agonists combined with binding to this protein (DeVito et al., 1997; 2000). Currently, it is unclear if these compounds when bound to the CYP1A2 protein can easily become bioavailable to activate the AhR and cause dioxin-like responses. REPs calculated on total hepatic tissue concentration, instead of the “free” available concentrations, may lead to either an over- or under-estimation of the potency of a congener, depending on the relative degree of hepatic sequestration compared to TCDD. 4-PeCDF, sequesters in the liver to a much greater degree than TCDD, reflecting a significantly higher degree of binding to the CYP1A2 protein (Devito et al., 1998; Diliberto et al., 1999; Yoshimura et al., 1984). Assuming that protein-bound 4-PeCDF is not available for AhR activation, an EC50 value for hepatic responses based on total

104


Tissue distribution of DLCs: Impacts Systemic REPs

hepatic concentration would be misleadingly high, resulting in a lower calculated REP compared to TCDD. In contrast, PCB 118 and PCB 156 may display little affinity for the CYP1A2 protein and do not significantly sequester in the liver. As a result both PCB congeners can look more potent compared to TCDD when based on hepatic concentrations because all or nearly all of the compound present in liver is available for interaction with the AhR.

Very minor hepatic sequestration occurs within the human population exposed to environmentally relevant concentrations (Iida et al., 1999; Thoma et al., 1990; Watanabe et al., 2013; Weistrand and NorĂŠn, 1998). See also Fig. 3. It seems likely that background exposure levels are just not high enough to induce CYP1A2 induction and subsequent hepatic sequestration for e.g. TCDD and 4-PeCDF in humans. This is supported by studies using the caffeine breath test as measure for CYP1A2 activity (Abraham et al., 2002). However, theoretically, congener-specific hepatic sequestration in humans is possible as CYP1A2 is one of the more prominent P450 enzymes present in the liver (Bieche et al., 2007). Indeed, some studies in populations with elevated exposure does demonstrate alterations in caffeine metabolism suggesting CYP1A2 induction (Abraham et al., 2002; Lambert et al., 2006). In extrahepatic tissues, there is relatively little CYP1A2 protein available (Bieche et al., 2007). In the view of this, selecting blood as matrix to identify tissue concentrations for calculating relative potencies may reflect more directly the available concentration causing an AhR response, because potential sequestration of compounds due to CYP1A2 protein binding is less likely an issue. Furthermore, toxic responses that are most relevant for the human population are generally responses in extra-hepatic tissues, and the available systemic concentrations are generally those of blood rather than hepatic tissues. This suggests that the use of extra-hepatic responses with blood serum/plasma concentrations may be more appropriate to determine systemic REPs for human risk assessment, reducing or eliminating the impact of congenerspecific differences in hepatic CYP1A2 induction and binding. In conclusion, this study shows that distribution patterns are generally similar for TCDD, PeCDD, 4-PeCDF, PCB 126, PCB 118, PCB 156 and PCB 153 between studies using a single dose or subchronic dosing. Furthermore, the responding concentration for TCDD in single dose studies is comparable to the responding concentrations reported in subchronic studies. In contrast with data for laboratory rodents, available distribution data for humans in the general population display little or no hepatic sequestration. Therefore, calculating systemicREPs based on total hepatic concentration and responses could result in REPs that are not fully applicable to the relevant toxic endpoints and systemic exposure measures in most human studies. SystemicREPs based on blood serum/ plasma concentration with an extra-hepatic response might give a more accurate 105

4


prediction of the relative potency of a congener for humans under environmental exposure conditions.

106


Tissue distribution of DLCs: Impacts Systemic REPs

4

107


Part

III

Rodent- versus human-REPs

Man is distinguished from all other creatures by the faculty of laughter.

Joseph Addison


Chapter

5

Differential Relative Effect Potencies of Some Dioxin-like Compounds in Human Peripheral Blood Lymphocytes and Murine Splenic Cells Karin I. van Ede Konrad P.J. Gaisch Martin van den Berg Majorie B.M. van Duursen

Endocrine toxicology group, Institute for Risk Assessment Sciences (IRAS), Utrecht University, the Netherlands

Toxicology Letters 226: 43 â&#x20AC;&#x201C; 52 (2014)


Abstract Human risk assessment for dioxin-like compounds is typically based on the concentration measured in blood serum multiplied by their assigned toxic equivalency factor (TEF). Consequently, the actual value of the TEF is very important for accurate human risk assessment. In this study we investigated the effect potencies of 3 polychlorinated dibenzop-dioxins (PCDDs), 6 polychlorinated dibenzofurans (PCDFs) and 10 polychlorinated biphenyls (PCBs) relative to the reference congener 2,3,7,8-tetrachloro-dibenzop-dioxin (TCDD) in in vitro exposed primary human peripheral blood lymphocytes (PBLs) and mouse splenic cells. REPs were determined based on cytochrome P450 (CYP) 1A1, 1B1 and aryl hydrocarbon receptor repressor (AhRR) gene expression as well as CYP1A1 activity in human PBLs and Cyp1a1 gene expression in murine splenic cells. Estimated median human REPs for 1,2,3,4,6,7,8-heptachlorodibenzop-dioxin (1234678-HpCDD), 2,3,4,7,8,-pentachlorodibenzofuran (23478PeCDF), 1,2,3,4,7,8-hexachlorodibenzofuran (123478-HxCDF) and 1,2,3,4,7,8,9-heptachlorodibenzofuran (1234789-HpCDF) were with 0.1, 1.1, 1 and 0.09 respectively, significantly higher compared to those estimated for mouse with REPs of 0.05, 0.45, 0.09 and 0.04, respectively. Opposite to these results, the estimated median human REP of 3,3’,4,4’,5-pentachlorobiphenyl (PCB 126), was with 0.001 30-fold lower compared to the mouse REP of 0.03. Furthermore, human REPs for 1234678-HpCDD, 23478-PeCDF, 123478-HxCDF, 1234789-HpCDF and PCB 126 were all outside the ± half log uncertainty range that is taken into account in the WHO-assigned TEFs. Together, these data show congener- and species-specific differences in REPs for some, but not all dioxin-like congeners tested. This suggests that, more emphasis should be placed on human-tissue derived REPs in the establishment of a TEF for human risk assessment.

112


Differential REPs of DLCs in human and murine lymphocytes

Introduction

T

he estimation of human risk for polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs) and polychlorinated biphenyls (PCBs) is typically based on the concentration measured in blood serum multiplied by their toxic equivalency factor (TEF) assigned by the World Health Organization (Van den Berg et al., 2006). Consequently, the actual value of the TEF is crucial for accurate human risk assessment. Each congener-specific TEF expresses the relative aryl hydrocarbon receptor (AhR)-mediated potency of a dioxinlike compound (DLC) compared to 2,3,7,8-tetrachloro-dibenzo-p-dioxin (TCDD), the most potent and well-studied congener. Although each TEF is derived from a large number of relative effect potencies (REPs), these REPs are primarily based on rodent in vivo and in vitro data (Haws et al., 2006). However, human in vitro models show that the potency of a number of congeners may differ from those derived from rodent studies. For example, 2,3,4,7,8-pentachlorodibenzofuran (23478-PeCDF, WHOTEF 0.3), 1,2,3,4,7,8-hexachlorodibenzofuran (123478-HxCDF, WHO-TEF 0.1) and 1,2,3,6,7,8-hexachlorodibenzofuran (123678-HxCDF, WHO-TEF 0.1) were found to be as potent as TCDD in human lymphoblastoid cells reporting aryl hydrocarbon hydroxylase (AHH)-inducing potency or for induction of cytochrome P450 1A1 (CYP1A1) gene expression in human keratinocytes (Nagayama et al., 1985; Sutter et al., 2010). Also, the potency of 3,3â&#x20AC;&#x2122;,4,4â&#x20AC;&#x2122;,5-pentachlorobiphenyl (PCB 126) for ethoxyresorufin-Odeethylase (EROD) activity or CYP1A1 mRNA induction in primary human hepatocytes, keratinocytes, peripheral blood lymphocytes (PBLs) and human hepatoblastoma cells (HepG2) is generally up to 100-fold lower than expected based on its WHO-TEF of 0.1 (Silkworth et al., 2005; Sutter et al., 2010; Van Duursen et al., 2005; Zeiger et al., 2001). However, even though these species-specific differences in potency, in particular for PCB 126, were acknowledged by the expert panel during the WHO-TEF re-evaluation in 2005, it was concluded that more information regarding the difference between rodents and humans is needed (Van den Berg et al., 2006). Furthermore, in aforementioned studies only a few congeners assigned with a TEF-value have been analyzed, while these species-specific differences in potency might also concern other DLCs. Within this study, we determined species-specific differences in potency for CYP1A1 activity and CYP1A1, CYP1B1 and aryl hydrocarbon receptor repressor (AhRR) gene expression for 20 selected congeners consisting of four PCDDs, six PCDFs, eight dioxin-like PCBs and two non-dioxin-like (NDL) PCBs in primary human PBLs and murine splenic cells. As human peripheral blood is easy to collect, it is an interesting matrix for monitoring human health. Changes in AhR-mediated gene expressions in PBLs are widely used as biomarkers of human exposure to DLCs and polycyclic aromatic hydrocarbons (PAHs), despite the uncertainties and interindividual variability in their responses (Guida et 113

5


al., 2013; Hanaoka et al., 2002; Hu et al., 2006; McHale et al., 2007; Van Duursen et al., 2005). Furthermore, present concerns regarding responses in human populations upon DLC exposure are more and more focused on extra-hepatic responses, including (neuro) development, reproductive functions, immunotoxicity and extra-hepatic carcinogenic responses (ECSCF, 2001; IARC, 2012; JECFA, 2001; Lauby-Secretan et al., 2013; UKCOT, 2001; USEPA, 2012). Consequently, studies with respect to human responses that focus on extra-hepatic tissues might be of more interest from a human risk assessment point of view.

In this study, the 20 selected congeners were divided into two groups. Group one consisted of seven congeners, i.e. TCDD, 1,2,3,7,8-pentachlorodibenzodioxin (12378-PeCDD), 23478-PeCDF, PCB 126, 2,3’,4,4’,5-pentachlorodiphenyl (PCB118), 2,3,3’,4,4’,5-hexachlorobiphenyl (PCB 156) and the NDL-PCB, 2,2’,4,4’,5,5’hexachlorobiphenyl (PCB 153) that represent approximately 90% of the dioxin-like activity in the human food chain (Liem et al., 2000). These congeners have also been studied in an in vivo mouse study, where congener-specific REPs in liver and PBLs were calculated based on administered dose as well as based on liver, adipose tissue and blood plasma concentrations (van Ede et al., 2013a). This allows us to compare in vitro with in vivo derived REPs based on either the administered dose or systemic concentrations. A second group of 13 congeners consisted of two PCDDs, five PCDFs, five DL-PCBs and one NDL-PCB, which are commonly found in human tissues and the food chain, but are of lower toxicological meaning. Materials and Methods Chemicals TCDD, 12378-PeCDD, 1,2,3,6,7,8-hexachlorodibenzo-p-dioxin (123678-HxCDD), 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin (1234678-HpCDD), 2,3,7,8-tetrachlorodibenzofuran (2378-TCDF), 23478-PeCDF, 123478-HxCDF, 2,3,4,6,7,8-hexachlorodibenzofuran (234678-HxCDF), 1,2,3,4,6,7,8-heptachlorodibenzofuran (1234678HpCDF), 1,2,3,4,7,8,9-heptachlorodibenzofuran (1234789-HpCDF) and PCB 126 were purchased from Wellington Laboratories Inc. (Guelph, Ontario, Canada). PCB118, PCB156 and PCB153 were purchased from Cerilliant Corp. (Round Rock, TX, USA). 2,4,4’,5-tetrachlorobiphenyl (PCB 74), 3,3’,4,4’-tetrachlorobiphenyl (PCB 77), 2,3,3’,4,4’-pentachlorobiphenyl (PCB 105), 2,3’,4,4’,5,5’-hexachlorobiphenyl (PCB 167), 3,3’,4,4’,5,5’-hexachlorobiphenyl (PCB 169), 2,3,3’,4,4’,5,5’-heptachlorobiphenyl (PCB 189) were purchased from Larodan Fine Chemicals (Malmö, Sweden). All congeners had a purity > 99% except for 1234678-HpCDD (98.7%). The congeners were dissolved 114


Differential REPs of DLCs in human and murine lymphocytes

and diluted in dimethyl sulfoxide (DMSO) (Sigma-Aldrich, Stockholm, Sweden).

Cell preparation, culture and exposure Human buffy coat from 11 healthy volunteers consisting of six male (age 24-, 43-, 46-, 55-, 65-, 66-years old) and five female (age 25-, 25-, 38-, 46-, 57-years old) all living in The Netherlands were obtained from Sanquin Blood Supply (Sanquin Blood Supply, Amsterdam, The Netherlands). The study was evaluated and approved by the Sanquin Executive Board and a written informed consent was obtained from all donors. PBLs were isolated using Ficoll-Paque gradient according to the manufacturer’s instructions (GE Healthcare Europe, Diegem, Belgium). Murine splenic cells were isolated from 10-week old female C57Bl/6 mice purchased from Harlan laboratories (Venray, The Netherlands). Animals were euthanized by CO2/O2 and spleens were removed. To obtain a single cell spleen suspension, spleens were pressed through a 70µm cell strainer (BD Biosciences, Bedford, MA, USA) and red blood cells were lysed with lysis reagent (containing 155 mM NH4Cl, 10 mM KHCO3, 0.1 mM EDTA, pH 7.8). The animals were handled in a humane manner and the study was approved by the Animal Ethical Committee (DEC Utrecht, Utrecht, The Netherlands).

Human PBLs were suspended in culture medium consisting of phenol red-free RPMI 1640 supplemented with 10% fetal bovine serum (FBS) (Invitrogen, Breda, the Netherlands), 100 U/mL penicillin, 100 µg/mL streptomycin (Invitrogen) and 1.5% phytohaemagglutin (PHA) (Life Technologies, Bleiswijk, the Netherlands). Murine splenic cells were suspended in culture medium consisting of phenol red-free RPMI 1640 supplemented with 10% fetal bovine serum (FBS) (Invitrogen), 100 U/mL penicillin, 100 µg/mL streptomycin (Invitrogen) and 5 µg/mL Concanavalin A (Con A) (Calbiochem, Merck Millipore, Darmstadt, Germany). Cell concentrations were determined using a Beckman Coulter Counter (Beckman Coulter, Woerden, The Netherlands) and cell number was adjusted to 4 x 106 cells/mL. For ethoxyresorufinO-deethylase (EROD) activity, 500 µL cell suspension was seeded onto 24-well plates (Costar, Cambridge, MA, USA), while for gene expression analysis, 1 mL cell suspension was seeded onto 12-well plates (Costar). Standard curves of the 20 selected PCDDs, PCDFs and PCBs were prepared in culture medium containing twice the desired concentration. For EROD activity, 500 µL exposure medium was added in duplicate and for gene expression 1 mL exposure medium was added in duplicate. This resulted in a final solvent concentration of 0.1% v/v DMSO with the following concentrations of the congeners: TCDD, 12378-PeCDD and 23478-PeCDF (0.1, 0.25, 1, 2.5 and 10 nM), 123678-HxCDD (1, 2.5, 5, 10 and 25 nM) PCB 126, 2378-TCDF, 1234678-HpCDD, 123478-HxCDF, 234678-HxCDF, 1234678-HpCDF and 1234789-HpCDF (1, 2.5, 10, 25 and 100 nM), PCB 74, 77, 105, 167, 169 and 189 (0.5, 1, 2.5, 5 and 10 µM), PCB 118, 156, 115

5


153 (0.25, 1, 2.5, 10 and 25 µM).

To determine the optimal time point for CYP1A1 activity as well as CYP1A1 gene expression, human PBLs from two donors and splenic cells from two mice were exposed to 10nM TCDD for 1, 2, 4, 24, 48 and 72 hours in two independent experiments. Results showed for human PBLs a maximum induction for CYP1A1 activity as well as CYP1A1 gene expression after 48 hours. For mice splenic cells, maximum Cyp1a1 gene expression was reached after 2 hours (data not shown). Based on these data it was decided to expose human PBLs and mice splenic cells for 48- and 2-hours, respectively.

For gene expression, TCDD, 12378-PeCDD, 23478-PeCDF, PCB 126, 118, 156 and NDL-PCB 153 were tested in two independent experiments where each experiment consisted of human PBLs from one donor or pooled splenic cells from 18 mice. The other congeners were firstly screened for potency differences in induction of gene expression between the mouse and human model as well as deviations from the assigned WHOTEF values using PBLs from one human donor or pooled splenic cells from 18 mice. From this screening, the following congeners were selected to determine effects on EROD activity: TCDD, 12378-PeCDD, 23478-PeCDF, PCB 126, 118, 156, NDL-PCB 153, 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF. Effects on EROD activity by these selected congeners was tested in human PBLs of six individual donors within one experiment.

RNA isolation, quantitative real-time PCR Total RNA was isolated from human PBLs or murine splenic cells using a QIAGEN RNeasy kit (QIAGEN, Venlo, The Netherlands). Purity and concentration of the isolated RNA was determined by measuring the absorbance ratio at 260/280 nm and 230/260 nm with a Nanodrop 2000 spectrophotometer (Thermo Scientific, Asheville, NC, USA). RNA was reverse transcribed to complementary DNA (cDNA) using the iScript cDNA synthesis Kit (Bio-Rad, Veenendaal, the Netherlands). Quantitative real-time PCR analyses were performed using the iQ Real-Time PCR Detection System with SYBR green (Bio-Rad). Amplification reactions were set up with 15 µL mastermix containing 12.5 µL iQ SYBR Green Supermix (Bio-Rad), 0.5 µL distilled H2O, 1 µL (10 µM) forward primer, 1 µL (10 µM) reverse primer, and 10 µL first strand cDNA (10X diluted). Primer sequences were as follows: human CYP1A1 (NM_000499): forward-5’CAGAAGATGGTCAAGGAGCA-3’ and reverse-5’-GACATTGGCGTTCTCATCC-3’ (Andersson et al., 2011); human CYP1B1 (NM_000104): forward-5’-CGGCCACTATCACTGACATC-3’ and reverse-5’- CTCGAGTCTGCACATCAGGA-3’ (Andersson et al., 2011); human AhRR (NM_020731.4): forward-5’- CGCTGCTTCATCTGCCGTGT-3’ and reverse-5’CTGCATCGTCATGAGTGGCTCG-3’ (designed using the Primer designing tool (NCBI)); 116


Differential REPs of DLCs in human and murine lymphocytes

human β-actin (NM_001101.3): forward-5’-TTGTTACAGGAAGTCCCTTGCC-3’ and reverse-5’-ATGCTATCACCTCCCCTGTGTG-3’ (designed using the Primer designing tool (NCBI)); mouse Cyp1a1 (NM_009992.4): forward-5’-GGTTAACCATGACCGGGAACT-3’ and reverse 5’-TGCCCAAACCAAAGAGAGTGA-3’ (Schulz et al., 2012); mouse β-actin (NM_007393.3): forward-5’-ATGCTCCCCGGGCTGTAT-3’ and reverse-5’CATAGGAGTCCTTCTGACCCATTC-3’ (Schulz et al., 2012). All primers were run through the National Center for Biotechnology Information Primer-BLAST database (http:// www.ncbi.nlm.nih.gov/tools/primer-blast/) to confirm specificity and validated for optimal annealing temperature (60 °C for all primers) and efficiency in our laboratory (90 – 110%). The following program was used for denaturation and amplification of the cDNA: 3 min at 95 °C, followed by 40 cycles of 15 s at 95 °C and 45 s at 60 °C. Gene expression for each sample was expressed in terms of the threshold cycle (Ct), normalized to the reference gene β-actin ( Ct). Fold induction was calculated between the treated and vehicle-treated control group. EROD activity CYP1A1 activity in human PBLs was determined by means of EROD activity as described by Van Duursen et al. (2005). In contrast with human PBLs, EROD activity could not be determined in murine splenic cells or in murine PBLs. Even after variance experimental set-up changes (e.g. exposure time, cell density, Con-A concentration, ethoxyresorufin concentration and measuring time). Data analysis Dose-response curves were obtained using a sigmoidal dose-response nonlinear regression curve fit with variable slope (GraphPad Prism 6.01, GraphPad Software Inc., San Diego, CA) [1]. [1]

In this Hill equation, y is the dependent variable (mRNA induction or EROD activity), x the independent variable (concentration), E0 is the estimated background response, Emax is the maximum response, b is the estimated median effect concentration (EC50), and n is the shaping parameter of the Hill curve.

REPs were estimated as described previously by Van Ede et al. (2013). Briefly, effect concentrations were calculated at which a congener reached 20% response of TCDD (BMR20TCDD). Using the congener specific BMR20TCDD concentration REPs were calculated relatively to TCDD [2]. 117

5


[2] Results and Discussion Gene expression Probably the best-studied target genes for AhR-mediated effects upon exposure to DLCs are CYP1A1, CYP1B1 and AhRR. Upregulation of these genes precedes modifications of protein levels of for example CYP1A1 (Denison et al., 2011). Concentration-response curves of CYP1A1, CYP1B1 and AhRR gene expression in human PBLs as well as concentration-response curves of Cyp1a1 gene expression in murine splenic cells were determined upon exposure to DLCs (Figure 1, 2 and 3). Gene expression of CYP1A1, CYP1B1 and AhRR in human PBLs and Cyp1a1 gene expression in murine splenic cells was induced by all PCDDs and PCDFs tested. Although the maximal response of the different dioxins and furans was not always comparable to that of TCDD, it was for all congeners higher than 20% induction caused by TCDD. In contrast with the dioxins and furans, not all PCBs induced CYP1A1, CYP1B1 and AhRR gene expression in human PBLs or mice splenic cells. In human PBLs, mono-ortho PCBs 105 and 156 did not induce any of the genes tested. This lack of inducibility was similar to that observed for NDL-PCB 153. The other PCBs tested, did significantly induce gene expression. However, with the

exception of CYP1B1 gene expression, the maximal responses were low and generally below 20% induction of TCDD (See Figure 1 and 2G-I). A clear difference in response between human PBLs and murine splenic cells was seen for PCB 126. In human PBLs, PCB 126 only reached 20% induction of TCDD for AhRR gene expression in one donor, whereas in mouse murine splenic cells PCB 126 could induce Cyp1a1 gene expression as high as 70 to 90% of the maximal response caused by TCDD (See Figure 1 and 3A/C).

BMR20TCDD concentrations REPs that are driving the TEFs are typically calculated based on effect ratios between the individual congener and TCDD using the 50% effect concentrations (EC50). However, such estimations are only valid when the dose-response curves for the individual congeners are parallel to the standard curve (TCDD) and with a similar maximal response (Ymax). For nonparallel dose-response curves the ratio at a 50% effect can be significantly different from a ratio derived at another point in the curve, such as 20% or 80% (Villeneuve et al., 2000). There is no simple solution to deal with this issue, however some studies have suggested the use of other then EC50 values (Toyoshiba et al., 2004; Villeneuve et al., 2000). Villeneuve et al. (2000) describes a method to calculate REPs based on multiple point estimates over the range of response from EC20 to EC80. 118


Relative expression (fold induction over control)

0

20

40

60

80

100

0

20

40

60

-9

-11

-8

-7

-6

-9

-7

-6

Concentration (log M)

-10

-8

Concentration (log M)

-10

D. Donor 2, CYP1A1

-11

A. Donor 1, CYP1A1

-5

-5

-4

-4

2

4

6

8

10

12

14

16

18

20

2

4

6

8

10

12

-9

-11

-8

-7

-6

-9

-7

-6

Concentration (log M)

-10

-8

Concentration (log M)

-10

E. Donor 2, CYP1B1

-11

B. Donor 1, CYP1B1

-5

-5

-4

-4

2

4

6

8

10

12

14

16

18

20

2

4

6

8

10

12

-9

-11

-8

-7

-6

-9

-7

-6

Concentration (log M)

-10

-8

Concentration (log M)

-10

F. Donor 2, AhRR

-11

C. Donor 1, AhRR

-5

-5

-4

-4

TCDD 12378-PeCDD 23478-PeCDF PCB 126 PCB 118 PCB 156 PCB 153

Figure 1. Dose-response curves for CYP1A1 (A and D), CYP1B1 (B and E) and AhRR (C and F) gene expression of TCDD, 12378-PeCDD, 23478-PeCDF, PCB 126, PCB 118, PCB 156 and PCB 153 in human PBLs after 48 h exposure. Upper and lower lines represent two individual donors. Data are represented as mean Âą SD (n=2). BMR20TCDD is indicated with a black dotted line.

Relative expression (fold induction over control)

80

Differential REPs of DLCs in human and murine lymphocytes

5

119


Relative expression (fold induction over control)

200

30

A. CYP1A1

26

160

0

2 -11 -10

-9

-8

-7

-6

-5

-4

Relative expression (fold induction over control)

8 6 4 2 -9

-8

-7

-6

-5

-4

10

D. CYP1A1

240 200

34

E. CYP1B1

30

120 40 -8

-7

-6

-5

-4

6 2

-9

-8

-7

-6

-5

-4

-11 -10

-9

-8

-7

-6

-5

-4

Concentration (log M)

22

H. CYP1B1

I. AhRR

TCDD PCB 74 PCB 77 PCB 105 PCB 167 PCB 169 PCB 189

18

6

14

120

4

10

80 40 0

TCDD 123678-HxCDD 1234678-HpCDD 1234678-HpCDF 1234789-HpCDF

10

200 160

-4

14

-11 -10

8

-5

F. AhRR

Concentration (log M)

G. CYP1A1

-6

18

2 -9

-7

22

4

80

-8

26

6

160

-9

Concentration (log M)

8

-11 -10

-11 -10

Concentration (log M)

Concentration (log M)

Relative expression (fold induction over control)

10

-11 -10

Concentration (log M)

240

12

10 6

0

14

14

40

TCDD 2378-TCDF 123478-HxCDF 234678-HxCDF

C. AhRR

16

18

80

280

18

22

120

320

20

B. CYP1B1

6

2

2 -11 -10

-9

-8

-7

-6

Concentration (log M)

-5

-4

-11 -10

-9

-8

-7

-6

Concentration (log M)

-5

-4

-11 -10

-9

-8

-7

-6

-5

-4

Concentration (log M)

Figure 2. Dose-response curves for CYP1A1 (A, D and G), CYP1B1 (B, E and H) and AhRR (C, F and I) gene expression of TCDD, 2378-TCDF, 123478-HxCDF, 234678-HxCDF (upper line), TCDD, 123678-HxCDD, 1234678-HpCDD, 1234678-HpCDF, 1234789-HpCDF (middle line), TCDD, PCB 74, 77, 105, 167 169 and 189 (lower line) in human PBLs after 48 h exposure. Data were obtained from one experiment and are represented as mean Âą SD (n=2). BMR20TCDD is indicated with a black dotted line.

Toyoshiba et al. (2004) suggest three solutions to deal with non-parallel dose response curves: (I) Base the REPs on EDx values, where X is chosen such that the shape of the curve has less influence on the outcome of the REP, for example EC1; (II) Rescale the response to have equal maximum response estimates; (III) Choose a single reference response for TCDD (benchmark response) and compare ratios of predicted doses at the given response regardless agreement in shape. For this study, it was decided to choose the latter approach and make use of a benchmark response as earlier described by Van Ede et al. (2013). With this approach, concentrations were calculated at which a congener reached 20% induction caused by TCDD (BMR20TCDD). The BMR20TCDD value was preferred above a lower BMR value, e.g. 5% or 10% of TCDD max, because these would usually fall within the noise of the background or in the bend of the dose-response 120


Differential REPs of DLCs in human and murine lymphocytes

30

TCDD 12378-PeCDD 23478-PeCDF PCB 126 PCB 118 PCB 156 PCB 153

A.

25 20 15 10 5 0

-11 -10

-9

-8

-7

-6

-5

Relative expression (fold induction over control)

Relative expression (fold induction over control)

curves and not in the lower area of the linear part of the curve.

-4

40 35 30 25

234678-HxCDF 1234678-HpCDF 1234789-HpCDF

20 15 10 5 0

-11 -10

PCB 126

60

PCB 118 PCB 156

40

PCB 153

20

-11 -10

-9

-8

-7

-6

Concentration (log M)

-5

-4

Relative expression (fold induction over control)

Relative expression (fold induction over control)

TCDD 12378-PeCDD 23478-PeCDF

C.

80

0

-9

-8

-7

-6

-5

-4

Concentration (log M)

Concentration (log M) 100

TCDD 1234678-HpCDD 2378-TCDF 123478-HxCDF

B.

80

TCDD 123678-HxCDD

D.

PCB 74 PCB 77 PCB 105 PCB 167 PCB 169 PCB 189

60 40 20 0

-11 -10

-9

-8

-7

-6

-5

-4

Concentration (log M)

Figure 3. Dose-response curves for Cyp1a1 gene expression of TCDD, 12378-PeCDD, 23478-PeCDF, PCB 126, PCB 118, PCB 156 and PCB 153 (A and C), TCDD, 2378-TCDF, 123478-HxCDF, 234678-HxCDF, 1234678-HpCDD, 1234678-HpCDF, and 1234789-HpCDF (B), TCDD, 123678-HxCDD, PCB 74, 77, 105, 167, 169, and 189 (D) in mice splenic cells after 2 h exposure. Graph A and C represent two independent experiments of pooled splenic cells from 18 mice. Graph B and D represent one experiment of pooled splenic cells from 18 mice. Data are represented as mean Âą SD (n=2). BMR20TCDD is indicated with a black dotted line.

The BMR20TCDD concentrations for CYP1A1 induction by TCDD were similar for human PBLs and murine splenic cells, with 0.21 and 0.26 nM respectively (Table 1). Typically, humans are considered to be relatively insensitive towards dioxin-induced effects in contrast with most laboratory species (Connor and Aylward, 2006). This is mostly attributed to an, at least 10 times, less sensitive human AhR compared to rodent species (Black et al., 2012; Carlson et al., 2009; Silkworth et al., 2005; Wiebel et al., 1996; Xu et al., 2000). Also, in a study performed by Nohara et al. (2006), higher EC50 values for CYP1A1 mRNA induction by TCDD were found in in vitro exposed human lymphocytes compared with murine and rat lymphocytes, with EC50 values of 1.43 versus 0.33 and 0.14 nM, respectively. In contrast with our study, Nohara et al. (2006) did not activate lymphocytes with a mitogen, which has been shown to affect AhR-mediated responses (Whitlock Jr. et al., 1972). These differences in experimental set-up might explain why in our study human PBLs appear to be similarly sensitive as murine splenic cells to TCDD exposure. 121

5


122 CYP1B1 mRNA

AhRR mRNA

0.734

Mouse exp. 1

0.371

7.13

Mouse exp. 1

NA

NA

ND

ND

NA

NA

0.087

0.096

NA

NA

0.199

1.6

1.7

0.7

0.6

NA

NA

NA

ND

53.84

NA

NA

0.085

0.078

NA

NA

0.621

0.107

0.002

2.8

1.5

0.4

1.1

1

Mouse exp. 2

8,66

14.67

0.02

0.02

NA

0.948

0.08

NA

3.53

0.04

Donor 3

0.03

0.03

0.3

0.6

2.0

0.8

0.6

0.3

1.6

NA

0.267

NA

0.143

123678-HxCDD

9.10

ND

1

0.7

1

1

1

Mouse exp. 2

Donor 2

ND

0.973

Donor 1

Mouse exp. 2

0.104

0.245

Mouse exp. 1

Donor 2

Donor 1

0.487

0.133

Mouse exp. 2

Donor 2

0.292

0.300

Donor 1

Mouse exp. 2

NA

0.077

0.116

0.238

Congeners group 2

PCB-126

23478-PeCDF

12378-PeCDD

1

1

1

1

0.213

0.200

0.164

0.141

Mouse exp. 1

Donor 3

1

1

0.207

0.212

Donor 1

TCDD

Donor 2

0.1

0.1

0.3

1

1

BMR20TCDD (nM) REP BMR20TCDD (nM) REP BMR20TCDD (nM) REP WHO-TEFa

CYP1A1 mRNA

Congeners group 1

Table 1: BMR20TCDD concentrations and corresponding REPs for PCDDs, PCDFs and PCBs tested in human PBLs and mouse splenic cells.


Mouse exp. 2

Donor 3

Mouse exp. 2

Donor 3

Mouse exp. 2

Donor 3

Mouse exp. 2

Donor 3

Mouse exp. 2

Donor 3

Mouse exp. 1

Donor 3

Mouse exp. 1

Donor 3

Mouse exp. 1

Donor 2

Mouse exp. 1

Donor 2

Mouse exp. 1

Donor 2

Mouse exp. 1

Donor 3

ND

ND

ND

ND

ND

ND

ND

ND

ND

ND

4.79

0.710

572.8

37.63

2.54

4.51

2.35

0.146

0.331

1.13

4.64

1.33

0.04

0.3

0.0004

0.005

0.08

0.05

0.09

1.5

0.6

0.2

0.05

0.1

0.006

13.70

0.00005

NA

667.7

NA

779.5

NA

NA

0.0001

0.0001

0.00003

1665

2292

NA

0.000004

0.2

21556

NA

0.313

NA

NA

2.85

NA 0.05

1.4

NA

0.1

0.101

1.07

NA

0.2

0.493

NA

ND

NA

3037

NA

15537

NA

50359

NA

ND

NA

0.235

NA

22.85

NA

ND

NA

0.442

NA

0.920

NA

0.980

0.0001

0.01

0.00001

0.0001

-

0.01

0.01

0.1

0.1

0.1

0.01

0.00005

0.000009

0.000003

0.6

0.006

0.5

0.3

0.1

ND, not determined because BMR20TCDD was not reached; NA, not analysed. BMR20TCDD and REPs were calculated as described in “Materials and Methods”. a Current WHO-TEF (Van den Berg et al., 2006). For PCB 105, 118, 153 and 156, no BMR20TCDD and REPs could be determined for the various biomarkers tested in human PBL and mouse splenic cells, for this reason they are not presented in this table.

PCB 189

PCB 169

PCB 167

PCB 77

PCB 74

1234789-HpCDF

1234678-HpCDF

234678-HxCDF

123478-HxCDF

2378-TCDF

1234678-HpCDD

Differential REPs of DLCs in human and murine lymphocytes

123

5


Human REPs versus mouse REPs When calculating REPs based on these BMR20TCDD concentrations, the estimated human REPs show a rank order of 23478-PeCDF > 123478-HxCDF > TCDD ≈ 12378-PeCDD > 1234789-HpCDF > 1234678-HpCDD ≈  2378-TCDF > 123678-HxCDD ≈   234678-HxCDF > 1234678-HpCDF > PCB 126 > PCB 169 ≈   PCB 189 > PCB 77 ≈   PCB 167. Whereas the mouse BMR20TCDD-derived REPs were in the rank order TCDD > 12378-PeCDD ≈a   ≈   23478-PeCDF ≈  2378-TCDF > 123478-HxCDF ≈   234678-HxCDF > 1234678-HpCDDaa 1234789-HpCDF ≈   PCB 126 ≈  123678-HxCDD > 1234678-HpCDF. Most noticeable are the higher human PBL-derived REPs for 23478-PeCDF (0.8 – 2.8), 123478-HxCDF (0.5 – 1.5), 1234678-HpCDD (0.1 – 0.2), and 1234789-HpCDF (0.2 – 0.6) compared to those derived for the mouse with REPs for 23478-PeCDF (0.3 – 0.6), 123478-HxCDF (0.09), 1234678-HpCDD (0.05), and 1234789-HpCDF (0.04). In contrast, human REPs for PCB 126 could only be derived for AhRR gene expression in one donor and was with 0.002 lower compared to the mouse REPs (0.03). These results suggest that human PBLs are more sensitive for 23478-PeCDF, 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF and less sensitive for PCB-126 compared to mouse splenic cells. Another remarkable observation is that for CYP1A1 gene expression, none of the PCBs tested were capable to induce a response in human PBLs that was high enough to calculate a REP, while only PCB 126 could induce such a response in mouse splenocytes. These differences in response between dioxins and furans on the one side and PCBs on the other side may be due to differences in AhR binding mechanisms that are governed by the physic-chemical properties of either the dioxin or biphenyl structure (Petkov et al., 2010). It should also be pointed out that serum in culture medium could have an effect on the bioavailability and REPs of DLCs in an in vitro system (Hestermann et al., 2000). In our study, the serum concentrations in murine and human lymphocyte cultures were similar, therefore did not contribute to species-differences in potencies observed. The very low or lack of response by PCBs in human PBLs are in agreement with earlier studies using human primary cells or cell lines derived from liver, breast, prostate, lymphocytes, or keratinocytes (Endo et al., 2003; Silkworth et al., 2005; Spink et al., 2002; Sutter et al., 2010; Van Duursen et al., 2003; 2005; Zeiger et al., 2001). EROD activity Based on the observed differences in potency of the various DLCs in human PBLs and murine splenic cells, it was decided to select 123478-HxCDF, 1234678-HpCDD and 1234789-HpCDF together with the congeners TCDD, 12378-PeCDD, 23478-PeCDF, PCB 126, 118, 156 and 153 to determine REPs based on EROD activity in the PBLs of six human donors. EROD activity a highly sensitive indicator for AhR-mediated induction of CYP1A1 protein (Schrenk et al., 1995), it is also a faster measurement compared to gene expression analysis. EROD activity could not be determined in mouse splenic cells

124


Differential REPs of DLCs in human and murine lymphocytes

(See Materials and Methods).

EROD activity (pmol RSF/min/mg protein)

As described previously by Van Duursen et al. (2005), significant differences in minimum and maximum EROD activity between individual human donors were observed. For TCDD, these inter-individual differences are shown in Figure 4. 16

donor 1

14

donor 2 donor 3 donor 4

12 10

donor 5 donor 6

8 6 4 2 0 -2

-12

-11

-10

-9

-8

-7

Concentration (log M)

Figure 4. Dose-response curves for EROD activity of TCDD in human PBLs of six donors after 48 h exposure. Data are represented as mean Âą SD (n=2).

Similar variation in minimum and maximum responses between donors were also observed for the other congeners tested (data not shown). As can be seen in Figure 4, Donor 5 and 6, were less responsive to the DLCs tested and it was not possible to generate reliable dose-response curves for these donors and therefore excluded from further calculations. All congeners tested, except PCB 118, 156 and NDL-PCB 153, dosedependently induced EROD activity above the benchmark response of 20% TCDD in the remaining donors (See Figure 5). Within one individual donor the maximum responses for the PCDDs and PCDFs were generally similar to TCDD and between 10 to 14 pmol RSF/min/mg protein. In contrast, the maximum response of PCB 126 was with 3 to 4 pmol RSF/min/mg protein significantly lower and reached only 30 to 40% of the maximal response of TCDD. This observation is in agreement with consistently lower Ymax values for PCB 126-induced AhR-mediated responses in human models (Silkworth et al., 2005; Van Duursen et al., 2005; Zeiger et al., 2001). BMR20TCDD and corresponding REPs that have been calculated from the dose-response curves for EROD activity are presented in Table 2. Although deviations between BMR20TCDD concentrations are observed between donors, the REPs based on EROD activity for the different DLCs are 125

5


generally similar and comparable to the human REPs calculated for gene expressions.

EROD activity (pmol RSF / min / mg protein)

14

14

Donor 4

12

12

10

10

8

8

6

6

4

4

2

2

0

-12

-11

-10

-9

-8

-7

-6

-5

-4

0

Donor 7

123478-HxCDF 1234789-HpCDF PCB 126 PCB 118 PCB 156 PCB 153

-12

-11

EROD activity (pmol RSF / min / mg protein)

14

Donor 8

16

-9

-8

-7

-6

-5

-4

-6

-5

-4

Donor 9

12

14

10

12 10

8

8

6

6

4

4

2

2 0

-10

Concentration (log M)

Concentration (log M)

18

TCDD PeCDD 1234678-HpCDD 4-PeCDF

-12

-11

-10

-9

-8

-7

Concentration (log M)

-6

-5

-4

0

-12

-11

-10

-9

-8

-7

Concentration (log M)

Figure 5. Dose-response curves for EROD-activity of TCDD, 12378-PeCDD, 1234678-HpCDD, 23478-PeCDF, 123478-HxCDF, 1234789-HpCDF, PCB 126, PCB 118, PCB 156 and PCB 153 in human PBLs of 4 individual donors after 48 h exposure. For Donor 9, only the congeners TCDD, 12378-PeCDD, 23478-PeCDF, PCB 126, 118, 156 and PCB 153 were tested, due to a lower amount of PBLs. Data are represented as mean Âą SD (n=2). BMR20TCDD is indicated with a black dotted line.

REPs versus WHO-TEFs When comparing REPs from this study with WHO-TEFs, it is clear that based on gene expression the ranking order of mouse REPs is more in line with that of the WHO-TEFs than the human REPs determined in our study with PBLs (See Table 1) (Van den Berg et al., 2006). In Figure 6, the ratios between calculated human REPs and WHO-TEFs are shown for 12378-PeCDD, 23478-PeCDF, PCB 126, 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF. Here, a ratio of 1 indicates that a derived REP is similar to the WHO-TEF. Noticeable, median human REPs of 23478-PeCDF (median REP 1.1), 1234678-HpCDD (median REP 0.1), 123478-HxCDF (median REP 1), and 1234789-HpCDF (median REP 0.09) were 4 to 10 times higher than their WHO-TEFs. Moreover, these REPs were 126


Differential REPs of DLCs in human and murine lymphocytes

Table 2: BMR20TCDD concentrations and corresponding REPs for the 10 selected congeners derived from EROD activity in human PBLs.

Selected congeners TCDD

12378-PeCDD

23478-PeCDF

PCB-126

1234678-HpCDD

123478-HxCDF

1234789-HpCDF

EROD activity

REP WHO-TEFa BMR20TCDD (nM)  

Donor 4

0.186

Donor 9

0.151

Donor 7 Donor 8 Donor 4 Donor 7 Donor 8 Donor 9

Donor 4

Donor 7 Donor 8 Donor 9

Donor 4

Donor 7 Donor 8 Donor 9

Donor 4

Donor 7 Donor 8 Donor 9

Donor 4

Donor 7 Donor 8 Donor 9 Donor 4

Donor 7 Donor 8

Donor 9

0.103 0.055

1 1 1 1

0.168

1.1

0.072

2.1

0.081 0.046

0.087

0.135 0.088 0.213

1.3 1.2

2.2

0.8 0.6 0.7

29.18

0.006

165.5

0.001

141.8 31.10

0.733

1.526

0.001 0.002

0.1

0.1

0.01

0.9

NA

1.306

1.467 0.711 NA

0.1

1.3

0.1

0.115 0.143

0.3

0.01

0.07

NA

1

0.3

0.740

0.140

1

0.4

0.07 0.08

5

NA, not analysed because there were not enough PBLs. For PCB 118, 153 and 156, no BMR20TCDD and REPs could be determined, for this reason they are not presented in this table. BMR20TCDD and REPs were calculated as described in “Materials and Methods”. a Current WHO-TEF (Van den Berg et al., 2006).

127


outside the half log uncertainty range that is assumed for the WHO-TEF (Van den Berg et al., 2006).

The fact that 123678-HxCDD, 2378-TCDF, 234678-HxCDF and 1234678-HpCDF do not show the same deviation between human PBLs and mouse splenic cells as well as towards their WHO-TEF suggests that there is a species- and congener-specific difference in REPs. Another study with human keratinocytes found similar results, with 10-fold higher REPs for 123678-HxCDF compared to its WHO-TEF of 0.1 and no deviation from the WHO-TEF for 2378-TCDF and 123678-HxCDD (Sutter et al., 2010). For PCB 126, the median human REP in our study was 0.001, which is 100 times lower than the WHO-TEF and far outside the suggested uncertainty range (Figure 6). Several other studies have also shown that the potency of PCB 126 is approximately a 100-fold lower in human in vitro models compared to its assigned WHO-TEF of 0.1 (Carlson et al., 2009; Silkworth et al., 2005; Sutter et al., 2010; Van Duursen et al., 2003; 2005; Westerink et al., 2008; Zeiger et al., 2001). A toxicogenomic study where primary rat and human hepatocytes were exposed to PCB 126 indicated that only five of the 4000 orthologous genes tested were shared between the rodent species and humans (Carlson et al., 2009). Ratio REPs / WHO-TEFs

100 10 1 0.1 0.01

0.001

C Pe

DD

eC 4-P

DF

26 DF DF DD xC pC B-1 pC H H C H P 9 8 78 78 67 34 34 34 12 12 12

Figure 6. Ratios between REPs determined in human PBLs for EROD activity (closed square) and gene expression (open square) and their assigned WHO-TEFs. Each symbol represents an individual donor. For gene expression, symbol represents the mean ratio REPs for the biomarkers determined. The black lines represent the median of the REPs. Gray shaded area represents the half log uncertainty range around the WHO-TEF.

In this study, REPs have been calculated based on a molar concentration, as is done in most in vitro studies. In contrast, the WHO-TEFs are mostly derived from mass-based 128


Differential REPs of DLCs in human and murine lymphocytes

REPs. Human risk assessment is often based on biomonitoring data, which is generally expressed in terms of mass concentrations. The REPs from this study based on molar concentrations could be modified by the ratio of the molecular weights of TCDD and the other tested congeners. Within this study, differences between molar and mass-based REPs would be at most a factor of 0.75 (for 1234678-HpCDD). With respect to the half log uncertainty range assumed to apply to TEF values, this difference is considered negligible. In vitro REPs versus in vivo systemic REPs Translating in vitro derived REPs to an in vivo situation is challenging, as pharmacokinetic properties cannot be taken into account. However, human risk assessment of DLCs is often based on blood concentrations, rather than on the administered dose. This means that relative potencies determined at the target tissue, like in in vitro studies, may give a better prediction of the actual potency of a congener in an in vivo situation when based on systemic concentrations. This might be in particular true for congeners like 23478-PeCDF, for which the hepatic disposition due to strong CYP1A2 binding is very different compared to the reference compound TCDD. In a single dose in vivo study with C57Bl/6 mice performed in our lab, we compared REPs of DLCs based on administered dose with those calculated based on liver, adipose tissue or blood plasma levels (van Ede et al. 2013a). We found for 23478-PeCDF, a 10-fold higher systemic REP based on plasma levels compared to those based on administered dose. It is noticeable that the mean REPs established in our in vitro study for 23478-PeCDF, 12378-PeCDD and PCB126 with mouse primary splenic cells are with 0.5, 0.5 and 0.03 respectively, similar to their mean systemic REPs based on plasma levels in the mouse in vivo study with 0.4, 0.8 and 0.02, respectively (Van Ede et al. 2013a). These similarities in REPs might indicate that in vitro derived REPs can potentially be used as a surrogate for in vivo systemic derived REPs. Conclusion All together, these data show congener- and species-specific differences in REPs between the mouse and human for some DLCs. Our study again confirms the possibility that the present WHO-TEF for PCB 126 may significantly overestimate its potency for humans. In addition, we also showed that the human PBL-derived REPs of 23478-PeCDF, 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF deviate from those observed in the mouse splenocyte model and from their WHO-TEFs. The results from this study indicate that more emphasis should be placed on human-tissue derived REPs in the establishment of a TEF for human risk assessment. For that, additional studies including other human tissues and endpoints would be desirable. 129

5


Chapter

6

In Vitro and In Silico Derived Relative Effect Potencies of Ahreceptor Mediated Effects by PCDD/Fs and PCBs in Human, Rat, Mouse and Guinea pig CALUX Cell-lines Karin I van Ede‡*, Mehdi Ghorbanzadeh†*, Malin Larsson†, Majorie BM van Duursen‡, Lorenz Poellinger§, Sandra Lücke§, Miroslav Machala#, Kateřina Pěnčíková#, Jan Vondráček#, Martin van den Berg‡, Michael S Denison┴, Tine Ringsted†, Patrik L Andersson†*

Endocrine toxicology group, Institute for Risk Assessment Sciences (IRAS), Utrecht University, the Netherlands † Department of Chemistry, Umeå University, Umeå, Sweden § Institute of Environmental Medicine, Karolinska Institute, Stockholm, Sweden # Department of Chemistry and Toxicology, Veterinary Research Institute, Brno, Czech Republic ┴ Department of Environmental Toxicology, University of California, Davis, California ‡

* Both authors contributed equally to this study

Manuscript in preparation


Abstract For a better understanding of species-specific relative effect potencies (REPs), responses of 20 dioxin-like compounds (DLCs) were assessed using chemical-activated luciferase gene expression assays (CALUX) derived from rat, mouse, guinea pig and human cell lines. These data show that polychlorinated dibenzo-p-dioxin (PCDD), polychlorinated dibenzofuran (PCDF) and polychlorinated biphenyl (PCB)-mediated responses in the human CALUX cell line differ significantly from responses in the rat, mouse and guinea pig derived CALUX cell lines. The human cell line is the least sensitive as indicated by the 20% effect concentrations of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) that were 1.5, 5.6, 11.0 and 190.0 pM for guinea pig, rat, mouse and human cells, respectively. Also apparent congener-specific species differences in potency were observed between human and rodent CALUX cell lines which was most clearly reflected by a lower human REP for PCB 126 (0.003) compared to guinea pig (0.2), rat (0.07) and mouse (0.05). Quantitative structure-activity relationship (QSAR) models were developed using orthogonal projections to latent structures and a variety of calculated and measured chemical descriptors. These models show that electronic properties and molecular surface characteristics play an important role in aryl hydrocarbon receptor (AhR) binding of the studied congeners. Furthermore, the human QSAR model showed different critical descriptors compared to the rodent QSAR models. This might indicate that the ligand-receptor interaction is different between the human and the rat, mouse and guinea pig cells. The present study established in vitro REPs for 18 congeners assigned with a WHO-TEF value in rodents and human species and in silico rodent-REPs for all congeners assigned with a TEF, which will aid to improve risk assessment of DLCs for humans and the environment.

132


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

Introduction

P

olychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs) and polychlorinated biphenyls (PCBs) include a range of highly toxic and persistent environmental pollutants originating from industrial products and combustion activities. In total, there are theoretically 209 PCB and 210 PCDD/F congeners based on the number of chlorine atoms and their positions on the aromatic rings. Owing to their chemical characteristics, high resistance to biodegradation and high lipophilicity, these compounds are widely distributed in the environment and human food chain (Cleverly et al., 2007; Liem et al., 2000; Schecter et al., 1998). Exposure to PCDDs, PCDFs and dioxin-like PCBs can cause a wide variety of adverse health effects including (neuro)developmental defects, endocrine disruption, skin toxicity, immune deficiencies and carcinogenic responses (Bavithra et al., 2012; Charnley and Kimbrough, 2006; Eubig et al., 2010; Safe et al., 1985a; 1986; Schantz et al., 2001; White and Birnbaum, 2009). Most, if not all, biological effects of these dioxinlike compounds (DLCs) are mediated through a common mechanism of action initiated by binding to and activation of the aryl hydrocarbon receptor (AhR) (Denison et al., 2011; Hankinson, 1995b; Okey et al., 1994; Safe, 1993; Sewall and Lucier, 1995). Risk assessment of DLCs is challenging since these compounds exist in the environment as complex mixtures. In order to simplify risk assessment for this class of compounds the toxic equivalency (TEQ) concept has been developed. The TEQ value of a sample reflects the overall toxicity due to DLCs and is the sum of congener-specific toxic equivalency factors (TEFs) multiplied by the concentration in a matrix, such as blood. In total 29 PCDDs, PCDFs and PCBs have been assigned with a TEF value by the World Health Organization (WHO) (Van den Berg et al., 1998; 2006). This means that those compounds must (1) have some similarity in structure to the 2,3,7,8 substituted PCDDs and PCDFs, (2) bind to and activate the AhR, (3) be persistent and accumulate in the food chain, and (4) show AhR-mediated biological/toxic response (Ahlborg and Hanberg, 1994; Van den Berg et al., 2006). Each TEF value is derived from multiple toxic and biologic relative effect potencies (REPs) of an individual DLC compared to the most potent congener, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (Van den Berg et al., 2006).

While TEF values for DLCs are mainly derived from REPs determined in in vivo and in vitro animal studies, they are widely used for human risk assessment with the prerequisite that human REPs are comparable to those derived from animal studies. Yet, there is information from human in vitro models indicating that, for some DLCs, the REPs may be significantly different compared to those derived from animal studies (Silkworth et 133

6


al., 2005; Sutter et al., 2010; Van Duursen et al., 2005). However, in these studies only a few congeners were tested. Currently, more knowledge is needed on species sensitivity to DLCs, in particular with respect to differences between humans and experimental animal species. The use of mouse, rat, guinea pig and human recombinant cell lines containing an AhR responsive reporter gene (firefly luciferase) in combination with quantitative structure-activity relationship (QSAR) analysis can potentially provide more insight on this issue. A QSAR represents a statistical model that quantifies the relationship between the structures of the compounds and the corresponding biological activity. The model provides a prediction of the biological activity of structurally similar but untested compounds as well as discovering structural analogies that influence the activity of a group of compounds. A number of QSAR models have been reported to estimate different biochemical and toxicological responses for PCBs, PCDFs and PCDDs (Almenningen et al., 1985; Almlof, 1974; Bandiera et al., 1983; Cheney and Tolly, 1979; Dynes et al., 1985; Field et al., 1985; Hafelinger and Regelman, 1985; Li et al., 2011; McKinney and Singh, 1981; Mekenyan et al., 1996; Safe et al., 1985b; Tsuzuki et al., 1988; Tuppurainen and Ruuskanen, 2000; Van Der Burght et al., 1999; Van der Burght et al., 2000). In the present study, the potencies of a set of 20 selected PCDD/Fs and PCBs were determined using AhR-dependent luciferase reporter gene bioassays from rat, mouse and human hepatoma cells, and guinea pig intestinal adenocarcinoma cells (Denison et al., 2004). Based on the resulting in vitro data, species sensitivity and variation were examined using effect concentration ratio plots and principal component analysis (PCA). QSAR models were developed to relate the calculated REP values of the tested compounds with the calculated descriptors using orthogonal projection to latent structures (OPLS) to finally predict the REPs for the DLCs that have been assigned with a TEF value by the WHO. The most significant descriptors of the derived models were identified to study differences in their structure-activity relationships in the tested species. Finally, derived REP values were compared and discussed in relation to their assigned TEF values. Materials and Methods Chemicals A set of four PCDDs, six PCDFs and ten PCBs, were selected based on TEF values, number of chlorine atoms, substitution pattern, and environmental abundance. In addition two non-dioxin like (NDL) PCBs (PCB74 and PCB153) were selected. Selected compounds are displayed in Figure S1 of the Supporting Information. 2,3,7,8-tetrachlorodibenzop-dioxin (TCDD), 1,2,3,7,8-pentachlorodibenzo-p-dioxin (12378-PeCDD), 1,2,3,6,7,8hexachlorodibenzo-p-dioxin (123678-HxCDD), 1,2,3,4,6,7,8-heptachlorodibenzo-p-

134


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

dioxin (1234678-HpCDD), 2,3,7,8-tetrachlorodibenzofuran (TCDF), 2,3,4,7,8,-pentachlorodibenzofuran (23478-PeCDF), 1,2,3,4,7,8-hexachlorodibenzofuran (123478HxCDF), 2,3,4,6,7,8-hexachlorodibenzofuran (234678-HxCDF), 1,2,3,4,6,7,8-heptachlorodibenzofuran (1234678-HpCDF), 1,2,3,4,7,8,9-heptachlorodibenzofuran (1234789-HpCDF) and 3,3’,4,4’,5-pentachlorobiphenyl (PCB126) were purchased from Wellington Laboratories Inc. (Guelph, Ontario, Canada). 2,3’,4,4’,5-pentachlorobiphenyl (PCB118), 2,3,3’,4,4’,5-hexachlorobiphenyl (PCB156) and 2,2’,4,4’,5,5’-hexachlorobiphenyl (PCB153) were purchased from Cerilliant Corp. (Round Rock, TX, USA). 2,4,4’,5-tetrachlorobiphenyl (PCB74), 3,3’,4,4’-tetrachlorobiphenyl (PCB77), 2,3,3’,4,4’-pentachlorobiphenyl (PCB105), 2,3’,4,4’,5,5’-hexachlorobiphenyl (PCB167), 3,3’,4,4’,5,5’-hexachlorobiphenyl (PCB169), 2,3,3’,4,4’,5,5’-heptachlorobiphenyl (PCB189) were purchased from Larodan Fine Chemicals (Malmö, Sweden). All congeners had a purity > 99% except for 1234678-HpCDD (98.7%). The congeners were dissolved and diluted in dimethyl sulfoxide (DMSO) (Sigma-Aldrich, Stockholm, Sweden).

Molecular descriptors The 3D molecular structures of the compounds were constructed using the software Scigress program (Scigress Version 2.2.0., 2008). All molecular structures were geometrically optimized using the Austin Model 1 (AM1), a semi empirical method incorporated in the MO-G application of the software Scigress. Prior to the geometry optimization the initial dihedral angle was set; 44° for non-ortho (no) PCBs and 50° for mono-ortho PCBs based upon crystallographic data of the PCBs (Li et al., 2011). The 2,3,7,8 substituted PCDD/Fs were optimized with the same procedure, but with a planar structures. The chemical descriptors included in the current study are related to molecular size as starting point, conformation, connectivity, hydrophobicity, and electronic properties. Detailed information on all 98 calculated and measured descriptors has been descripted earlier by Larsson et al. and only a brief summary will be given here (Larsson et al., 2013). The two-dimensional molecular descriptors size, conformation and connectivity were calculated in MOE (MOE 2006.08., 2008) and the octanol-water partition coefficient (log Kow) from KowWIN (www.epa.gov). Included three-dimensional molecular descriptors were dipole moments, molecular orbital (MO) energies, atom-specific electron density coefficients of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO), atomic electrostatic potential charges, atom-specific nucleophilic, electrophilic and radical susceptibility. Note that the atom-specific descriptors are calculated for the lateral positions of the three chemical classes, i.e. positions 2,3,7,8 and 345/3’4’5’ for the PCDD/Fs and PCBs, respectively. This was done to compare these three groups of compounds (due to the structural differences in the chemical skeletons, Figure S1 135

6


of the Supporting Information) and to capture atomic specific characteristics of the lateral positions, which are critical for AhR mediated responses (Safe, 1986). Due to the different number of lateral positions for these chemical classes, the highest and lowest values concerning these positions were used as descriptors. Calculations for the electronic descriptors were performed in Scigress using AM1 (MO energies, dipole moment, susceptibilities) and in Gaussian 09 suite of programs using B3-LYP 6-31G** (MO energies, dipole moment, atomic ESP charges) (Gaussian 09 Revision A.1, 2009). From the MO energies, the differences between the energy of the two highest HOMO (EHOMO, EHOMO-1) energies and the LUMO (ELUMO) energy were created (GAP and GAP1). The experimental digitalized ultraviolet (UV) absorption spectra were previously measured in our laboratory for all studied compounds in the range from 200 to 350 nm and used as descriptors to describe molecular size and substitution pattern related properties (Andersson et al., 1996; Larsson et al., 2013). The earlier not published UV spectra of PCBs 74 and 153 are included in Supporting Information (Figure S2).

Biological data The biological data used in this study to build QSAR models was based on effect concentrations and REPs determined in chemically activated luciferase expression (CALUX) bioassay systems containing a stably integrated DRE-driven firefly luciferase reporter gene. In total five different cell lines were used of four different species, namely, rat, mouse, guinea pig and human. The rat hepatoma (H4IIe) cells and guinea pig intestinal adenocarcinoma cells (GPC16) contain the stably transfected plasmid pGudLuc 1.1, whereas, the mouse hepatoma (Hepa1c1c7) cells contain the stably transfected plasmid pGudLuc 6.1 (Garrison et al., 1996; Han et al., 2004). The names of the rat, guinea pig and mouse, clonal cell lines are H4L1.1c4, G16L1.1c8 and H1L6.1c2, respectively. The pGudLuc1.1/6.1 plasmids contain the luciferase reporter gene under AhR-dependent control of 4 xenobiotic responsive elements. Two human CALUX bioassays were used. Human hepatocellular carcinoma cells (HepG2) were stably transfected with an AhRcontrolled luciferase reporter gene construct of either pGL-4.27-DRE (AZ-AhR cells) or a pTX.DIR luciferase reporter under the control of two xenobiotic response elements of the rat CYP1A1 gene (HepG2-XRE-Luc) (Berghard et al., 1993; Novotna et al., 2011). The H4L1.1c4, H1L6.1c2 and G16L1.1c2 cell lines were cultured in a-MEM culture medium (Gibco / Invitrogen, Breda, The Netherlands) supplemented with 10% fetal bovine serum (FBS) (Gibco / Invitrogen, Breda, The Netherlands), 50 IU/mL penicillin and 50 mg/mL streptomycin (Gibco / Invitrogen, Breda, The Netherlands). The human AZ-AhR cells were cultured in Dulbeccoâ&#x20AC;&#x2122;s modified Eagle medium (Life Technologies, Carlsbad, CA, USA), supplemented with 10% FBS (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA), 24 mM NaHCO3 (Sigma-Aldrich), 10 mM HEPES (Sigma-Aldrich), non-essential amino acids (Sigma-Aldrich) and 40 mg/ml gentamicin sulfate (Life Technologies).

136


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

HepG2-XRE-Luc cells were propagated in RPMI 1640 supplemented with 10% FBS, 100 IU/mL penicillin and 100 µg/mL streptomycin as well as 800µg/ml geneticin. All cell culture materials for HepG2-XRE-Luc cells were purchased from Life Technologies (Carlsbad, CA). The cell lines were grown confluent in white clear-bottomed 96 well microplates (Costar, Cambridge, MA, USA) at 37°C in a humidified 5% CO2 atmosphere. Standard curves of the 20 selected PCDDs, PCDFs and PCBs were prepared in culture medium containing twice the desired concentration. For exposure, 100µl was added in triplicate to the 96MW-plate containing 100µl medium. The outer edge of the MW-plate was filled with medium only to avoid concentration differences due to evaporation. The final DMSO concentration was 0.1% v/v with the following concentration ranges of the congeners: TCDD, PeCDD and 23478-PeCDF (0.0005 – 1 nM), 2378-TCDF, 123478-HxCDF, 234678-HxCDF and PCB 126 (0.005 – 10 nM), 123678-HxCDD (0.005 – 25 nM), 1234678-HpCDD, 1234678-HpCDF and 1234789-HpCDF (0.05 – 100 nM), PCB 169 (0.005 – 1000 nM), PCB 77 and PCB 189 (10 – 5000 nM), PCB 74, PCB 105, PCB 118, PCB 153, PCB 156, PCB 167 (10 – 10000 nM). For the G16L1.1c8 cell line, some congeners were exposed with a different concentration range; TCDF, 123478-HxCDF, 234678-HxCDF, 1234678-HpCDD, 1234678-HpCDF, 1234789-HpCDF (0.0005 – 1 nM), PCB 169 (0.05 – 50 nM) and PCB 77 (0.5 – 500 nM). In each experiment a reference curve of TCDD was included. After an exposure period of 24 h, cells were washed with phosphate buffered saline (PBS) and lysed with lysis reagent (Promega, Fitchburg, WI, USA, pH 7.8). For the H4L1.1c4, H1L6.1c2 and G16L1.1c2 cell line luciferase activity was measured 20 minutes after the cells were lysed using the Luminostar Optima from BMG Labtech (Offenburg, Germany). The human AZ-AhR cells were lysed for 15 minutes and stored at -80 °C until luciferase activity was measured on a luminometer using Luciferase Assay Kit (BioThema, Handen, Sweden) according to the manufacture’s recommendations. For the HepG2-XRE-Luc cells, luciferase activity was analyzed 15 minutes after the cells were lysed on a GloMax® luminometer (Promega) using Luciferase Assay Kit (BioThema, Handen, Sweden) according to the manufacture’s recommendations. Luciferase activity was normalized to total protein concentration of whole cell extracts as determined by a colometric method (Bio-Rad, Hercules, CA).

Dose-response modeling The dose-response curves were fitted by a four-parameter log-logistic model in GraphPad Prism version 6.00 (GraphPad Software, La Jolla California USA, www. graphpad.com) (Ritz, 2010). The equation used in GraphPad was the “log (agonist) vs. response -Variable slope” with a fixed bottom plateau set to 0. It should be noted that not all congeners had a similar Ymax or Hill slopes as seen for TCDD. This difference has a profound influence on the EC50 calculations, which generally form the basis for REP determination. Therefore, it was decided to calculate the concentration needed for 137

6


a congener to reach a benchmark response (BMR) of 20% and 50% of the maximum TCDD response (BMR20TCDD and BMR50TCDD) (Sebaugh, 2011). Prerequisites for BMR20TCDD and / or BMR50TCDD calculation;

• For BMR20TCDD, Ymax had to reach at least 25% of TCDD maximum response. • For BMR50TCDD, Ymax had to reach at least 55% of TCDD maximum response. • If maximum response did not reach a clear Ymax, top plateau was fixed at the Ymax of TCDD. • If the slope of a dose-response curve could not be defined, the slope was fixed to 1, assuming a one-to-one relationship between agonist and receptor (Wenner et al., 2011). • Coefficient of determination (R2) value of above 0.80.

The dose response curves of the 20 selected PCDDs, PCDFs and PCBs were defined by taking the average of two independent experiments in which each concentration was tested in triplicate (with the exception of PCB169 in rat and mouse where only one experiment could be used due to experimental circumstances). To exclude the background luciferase activity, the DMSO blank response was subtracted from the compound response.

Multivariate data analysis In order to develop QSAR models the multivariate OPLS method was applied, which uses the descriptor matrix X to predict the response matrix Y (Eriksson et al., 2012; Trygg and Wold, 2002). It is a modification of the partial least squares (PLS) method and it divides the systemic variation of X into two parts; one predictive variation correlated to Y and one orthogonal variation uncorrelated to Y. Compared to PLS, OPLS does not change the predictive power but improves model interpretation and reduces model complexity. The response values used to build QSAR models were log BMR20TCDD based REP (log REPBMR20TCDD) and log BMR50TCDD based REP (log REPBMR50TCDD). By definition the REP value was set to 1 for TCDD in every experimental model. The developed QSARs were evaluated by internal and external validation tests, and then applied to predict the response values of non-tested compounds. In addition, principal component analysis was applied to choose the training set compounds and to analyze the variation in the measured responses. With PCA one single matrix (X) is decomposed into the product of two smaller matrices, scores (T) and loadings (P), plus a matrix of residuals (E): X= TP´+E

(1)

The scores express the systemic behavior of the objects (here, compounds) and the 138


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

loadings comprise information on variables. The plot of orthogonal vectors of scores and loadings reflect the variation between the compounds and variables, respectively. The PCA and OPLS calculations were done using SIMCA version 13.0 software (Umetrics AB, Umea, Sweden) (Eriksson et al., 2012).

Training and validation sets The studied set of 20 PCDD/Fs and PCBs was split into a training set and a validation set. The training set of 12 compounds was selected based on the chemical diversity of the compounds as analyzed using PCA on the compiled set of chemical descriptors. Figure S3 of the supporting information displays the PCA score plot of the 31 compounds in the data set where each group of chemicals clustered together. All calculated descriptors (listed in Table S1) were used for the PCA, which resulted in a model with two significant principal components explaining 36% and 24%, of the variation in the data set, respectively. In order to have a diverse training set covering the whole chemical space, the congeners were selected from all three classes of compounds. In addition, compounds were selected from the different areas of the score plot including compounds with high and low PC1 and PC2 scores, respectively, to reach representatives with different number of chlorine atoms and from each chemical class. The training set consisted of six PCBs, four PCDFs and two PCDDs. As shown in Figure S3 the compounds of the training set were representative of each chemical class. The remaining eight tested compounds, including four PCBs, two PCDDs and two PCDFs, were used as validation set. The training set participated in the modeling process and the validation set was used to evaluate the predictive capacity of the resulting QSAR models.

Development and validation of QSAR models Models were developed including both responding and non-responding compounds. Non-responding compounds were assigned a REP one order of magnitude lower than the lowest REP calculated in the corresponding assay (Andersson et al., 2000; Harju et al., 2007) except in human CALUX model where the non-responding PCBs were assigned a REP two orders of magnitude lower (identical with lowest TEF value). This procedure was done for being able to model the chemistry of non-activity. The same training and validation sets were applied to develop and validate all QSAR models. The fitting of the models was assessed by the coefficient of determination (R2) and the root mean square error (RMSE). The ability to predict new compounds was evaluated by internal crossvalidation test and by using the external validation set. With cross-validation, a group of compounds is excluded from the model development and the developed model predicts the target values corresponding to the removed compounds. This procedure is repeated several times until each observation has been removed once and the predictive ability of the model is expressed as cross-validated explained variation (Q2). A calculated 139

6


Q2 value larger than 0.5 indicates that the developed model could be regarded as predictive (Golbraikh and Tropsha, 2002). In addition, the root mean square error of cross-validation (RMSEcv) was calculated. Based on the predictions of the validation set, the root mean square error of prediction (RMSEP) was also calculated. RMSEP is a measure of the predictive power of the developed model and is calculated as the standard deviation of the predicted residuals. Outliers in the models were searched using the model membership probability. It calculates the probability that a compound belongs to the model. With a confidence level of 0.95, a compound with a membership probability less than 0.05 is considered to be moderate outlier. Variable influence on projection (VIP) was used to show the importance of each chemical descriptor in the models. In order to find how the descriptors influence the developed models, the correlation plot for each important descriptor and the corresponding response value was investigated. The applicability domain of the developed models was analyzed as recommended by the organization for economic cooperation and development (OECD). The approach used to determine the applicability domain of the models was based on the membership probability. According to this method, a compound with a membership probability higher than 0.05 is considered as being inside the applicability domain of the model. Results and discussion CALUX assays The concentrations calculated for the 20 selected congeners to reach the benchmark response of 20% and 50% of TCDD maximum (BMR20TCDD and BMR50TCDD), as well as Ymax, for the rat, mouse, guinea pig and human CALUX assays are listed in Table 1 and exemplified with two dose-response curves of TCDD and PCB126 in Figure 1. In the guinea pig CALUX assay, all congeners, except the NDL PCB153, were able to induce AhRmediated luciferase activity. In the rat and mouse CALUX assay, all congeners, except the mono-ortho substituted PCB189 and the NDL PCB153, induced AhR-mediated luciferase activity. In contrast, both human bioassays showed an AhR-mediated effect for the PCDDs, PCDFs and PCB126; no induction was observed with the six mono-ortho substituted PCBs nor the non-ortho substituted PCBs 77 and 169. This difference in sensitivity towards PCBs clearly distinguished the rat, mouse and guinea pig based CALUX cell lines from the human CALUX cell lines. The maximum level of induction by the active DLCs (indicating differences in their efficacy as AhR activators) did not always reach the same Ymax as TCDD. However, in the rat, mouse, guinea pig and human AZ-AhR CALUX assays, the response was generally above 55% induction of TCDD and consequently the BMR20TCDD and BMR50TCDD could be calculated (Table 1). However, in 140


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

A) T C D D

B)

100

100

80

80

R e la tiv e e f f e c t ( % )

 

R e la tiv e e f f e c t (% )

the human HepG2-XRE-Luc CALUX assay only the responses of 1234678-HpCDD, TCDF, 23478-PeCDF and 123478-HxCDF were above the 55% induction level observed with TCDD and consequently the BMR50TCDD could only be calculated for these congeners. For all tested congeners in the human HepG2-XRE-Luc CALUX assay, except for PCB126 that reached a 14% of the TCDD Ymax, a BMR20TCDD could be calculated.

60 40 20

PC B 126

G u in e a p ig M ouse R at

60

H u m a n A Z -A h R H u m a n H e p G 2 -X R E -L u c

40 20 0

0 -4

-2

0

C o n c e n tr a tio n [ lo g ( n M ) ]

2

-4

-2

0

2

C o n c e n tr a tio n [ lo g ( n M ) ]

Figure 1. Dose-response curves of (a) TCDD and (b) PCB126 for rat H4L1.1c4, mouse H1L6.1c2, guinea pig G16L1.1c8, human AZ-AhR and human HepG2-XRE-Luc cells.

Figure 1 clearly illustrates that the human cell line is less sensitive to AhR-activation by TCDD as well as PCB126 compared to the rat, mouse and guinea pig cell lines. When BMR20TCDD ratios were compared between two species for all congeners tested, it showed that BMR20TCDD concentrations for guinea pig were one order of magnitude lower for the PCDDs and PCDFs tested and up to two orders of magnitude lower for the PCBs tested compared to rat and mouse (Figure S4a-b). Variation in the BMR20TCDD concentrations in the rat and mouse CALUX assay were within one order of magnitude. Generally, the BMR20TCDD for PCDDs and PCDFs were lower in mice and BMR20TCDD for PCBs, except PCB77, were higher in mice compared to rat (Figure S4c). In the human CALUX assay, BMR20TCDD concentrations were up to two orders of magnitude higher compared to rat, mouse and guinea pig CALUX assays for all congeners tested (Figure S4d-f). From the data presented in Figures 1 and S4, the rank order in sensitivity for the different CALUX assays is guinea pig > mouse ~ rat > human. These data are in excellent agreement with the lower affinity that the human AhR exhibits for TCDD in vitro compared to the C57Bl mouse AhR. In fact, the dissociation constant (Kd) of human AhR for TCDD is comparable to that of the AhR of TCDD-resistant DBA/2 mice (Connor and Aylward, 2006; Ema et al., 1994; Micka et al., 1997). In line with these observations, knock-in mice homozygous for the human AhR show considerable resistance to TCDD-induced toxicity and induction of target gene expression in comparison to TCDD-sensitive C57BL/6 mice (Moriguchi et al., 2003). Moreover, Silkworth et al. noted that human primary hepatocytes and HepG2 human hepatoma cells to be 10-1000 fold less sensitive toward TCDD and PCB126 141

6


Table 1: Benchmark response (BMR) concentrations in nM or µM (mono-ortho PCBs only) and efficacy (Ymax) in percentage relative to TCDD.

BMR

Compound

a

Chlorinated dibenzo-p-dioxin

20TCDD

TCDD

0.0056

1234678-HpCDD

0.20

12378-PeCDD

123678-HxCDD

0.012 0.05

Chlorinated dibenzofurans

Rat H4L1.1c4 BMR

Ymax

50TCDD

0.011

0.029

100

1.1

94

0.095

0.057

71

0.30

100 91

-c

234678-HxCDF

0.096

0.58

86

0.075

1234678-HpCDF

0.38

1234789-HpCDF

0.15

Non-ortho-substituted PCBs PCB77

40

PCB126 PCB169

Mono-ortho-substituted PCBs PCB74

0.076 2.6

4.2

PCB105

1.3

PCB167

1.5

PCB118

1.6

PCB156 PCB189

Di-ortho-substituted PCB PCB153

0.048

-b -

Ymax

100

0.057

0.97

123478-HxCDF

50TCDD

0.027

0.10

0.036

BMR

20TCDD

TCDF

23478-PeCDF

Mouse H1L6.1c2

BMR

0.21 0.40 1.9

0.83 290

0.43 11

15

81

90

91

0.43

2.8

68

88

94

100 86 79

0.025 0.018 0.066 6.0

0.24 21

0.09 0.05 1.2

1100 1 -

b

87 82 59 -c

76

46

-c

4.5

30

-c

74

0.28

2.3

62

-

-

0.06

82

0.03

12

-b

0.013

0.092

99

0.0096

-c

0.36

0.030

0.024

100

7.7 10

0.0091

-

c c

-b -

6.1 2.9 7.8 -b -

38 15 39 -b -

-c -c -c

-b

-

The names of compounds are in abbreviated form according to Materials and Methods. Induction too low to calculate BMR20TCDD and/or BMR50TCDD (see Materials and Methods). c Congener did not reach a clear top plateau, the top plateau was fixed at the Ymax of TCDD (100). a

142

b


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

Table 1: Benchmark response (BMR) concentrations in nM or µM (mono-ortho PCBs only) and efficacy (Ymax) in percentage relative to TCDD Guinea pig G16L1.1c8

BMR

20TCDD

0.0015 0.0018 0.014 0.019

0.0056 0.0012 0.0050 0.0059 0.020

0.014

BMR

Human AZ-AhR

BMR

Ymax

50TCDD

BMR

20TCDD

0.0049

100

0.19

0.052

100

0.59

0.0066 0.044 0.022

101 92 79

8.8

92

0.12 2.3

0.39

5.8 2.3 1.8 0.4

10.8

38.4

1.7

58

-b

1.1

7.0

-c

-b

-b

0.011 0.12 0.12 -

0.039

0.47

0.11 0.74

100

95

98 80

-

a

-b -

b

0.037

94

-b

-

-

-

0.59 1.1

-c -c

69

0.085

0.48

94

63

234

-c

1.2

-c

0.033

18

87

0.081

86

3.1

9.2

0.48

0.024

100

1,8

1.3

-b -b

-b -

b

-b -

b

-b -b -b

-

140 120 110 120 110

120

130

15

0.26 19

3.0

-b -

b

6.0

1.4 -

-

b b

-b

-b

-b

-b

-b

-b

-b

-b

-b

54 -

a

-b

-

b

-b -b

-

Ymax

0.60

0.93

-b

0.13

50TCDD

0.16

110

96

85

100

0.3

0.0050 0.022

0.72

0.070

94

0.012

BMR

20TCDD

2.8

0.11

Ymax

50TCDD

0.71

0.0081

HepG2-XRE-Luc

BMR

-b

-b -

b

-b -b -

-b

-b -

b

-b -b -

31 34 73 67 29 33 -b

14 -b -b -b -b -b

6

-b -b -

143


mediated AhR activation than rat cells (Silkworth et al., 2005). Besides ligand-binding affinity, ligand-specific recruitment of corepressors, coactivators and other cell factors and signaling pathways upon AhR activation and overall AhR response are highly cell type- and species-dependent (Carlson et al., 2009; Lonard and Oâ&#x20AC;&#x2122;Malley, 2007). Furthermore, differences in the luciferase responsive plasmid between human and rodent cell lines can also contribute to some of the observed variation.

Next, the REP values were calculated using BMR20TCDD and BMR50TCDD concentrations obtained in each CALUX cell line (Table 2). In Figure 2, the ratio between the calculated REPs (based on BMR20TCDD) and the WHO-TEFs are illustrated for the different species. A ratio of 1 indicates that a derived REP is comparable to its WHO-TEF. This graph shows that in general REPs for PCDDs and PCDFs in the rat, mouse, guinea pig and human AZAhR cell lines were similar or somewhat higher than the WHO-TEFs but still within the half order of magnitude of uncertainty around the WHO-TEF value (Figure 2a) (Van den Berg et al., 2006). Exceptions are 1234678-HpCDD and 1234789-HpCDF for which, for all species, REPs were calculated outside the uncertainty range and up to 50-fold higher compared to the WHO-TEFs. Also the AZ-AhR human REP for 123478-HxCDF was 17fold higher than its WHO-TEF value. In contrast to the PCDDs and PCDFs, REPs for the different PCBs were generally below the WHO-TEF and even outside the uncertainty range for rat and mouse cell lines (Figure 2b). The guinea pig cell line showed a wide variation around the WHO-TEFs for the different PCBs, with PCBs 77, 126, 105 and 156 having higher and PCBs 169, 118 and 167 having lower REPs than their respective WHO-TEFs. For the human AZ-AhR cell line only PCB 126 could be compared as this was the only active PCB, and this congener had a 30-fold less potent REP as compared to its WHO-TEF value. Although the BMR20TCDD for the 2 human cell lines were similar (Table 1), calculated REPs for the same congener deviated up to one order of magnitude from each other, with for most congeners having lower REPs in the human HepG2XRE-Luc cell line as compared to the AZ-AhR cell line (Figure 2c). REPs calculated for the human HepG2-XRE-Luc cell line were mostly outside the uncertainty range of the WHO-TEFs with 5-, 6-, 6- and 5-fold higher REPs for 1234678-HpCDD, 23478-PeCDF, 123478-HxCDF and 1234789-HpCDF, respectively and 6-, 10-, 10- and 4-fold lower REPs for 12378-PeCDD, 123678-HxCDD, 234678-HxCDF and 1234678-HpCDF, respectively (Figure 2c). These differences might be related to different plasmid constructs being used in individual CALUX assays.

144


t

t

v

OCDFnt

nt

123789-HxCDF

nt

123678-HxCDF

12378-PeCDF nt

1234789-HpCDFt

1234678-HpCDF

v

234678-HxCDF

123478-HxCDF

23478-PeCDFt

TCDFt

Chlorinated dibenzofurans

nt

OCDD

123789-HxCDDnt

nt

123478-HxCDD

1234678-HpCDD

v

123678-HxCDD

12378-PeCDDt

TCDDv

Chlorinated dibenzo-p-dioxins

Compound

0.04

0.01

0.1

0.1

0.2

0.1

0.03

0.1

0.5

1

REP

Rat

na

0.1

0.04

0.09

0.04

0.03 0.2

0.02

0.6

0.4

0.05

0.1

0.1

0.8 1.1

0.1

0.3

0.3

na

0.2

0.1

0.04

0.2

1

1.2

0.1

0.4

REP

REPQSAR

na

1.1

0.2

0.7

0.2

0.09

0.3

0.4

0.6

6.9

na

na

0.3

0.06

na

na

0.5

REPQSAR

Mouse

0.2

0.05

0.3

0.3

1.3

0.3

0.1

0.1

0.8

1

REP

na

0.8

0.2

0.6

0.2

0.1

0.3

0.4

0.5

2.3

na

na

0.2

0.06

na

REPQSAR

Guinea Pig

0.5

0.02

0.1

1.7

0.4

0.1

0.3

0.1

3

1

REP

Human

0.0003

0.1

0.1

0.03

0.01

0.01

0.1

0.1

0.3

0.1

0.0003

0.1

0.1

0.01

0.1

1

1

WHO-TEFa

Table 2: Relative effect potencies calculated based on measured data (REP) and predicted by developed QSAR models (REPQSAR), based on BMR20TCDD determined in rat H4L1.1c4, mouse H1L6.1c2, guinea pig G16L1. 1c8 and human AZ-AhR CALUX along with the WHO-TEFs.

REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

6

145


146

PCB169

t

PCB153t

Di-ortho substituted PCBs

nt

PCB157

PCB114nt

PCB189v

t

PCB167

t

PCB156

nt

PCB123

PCB118t

PCB105

PCB74v

Mono-ortho substituted PCBs

PCB81nt

v

0.000003

0.0001

0.00001

0.000003

0.000001

0.002

0.07

0.0001

REP

Rat

na

0.00001

0.00003

0.000002

0.000001

0.00001

0.000001

0.00001

0.000004

0.00003

0.0001

0.0001

0.0001

REPQSAR

0.000001

0.00004

0.000004

0.000002

0.000002

0.0005

0.05

0.002

REP

na

0.00001

0.00004

0.000001

0.00001

0.000001

0.00001

0.00001

0.00005

0.0005

0.0003

0.0007

REPQSAR

Mouse

0.00001 0.00001

0.0001

0.0001

0.00001

0.006

0.2

0.002

REP

na

0.00004

0.0002

0.00004

0.00006

0.00006

0.00018

0.002

0.001

0.002

REPQSAR

Guinea Pig

0.003

REP

Human

0.03

0.00003

0.00003

0.00003

0.00003

0.00003

0.00003

0.00003

0.00003

0.0003

0.1

0.0001

WHO-TEFa

Van den Berg et al., 2006. REPQSAR was not reported for human CALUX as the model was not valid. t Training set, v Validation set, nt Non-tested. na, not analyzed due to being outliers in validation set.

a

v

PCB126

PCB77t

Non-ortho substituted PCBs

Compound

Table 2: Continued.


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

         !          

3 12

Mouse  

1

78

-P

eC

1

DD

6 23

78

-H

xC

12

DD

6 34

78

-H

pC

DD

TC

DF

23

47

8-

Pe

F

CD

3 12

47

8-

Hx

CD

4 23

F

67

8-

Hx 12

CD

6 34

F 78

-H

pC

DF 1 ,7

-H

pC

Guinea Pig Human AZ-AhR

Human HepG2-XRE-Luc  

0 .1

DF

B

100

10

1

0 .1

0 .0 1

PC

B

77 PC

B

12

6 PC

B

16

9 PC

B

10

5 PC

B

11

8 PC

B

15

6 PC

B

16

7

C

100

R a tio R E P s / W H O -T E F s

!

Rat

10

0 .0 1

R a tio R E P s / W H O -T E F s

A

100

R a tio R E P s / W H O -T E F s

10

1

0 .1

0 .0 1

12

37

8-

Pe

CD

3 12

D

67

8-

H 1

xC

DD

4 23

67

8-

H

pC

DD

TC

DF

23

47

8-

Pe

CD

3 12

F

47

8-

H

xC

DF

4 23

67

8-

H 1

xC

DF

4 23

67

8-

H

pC

DF 1 ,7

-H

pC

DF

!

Figure 2. Ratios between REPs determined in the present study and their assigned WHO-TEFs ± half log uncertainty range. Ratios were determined for various PCDDs and PCDFs (a and c) or PCBs (b) in the rat, mouse, guinea pig, human AZ-AhR and human HepG2-XRE-Luc CALUX assays. Data represents the mean REP determined from 2 independent experiments. Gray shaded area represents the half log uncertainty range around the WHO-TEFs.

The variations in REPs between the different congeners tested in the rat, mouse, guinea pig and human CALUX assays were further investigated using PCA (Figure 3). Due to the better defined dose response curves of the DLCs for the human AZ-AhR cell line, further data analysis were performed using this human cell line only. The PCA model explains 90% of the variation by two principal components (PCs); 78% by the first PC and 12% by the second PC (Figure 3a). The first PC shows the variation related to the 147

6


4

A.

3 2

PC2

123478-HxCDF

1234789-HpCDF 1234678-HpCDD

1 0

12378-PeCDD

123678-HxCDD 1234678-HpCDF

-1

234678-HxCDF

TCDF

TCDD 23478-PeCDF

-2 PCB126

-3 -4

-6

-4

-2

0

2

4

6

PC1

0.8

B.

REPBMR50TCDD_H REPBMR20TCDD_H

0.6

LV2

0.4 0.2

REPBMR20TCDD_M REPBMR50TCDD_M REPBMR20TCDD_GP REPBMR20TCDD_GP REPBMR50TCDD_R REPBMR20TCDD_R

0.0 -0.2 -0.4

TEF 0.2

0.25

0.3

0.35

0.4

LV1

Figure 3. Principal component analysis calculated for the tested compounds based on the Y responses. The first two principal components (PC) are shown as (A) score plot of PC1 versus PC2 and (B) loading plot of loading vector (LV) 1 versus LV2, in which the variables are abbreviated by species (rat (R), mouse (M), guinea pig (GP) and human (H)) and response (REPBMR20TCDD and REPBMR50TCDD).

potency of congeners, taking into account all four CALUX assays studied. The congeners were positioned along the first principal component in order of their potency, such that the most potent ones (i.e. TCDD, 12378-PeCDD and 23478-PeCDF) were located on the right side of the score plot and the least potent ones (i.e. 1234678-HpCDF, 1234678-HpCDD and 1234789-HpCDF) on the left side. The second PC is related to the variation in REPs calculated in the human-CALUX assay and reflects the very low human REP value for PCB126 and relatively high REPs for 123478-HxCDF, 1234789-HpCDF, 148


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

and 1234678-HpCDD. The corresponding loading plot shows the variation in induction profiles of the four CALUX assays, where clearly the human cell system resulted in a distinct profile (Figure 3b). The slight separation between rat, guinea pig, and mouse cell line-derived assays in loading vector (LV) 2 is most likely attributed to the higher REPs for most PCDD/Fs in the mouse CALUX compared to the guinea pig and rat CALUX. The loading plot also shows that the REP value profiles calculated from rat, mouse and guinea pig CALUX assays are closer to the assigned TEF values than the REP values from the human CALUX assay.

The correlation between REP values from two different species-derived CALUX assays was also investigated (results are shown in Figure S5). Generally, REPs correlate well between the CALUX assays. Higher correlation coefficients are found when REPs are compared between rat and mouse (R = 0.97), guinea pig and rat (R = 0.98), and guinea pig and mouse CALUX assays (R = 0.97), than between human and any other species. Correlation coefficients with REPs from the human CALUX are 0.57, 0.69 and 0.52 for guinea pig, mouse and rat CALUX-derived REPs, respectively. Differences in REPs between CALUX assays may be due to biochemical, pharmacological and/or species-/cell-specific differences, such as variations in AhR ligand-binding affinity and specificity, differences in the binding of and regulation by cofactors (e.g. ARNT) or chaperone proteins, AhR DNA binding, and recruitment of transcriptional cofactors and/or differences in other factors that may modulate AhR functionality (Carlson et al., 2009; Petkov et al., 2010). Using in silico based methods, variations in ligand binding interactions with AhR were recently discussed where a biphenyl- and a dioxin-like mechanism were suggested (Petkov et al., 2010). In particular for PCBs, large species-variations in REPs and deviations from the WHO-TEF have been described (Van den Berg et al., 1998). For PCB126, REPs derived from studies using human primary cells or cell lines have been reported to be up to two orders of magnitude lower than the established WHO-TEF of 0.1 (Sutter et al., 2010; Van Duursen et al., 2005). The consistently lower human REPs for PCB126 and the absence of induction for the other PCBs tested in this study suggest that not only differences in the ligand-dependent affinity for the AhR, but also speciesand/or cell-specific differences in AhR functionality may play an important role in the observed variations in REPs (Carlson et al., 2009).

QSAR modeling REPs derived from rat, guinea pig, mouse and human AZ-AhR CALUX assays were used to calculated rat, guinea pig, mouse and human QSAR models, respectively. Internal and external predictivity of the QSAR models was higher when using log REPBMR20TCDD values for each CALUX assay than those based on log REPBMR50TCDD (results not shown). Therefore QSAR models were based on log REPBMR20TCDD and these showed Q2 and RMSECV values 149

6


from 0.81 to 0.92 and from 0.59 to 1.02, respectively (Table 3). Q2 values larger than 0.5 obtained after internal validation of each model generally mean that the developed models are predictive. In addition, the differences between obtained R2 and Q2 do not exceed 0.3 indicating that there was no over-fitting in the model development (Golbraikh and Tropsha, 2002). The plots of the predicted values versus the experimentally measured values for all CALUX assays are shown in Figure 4. Generally, there is a good agreement between the experimental and predicted REP values. While a better agreement was seen for PCDD/Fs for the rat (Figure 4a), mouse (Figure 4b), and guinea pig (Figure 4c) models, a lower prediction error was achieved for PCBs in the human model (Figure 4d) as compared with the other species. As described above, besides PCB126 none of the PCBs tested induced AhR-activity in the human CALUX assay. To be able to model and understand the characteristics of non-potent compounds, the nonresponding PCBs were assigned a REP value of one or two orders of magnitude lower than the lowest REP calculated in each assay (Table 2). However, since the human model was based on only one active PCB, it was not possible to capture enough information on PCBs chemical variation to allow prediction of non-tested compounds. On the other hand, attempts to develop a model using only the responding compounds in the human assay were not successful. Thus, the human QSAR model was not used for predicting effects of untested PCBs but will be discussed below in terms of descriptor dependence and outlying chemicals. The residuals obtained in the QSAR models for all compounds were plotted against the experimental values (Figure S6). The residuals of the training set compounds were randomly distributed indicating that no systematic error exists in the models. However, the residuals of some validation set compounds were large indicating that they may be outliers, as described in detail below. Interpretation of descriptors The most significant descriptors were studied using VIPs and correlation plots in order to gain insights in structure-specific related differences in mechanism of action (Table 3 and Figure S8). In most models the differences in LUMO and HOMO energy (GAP), total positive van der Waals surface area (PEOE_VSA_POS), Balabanâ&#x20AC;&#x2122;s connectivity topological index (Balabanâ&#x20AC;&#x2122;s index), selected UV descriptors and shape indices (e.g. Kier3) were the most significant descriptors. The potency of the compounds decreased with the increase of their GAP values. It has previously been demonstrated that the GAP descriptor is of importance in modeling the properties and activities of halogenated organic compounds and to be negatively correlated with REPs (Arulmozhiraja and Morita, 2004; Diao et al., 2010; Kobayashi et al., 1991; Niu et al., 2005). It was also concluded that compounds having PEOE_VSA_POS greater than 50 are more potent (Figure S8 of Supporting Information). Therefore, this important

150


0.96

0.94

0.98

Mouse

Guinea pig

Human

0.90

0.84

0.92

0.84

R2totb

0.38

0.64

0.58

0.61

RMSEEc

0.85 0.81 0.92

1.43 0.98

0.82

Q2e

0.97

1.32

RMSEPd

0.59

0.97

1.02

0.92

RMSEcvf

PEOE_PC+/PEOE_PC-/240/PEOE_VSA_POS/Esusc/ Gap_AM1/ Gap_B3LYP/Kier3

Gap_B3LYP/Gap_AM1/Balaban’s index / Kier3/215/ PEOE_VSA_POS/ 240/210/ PEOE_PC+/PEOE_PC-

Gap_B3LYP/Gap_AM1/ Balaban’s index/kier3/215/ PEOE_VSA_POS/240

Gap_B3LYP/Gap_AM1/240/ Kier3/Balaban’s index/ 215/PEOE_VSA_POS/PEOE_PC+/PEOE_PC-

Significant descriptorsg

d

a

R2tr: determination coefficient for the training set. b R2tot: determination coefficient for the whole data set. c RMSEE: root mean square error of the estimation. RMSEP: root mean square error of the prediction. e Q2: cross-validated R2. f RMSEcv: root mean square error for cross validation. g The most significant descriptors in order of VIP value. The meaning of each descriptor is given in Table S1.

0.94

Rat

R2tra

Table 3: Statistics of calculated OPLS models for each species.

REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

6

151


descriptor could be used to classify our studied compounds into low and high potent compounds. This descriptor indicates the surface characteristics of the congeners and may be related to their interaction at the receptor pocket. Balabanâ&#x20AC;&#x2122;s index has a negative effect on the model meaning that increasing the descriptor value decreases the predicted REP value. This descriptor value increases by increasing the number of chlorine in each chemical group. The compounds with Kier3 value above three show lower REP values. Both positive and negative correlations were found for selected UV absorbance wavelengths and REPs. This is due to spectral trend differences between the chemical classes. The main band for PCBs is found at 200-215 nm indicating that a large absorption at low wavelengths has a negative correlation with the REP whereas intensive absorption at 230-240 nm (second peak of PCDD/Fs) is positively correlated with the REP. Since the measured PCDD/Fs all have lower absorption than the majority of the PCBs at approximately 215 nm, low absorption at this wavelength is associated with high REP values. In the same manner, most measured PCBs have a local minimum around 230-240 nm while the PCDD/Fs have high absorption, therefore high absorption at 235 nm is associated with high REP values (Larsson et al., 2013).

Although GAP and PEOE_VSA_POS were among the important descriptors in human model, total positive and negative partial charges (PEOE_PC+ and PEOE_PC-) were determined as the most influential descriptors. The other significant descriptors in human QSAR model were absorption at the wavelength of 240 nm, PEOE_VSA_POS, Kier3 and the highest electrophilic susceptibility on a lateral carbon (Esusc). In conclusion, the existence of similar descriptors in the list of the most significant descriptors in rat, mouse and guinea pig models can be an indicator of similar interaction at the target for the species. Although the human model is not applicable to accurately predict the response values, the introduction of different descriptors as the most critical ones (e.g. PEOE_PC+, PEOE_PC-) into the model may indicate that the receptor interaction differs slightly in the human CALUX assay as compared to the other CALUX assays. Applicability Domain Analysis of the applicability domain was performed to evaluate whether the QSAR models produce reliable predicted data for the response of new compounds. In general, the developed QSAR models could be applied for rat, mouse and guinea pig cells to predict the response value of PCDD/Fs and PCBs containing at least four chlorine atoms. PCBs can be mono-ortho or non-ortho substituted while PCDDs and PCDFs must be substituted at the 2, 3, 7, and 8 positions with at most three chlorine atoms substituted at other positions. Moderate outliers in the models were searched using 152


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

the model membership probabilities with a set confidence level of 0.95 (Table S3). Among the tested congeners TCDF, 1234678-HpCDD and PCB169 and among the nontested ones, Octachlorodibenzo-p-dioxin (OCDD) and Octachlorodibenzofuran (OCDF) were considered to be moderate outliers in all models. Even though PCB126 was not an outlier based on the membership probability test, the congener showed large distance to the models in Y space. This means that the chemistry of the compound falls within the applicability domain but its response is outlying. The residuals obtained for PCB126 were more than two times higher than the standard deviation for residuals in the guinea pig and human models, and three times higher than the standard deviation for residuals in the rat and mouse models. This deviation can most likely be explained by the fact that REPs for PCB126 are more similar to those of PCDD/Fs compared to other PCBs, whereas its structure and chemistry is more similar to PCBs. QSAR models predict the potency of a congener based on chemical properties and their effect in an assay. As a consequence, developed QSAR models in this study tend to predict a considerably lower REP for PCB126, as has been determined based on the CALUX assays (Table 2, Figure 4).

b)  

1

A.

0

1

PCB126

-5

-6

PCB156

0

Exp. log REP

-1 -2 -3 -4 -5 -6 -7

-1

PCB77

-7

-5

-4

-3

-2

-1

0

PCB156

PCB118

PCB153

-7

1

-4

-3

-2

12378-PeCDD 234678-HxCDF

TCDF

0

1

123478-HxCDF

0

123478-HxCDF 1234789-HpCDF 1234678-HpCDD 1234678-HpCDF

TCDD

123478-HxCDF 1234678-HpCDD 23478-PeCDF 123678-HxCDD TCDF 234678-HxCDF

-1

PCB105

-1

12378-PeCDD

PCB156 PCB189

-5

Human

PCB77

PCB189

-6

D.

23478-PeCDF PCB126

PCB74

PCB105

Pred. log REP

Guinea Pig

1234678-HpCDF

-2

PCB126

-3

PCB118

PCB167 PCB74

 

-7

1

Pred. log REP

-7

-4

PCB167

C.

-3

PCB189

-6

1234678-HpCDF

PCB77

-6

PCB126

23478-PeCDF TCDF 123478-HxCDF

124789-HpCDF

-2

-5

PCB118

1234678-HpCDD

PCB74

PCB153

1

PCB105

PCB167

123678-HxCDD 123478-HxCDF 234678-HxCDF TCDF 1234789-HpCDF 1234678-HpCDD 1234678-HpCDF

Exp. log REP

-4

-7

23478-PeCDF

-3

12378-PeCDD 234678-HxCDF

0

TCDD

Exp. log REP

Exp. log REP

-1 -2

B.

Mouse 12378-PeCDD

-4

PCB153

-6

-5

-4

-3

Pred. log REP

-2

-1

0

-5

1

PCB 153 PCB167 PCB74 PCB18 PCB118 PCB105 PCB169 PCB156

-5

PCB77

-4

-3

-2

Pred. log REP

-1

0

1

Figure 4. The plots of QSAR-predicted (pred) log relative effect potency (REP) values against experimental (exp) log REP values for (A) rat H4L1.1c4, (B) mouse H1L6.1c2, (C) guinea pig G16L1.1c8, and (D) human AZ-AhR cells. The blue triangles refer to the training set compounds and red circles to the validation set compounds.

153

6


Similar findings have been observed by Larson et al (Larsson et al., 2013). It is worthy to note that QSAR models where PCB126 has been excluded shows lower prediction errors.

Prediction of REPs and comparison with TEFs TEFs are based on consensus decisions using mainly in vivo rodent studies but also in vitro information. Here, we compared the REPs calculated by our three QSAR models with the assigned TEF values. Correlation plots of predicted log REP versus log TEF values indicate a good agreement for most congeners and species; rat (R2 = 0.92), mouse (R2 = 0.91) and guinea pig (R2 = 0.76) (as shown in Figure S7 of the Supporting Information). Almost all QSAR-predicted REPs for PCDDs based on rat, mouse and guinea pig CALUX data were within one order of magnitude of the WHO assigned TEF values (Table 2). For the non-tested PCDDs, REPs predicted by QSAR models were close to given TEF values. Unfortunately, no value could be determined for OCDD as it was outside the applicability domain of the model. The QSAR-predicted REP values for PCDFs differed up to two orders of magnitude as compared to their TEFs. For instance, the predicted REP value for the non-tested 12378-PeCDF were 0.09, 0.7 and 0.6, in rat, mouse and guinea pig respectively, which are three to twenty times higher than expected based on the assigned TEF value of 0.03. However, it should be noted that this congener is in vivo very fast metabolized and only very low concentrations are found in blood (Brewster and Birnbaum, 1988). In vitro and in silico models might give an overestimation of the potency due to an overload of the in vitro system with high concentrations of the congener and the fact that toxicokinetics are not taken into account in these models. Among the non-tested PCDFs; 123789-HxCDF (TEF 0.1) was predicted to be the most potent congener with predicted REPs of 0.1, 1.1 and 0.8 in the rat, mouse and guinea pig, respectively. On the other hand, predicted REPs for 123678-HxCDF were close to the assigned TEF value of 0.1 (See Table 3). The difference between the measured and the predicted values was in general higher for PCBs than for PCDD/Fs. Predicted REP values of mono-ortho PCBs were lower than 0.0001 except PCB74 (0.0002) and PCB114 (0.0002) in the guinea pig model. Among the non-tested mono-ortho PCBs (PCBs 114, 123 and 157), PCB114 was predicted to be the most potent in each tested species with highest predicted REP in the guinea pig system. The developed QSAR models predicted REPs for PCB114 from 0.00003 for rat to 0.0002 for guinea pig. In previous studies, REP values of 0.0001 and 0.0002 have been calculated for PCB114 in monkey and pig systems, respectively (Van Der Burght et al., 1999; 2000). The REP values calculated for non-tested non-ortho PCB81 range from 0.0001 for rat cells to 0.002 for guinea pig cells. The REP value of this compound was calculated to be more than 0.01 for the induction of CYP1A activity in male castrated pig and in cynomolgus monkey hepatocytes (Dynes et al., 1985; Mekenyan et al., 1996). It was further assigned a REP value of 0.13 in fish 154


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

hepatoma cells and 0.00004 in rat hepatoma H4IIE cells (Hahn and Chandran, 1996; Sawyer and Safe, 1982). A large range of REP values has thus been reported of which those given in present study are in the range of the WHO-TEF value of 0.0003. Conclusion Taken these data together, it is evident that DLC-mediated responses in the human CALUX cell line differ significantly from responses in other rodent species derived CALUX cell lines. Not only is the human cell line less sensitive as indicated by the 20% effect concentrations of TCDD that were 1.5, 5.6, 11 and 190 pM for guinea pig, rat, mouse and human cells, respectively. Also apparent congener-specific species differences in potency were observed between human and rodent CALUX cell lines which was most clearly reflected by a lower human REP for PCB 126 (0.003) compared to guinea pig (0.2), rat (0.07) and mouse (0.05). Furthermore, human derived REPs for 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF were clearly higher compared to rodent derived REPs. PCA indicated that REPs derived from rat, mouse, and guinea pig revealed an induction pattern similar to each other and to the TEFs compared to human REP values. QSAR models identified differences in LUMO and HOMO energy, partial atomic charges, electrophilic susceptibility, Balabanâ&#x20AC;&#x2122;s index and surface area characteristics as the most important descriptors influencing the models. The most influencing chemical descriptors in the human model were different from the rodent models. This indicates differences in ligand-receptor interaction between the human and rodent CALUX assays. The predicted REP values for 11 non-tested compounds indicated that 123789-HxCDF was the most potent non-tested congener in each assay except in the rat assay in which 123789-HxCDD was predicted as the most potent congener. In vitro and in silico derived data from present study for all WHO-TEF assigned congeners could be used as a basis for a better understanding of species variations and development of risk assessment tools of DLCs.

155

6


Table S1: Short description of the quantum mechanical descriptors (QM), 2D descriptors (2D), ultraviolet absorption data (UV) and log Kow Variable number

Descriptor QM

Description

1

HOMOa1

Highest HOMO density on a lateral carbon

3

HOMOa3

Third highest HOMO density on a lateral carbon

2 4 5 6 7 8 9

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

HOMOa2 HOMOa4

Fourth highest HOMO density on a lateral carbon

LUMOa1

Highest LUMO density on a lateral carbon

LUMOa3

Third highest LUMO density on a lateral carbon

LUMOa2 LUMOa4 ESusc

NSusc

Second highest LUMO density on a lateral carbon Fourth highest LUMO density on a lateral carbon

Highest electrophilic susceptibility on a lateral carbon Highest nucleophilic susceptibility on a lateral carbon

RSusc

Highest radical susceptibility on a lateral carbon

QC_a2_ESP

Second highest partial charge on a lateral carbon (eV) derived from electrostatic potential

QC_a1_ESP QC_a3_ESP QC_a4_ESP Ha2/Ha1 Ha3/Ha1 Ha4/Ha1

Highest partial charge on a lateral carbon (eV) derived from electrostatic potential

Third highest partial charge on a lateral carbon (eV) derived from electrostatic potential

Fourth highest partial charge on a lateral carbon (eV) derived from electrostatic potential Ratio of the second highest and highest HOMO density Ratio of the third highest and highest HOMO density

Ratio of the fourth highest and highest HOMO density

La2/La1

Ratio of the second highest and highest LUMO density

La4/La1

Ratio of the fourth highest and highest LUMO density

La3/La1 HOMO_AM1

Ratio of the third highest and highest LUMO density

Energy of the HOMO (eV), based on AM1 (semi empirical method)

LUMO_AM1

Energy of the LUMO (eV), based on AM1 (semi empirical method)

GAP-1_AM1

Difference in energy of the LUMO and second highest occupied molecular orbital (HOMO-1), (LUMO-HOMO-1), based on AM1 (semi empirical method)

GAP_AM1

26

ChemPot_AM1

27

Dm_AM1

156

Second highest HOMO density on a lateral carbon

Difference in LUMO and HOMO energy (LUMO-HOMO, eV), based on AM1 (semi empirical method) Chemical potential, i.e. the average of the HOMO and LUMO energy (eV), based on AM1 (semi empirical method) Dipole moment, based on AM1 (semi empirical method)


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines - Supplemental Material

Table S1: Continued Variable number

Descriptor

Description

29

LUMO_ B3LYP

Energy of the LUMO (eV) calculated using DFT with the basis set B3LYP/6-31G**

31

GAP-1_ B3LYP

32

ChemPot_ B3LYP

Difference in energy of the LUMO and the second highest occupied molecular orbital (HOMO-1), (LUMO-HOMO-1), calculated using DFT with the basis set B3LYP/6-31G**

28

30

33

HOMO_B3LYP

GAP_ B3LYP

Dm_ B3LYP

Energy of the HOMO (eV) calculated using DFT with the basis set B3LYP/6-31G** Difference in LUMO and HOMO energy (LUMO-HOMO, eV) calculated using DFT with the basis set B3LYP/6-31G**

Chemical potential, i.e. the average of the HOMO and LUMO energy (eV) calculated using DFT with the basis set B3LYP/6-31G**

Dipole moment calculated using DFT with the basis set B3LYP/6-31G**

2D 34

VDistEq

35

VDistMa

36

Weight

38

chi1

39

VAdjEq

40

VAdjMa

41

balabanJ

43

PEOE_PC-

37

42

chi0

PEOE_PC+

If m is the sum of the distance matrix entries then VdistEq is defined to be the sum of log2 m - pi log2 pi / m where pi is the number of distance matrix entries equal to i. If m is the sum of the distance matrix entries then VDistMa is defined to be the sum of log2 m - Dij log2 Dij / m over all i and j. Molecular weight (including implicit hydrogens) with atomic weights taken from [CRC 1994].

Atomic connectivity index (order 0) from [Hall 1991] and [Hall 1977]. This is calculated as the sum of 1/sqrt(di) over all heavy atoms i with di > 0.

Atomic connectivity index (order 1) from [Hall 1991] and [Hall 1977]. This is calculated as the sum of 1/sqrt(didj) over all bonds between heavy atoms i and j where i < j.

Vertex adjacency information (equality): -(1-f)log2(1-f) - f log2 f where f = (n2 - m) / n2, n is the number of heavy atoms and m is the number of heavy-heavy bonds. If f is not in the open interval (0,1), then 0 is returned.

Vertex adjacency information (magnitude): 1 + log2 m where m is the number of heavy-heavy bonds. If m is zero, then zero is returned. Balaban’s connectivity topological index [Balaban 1982].

Total positive partial charge: the sum of the positive qi. Q_PC+ is identical to PC+ which has been retained for compatibility.

Total negative partial charge: the sum of the negative qi. Q_PC- is identical to PC- which has been retained for compatibility.

157

6


Table S1: Continued Variable number

Descriptor

45

PEOE_RPC-

46

PEOE_VSA_ FNEG

47

PEOE_VSA_FPOS

48

PEOE_VSA_HYD

49

PEOE_VSA_NEG

50

PEOE_VSA_POS

51

Kier1

53

Kier3

Third kappa shape index: (n-1) (n-3)2 / p32 for odd n, and (n-3) (n2)2 / p32 for even n [Hall 1991].

logS

Log of the aqueous solubility This property is calculated from an atom contribution linear atom type model [Hou 2004] with r2 = 0.90, ~1,200 molecules.

44

52

PEOE_RPC+

Kier2

54

KierFlex

56

apol

55

57

bpol

58

mr

59

vsa_hyd

158

Description

Relative positive partial charge: the largest positive qi divided by the sum of the positive qi. Q_RPC+ is identical to RPC+ which has been retained for compatibility.

Relative negative partial charge: the smallest negative qi divided by the sum of the negative qi. Q_RPC- is identical to RPC- which has been retained for compatibility.

Fractional negative van der Waals surface area. This is the sum of the vi such that qi is negative divided by the total surface area. The vi are calculated using a connection table approximation.

Fractional positive van der Waals surface area. This is the sum of the vi such that qi is non-negative divided by the total surface area. The vi are calculated using a connection table approximation.

Total hydrophobic van der Waals surface area. This is the sum of the vi such that |qi| is less than or equal to 0.2. The vi are calculated using a connection table approximation.

Total negative van der Waals surface area. This is the sum of the vi such that qi is negative. The vi are calculated using a connection table approximation.

Total positive van der Waals surface area. This is the sum of the vi such that qi is non-negative. The vi are calculated using a connection table approximation. First kappa shape index: (n-1)2 / m2 [Hall 1991].

Second kappa shape index: (n-1)2 / m2 [Hall 1991].

Kier molecular flexibility index: (KierA1) (KierA2) / n [Hall 1991].

Sum of the atomic polarizabilities (including implicit hydrogens) with polarizabilities taken from [CRC 1994].

Sum of the absolute value of the difference between atomic polarizabilities of all bonded atoms in the molecule (including implicit hydrogens) with polarizabilities taken from [CRC 1994].

Molecular refractivity (including implicit hydrogens). This property is calculated from an 11 descriptor linear model [MREF 1998] with r2 = 0.997, RMSE = 0.168 on 1,947 small molecules.

Approximation to the sum of VDW surface areas of hydrophobic atoms.


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines - Supplemental Material

Table S1: Continued Variable number

Descriptor

61

Density

63

vdw_vol

60

62

64

SMR

vdw_area

weinerPath

65

weinerPol

66

Zagreb

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

UV

Description

Molecular refractivity (including implicit hydrogens). This property is an atomic contribution model [Crippen 1999] that assumes the correct protonation state (washed structures). The model was trained on ~7000 structures and results may vary from the mr descriptor. Molecular mass density: Weight divided by vdw_vol.

Area of van der Waals surface calculated using a connection table approximation.

van der Waals volume calculated using a connection table approximation.

Wiener path number: half the sum of all the distance matrix entries as defined in [Balaban 1979] and [Wiener 1947].

Wiener polarity number: half the sum of all the distance matrix entries with a value of 3 as defined in [Balaban 1979]. Zagreb index: the sum of di2 over all heavy atoms i.

200

UV absorption at 200 nm.

210

UV absorption at 210 nm.

205 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290

UV absorption at 205 nm. UV absorption at 215 nm. UV absorption at 220 nm. UV absorption at 225 nm. UV absorption at 230 nm.

6

UV absorption at 235 nm. UV absorption at 240 nm. UV absorption at 245 nm. UV absorption at 250 nm. UV absorption at 255 nm. UV absorption at 260 nm. UV absorption at 265 nm. UV absorption at 270 nm. UV absorption at 275 nm. UV absorption at 280 nm. UV absorption at 285 nm. UV absorption at 290 nm.

159


Table S1: Continued Variable number

Descriptor

87

300

86 88 89 90 91 92 93 94 95 96 97 98

Description

295

UV absorption at 295 nm.

305

UV absorption at 305 nm.

310 315 320 325 330 335 340 345 350

Log Kow

UV absorption at 300 nm. UV absorption at 310 nm. UV absorption at 315 nm. UV absorption at 320 nm. UV absorption at 325 nm. UV absorption at 330 nm. UV absorption at 335 nm. UV absorption at 340 nm. UV absorption at 345 nm. UV absorption at 350 nm.

KOWWIN logKow from the software EpiSuite (available at www. epa.gov)

[Balaban 1979] Balaban, A.T.; Five New Topological Indices for the Branching of Tree-Like Graphs; Theoretica Chimica Acta 53 (1979) 355-375. [Balaban 1982] Balaban, A.T.; Highly Discriminating Distance-Based Topological Index; Chemical Physics Letters 89 No. 5 (1982) 399-404. [CRC 1994] CRC Handbook of Chemistry and Physics. CRC Press (1994). [Crippen 1999] Wildman, S.A., Crippen, G.M.; Prediction of Physiochemical Parameters by Atmoic Contributions; J. Chem. Inf. Comput. Sci. 39 No. 5 (1999) 868-873. [Hall 1991] Hall, L.H., Kier, L.B.; The Molecular Connectivity Chi Indices and Kappa Shape Indices in Structure-Property Modeling; Reviews of Computational Chemistry 2 (1991). [Hall 1977] Hall, L.H., Kier, L.B.; The Nature of Structure-Activity Relationships and Their Relation to Molecular Connectivity; Eur J. Med. Chem. 12 (1977) 307. [Hou 2004] Hou, T.J., Xia K., Zhang, W., Xu, X.J.; ADME Evaluation in Drug Discovery. 4. Prediction of Aqueous Solubility Based on Atom Contribution Approach; J. Chem. Inf. Comput. Sci. 44 (2004) 266-275. [MREF 1998] Labute, P.; MOE Molar Reflectivity Model unpublished. Source code in $MOE/lib/quasar.svl/q_ mref.svl (1998). [MOE] MOE 2006.08. Chemical Computing Group, Quebec, Canada, 2008; software available at http://www.chemcomp.com/ [Wiener 1947] Wiener, H.; Structural Determination of Paraffin Boiling Points; Journal of the Americal Chemical Society 69 (1947) 17-20.

160


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines - Supplemental Material Table S2: The membership probability values of the developed QSAR models for the TEF assigned compounds plus PCBs 74 and 153. Compound TCDD

v

1,2,3,7,8-pentachlorodibenzo-p-dioxin

t

rat

1,2,3,6,7,8-hexachlorodibenzo-p-dioxin

v

1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin t 123478-hexachlorodibenzo-p-dioxin 123789-hexachlorodibenzo-p-dioxin Octachlorodibenzo-p-dioxin

nt

TCDFt

2,3,4,7,8-pentachlorodibenzofurant 2,3,4,7,8-pentachlorodibenzofuran

t

2,3,4,6,7,8-hexachlorodibenzofuran

nt nt

1,2,3,4,7,8,9-heptachlorodibenzofuran 12378- pentachlorodibenzofuran 123678-hexachlorodibenzofuran

nt

nt

123789-hexachlorodibenzofurannt Octachlorodibenzofuran

nt

PCB77

t

PCB81

nt

PCB126v PCB169

v

PCB105

t

PCB114

nt

PCB118t PCB123

nt

PCB156

t

PCB157

nt

PCB167t PCB189 PCB74

v

v

PCB153

t

0.026042

0.004031

0.002210

0.646086 0.091838 0.503995 0.077976 0.917393

t

0.029821

0.041584

0.009696

1,2,3,4,6,7,8-heptachlorodibenzofuranv

guinea pig

0.161787

0.000073

v

mouse

0.992565 0.281916 0.983753 0.997211 0.881335

0.259352 0.279202 0.027359 0.000058 0.028964 0.951297 0.987736 0.273594 0.982288 0.99675

0.909834

0.019442 0.174605 0.002334 0.250513 0.021371 9.12E-05

0.020193 0.964374 0.979416 0.278265 0.980149 0.995418 0.928962

0.98857

0.985039

0.984584

0.0546626

0.094955

0.164796

0.696207 0.033193 0.050186 0.194472 0.001645 0.968773 0.984702 0.999920 0.620235 0.952758 0.974562 0.904536 0.946648 0.699053 0.571505

0.682518 0.032522 0.057165 0.178194 0.001144 0.941324 0.975163 0.999729 0.571375 0.962569 0.966126 0.921813 0.984874 0.678889 0.659423

0.689763 0.035137 0.082578 0.208924 0.001328 0.945539 0.975087 0.999697 0.573333 0.929019 0.957456 0.958460 0.991960 0.710041 0.741619

Those marked in grey have a probablity value below 0.05. They are regarded as moderate outliers. We didnâ&#x20AC;&#x2122;t predict REP for the validation and non-tested congeners with a membership probability below 0.05. t , v and nt indicate the training (t), validation (v) and non-tested (nt) sets, respectively.

161

6


Figure S1. Molecular structures and substitution pattern of studied compounds.

162


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines - Supplemental Material

Figure S2. The UV spectra of PCBs 74 and 153 in the range 200-350 nm. The spectra were normalized to the maximum absorption of each spectrum, to elucidate the peak patterns of each compound. 15

non-­‐tested compounds   training  compounds  

OCDF

10 1234678-HpCDD 123678-HxCDD 123789-HxCDD 123478-HxCDD

5

12378-PeCDD

PCB167

TCDD

PC2

PCB153

PCB189

PCB157 PCB123 PCB118 PCB105 PCB114

PCB156

0

OCDD 1234678-HpCDF PCB169

1234789-HpCDF

12378-PeCDF 234678-HxCDF 123789-HxCDF 23478-PeCDF 12378-PeCDF

123478-HxCDF

-­‐5

PCB74

PCB126

PCB77

PCB81

TCDF

-­‐10

-­‐15 -­‐20  

-­‐15

-­‐10

-­‐5

0 PC1  

5

10

15

20

Figure S3. The chemical diversity of studied compounds illustrated using a PCA score plot based on 98 chemical descriptors. The first 2 principal components explain 36 and 24% of the variation, respectively.

163

6


A.

D. PCDF

PCB

1000 human / rat ratio

100 10

B. PCDF

PCB

100 10 1

TCDD 12378-PeCDD 123678-HxCDD 1234678-HpCDD TCDF 23478-PeCDF 123478-HxCDF 234678-HxCDF 1234678-HpCDF 1234789-HpCDF PCB126 PCB77 PCB156 PCB169 PCB118 PCB74 PCB105 PCB167

mouse / guinea pig ratio

PCDD 1000

C. PCDD  

PCDF

PCB

E.                                 F.  

human / rat ratio

TCDD 12378-PeCDD 123678-HxCDD 1234678-HpCDD TCDF 23478-PeCDF 123478-HxCDF 234678-HxCDF 1234678-HpCDF 1234789-HpCDF PCB126 PCB77 PCB156 PCB169 PCB118 PCB74 PCB105 PCB167

rat / mouse ratio

0.1

PCB

100 10

 

10.0

1.0

PCDF

1

human / mouse ratio

1

PCDD

1000

TCDD 12378-PeCDD 123678-HxCDD 1234678-HpCDD TCDF 23478-PeCDF 123478-HxCDF 234678-HxCDF 1234678-HpCDF 1234789-HpCDF PCB126 PCB77 PCB156 PCB169 PCB118 PCB74 PCB105 PCB167

rat / guinea pig ratio

PCDD

PCDD

PCDF

PCB

PCDD

PCDF

PCB

1000 100 10 1

10000 1000 100 10 1

Figure S4. The variation in sensitivity between the four species, (A) BMR20TCDD ratios between rat and guinea pig, (B) BMR20TCDD ratios between mouse and guinea pig, (C) BMR20TCDD ratios between rat and mouse, (D) BMR20TCDD ratios between human and rat, (E) BMR20TCDD ratios between human and mouse, (F) BMR20TCDD ratios between human and guinea pig.

164


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines - Supplemental Material

R2 = 0.69

R2 = 0.97 -6.00

-5.00

-4.00

-3.00

-2.00

-1.00

-3.00

0.00 0.00

-2.00

-1.00

0.00

1.00 1.00 0.00

mouse

human

-2.00

-1.00

-4.00

-6.00

rat

-5.00

-4.00

-3.00

-2.00

-1.00

-3.00

mouse

R2 = 0.98 -6.00

-2.00

0.00 0.00

R2 = 0.52 -3.000

-2.500

-2.000

-1.500

-1.000

-0.500

-1.00

0.00 human

guinea pig

-2.00 -3.00 -4.00

-1.00 -2.00

-5.00 -6.00

rat

0.000 1.00

-3.00

rat

R2 = 0.97

R2 = 0.57 -2.00

-1.50

-1.00

-0.50

0.00

0.50 1.00

-6.00

-5.00

-4.00

-3.00

-2.00

guinea pig

1.00 1.00

-1.00 guinea pig

human

-0.50

0.00

0.00

0.50 0.00

-1.00

-2.00 -3.00

-1.00

-4.00

-1.50

-5.00

-2.00

mouse

-6.00

Figure S5. The relationship between the log REP values for each two species.

165

6


A. Rat  

B. Mouse  

3.0 2.0  

3.0

Training

Validation

2.0

1.0

PCB126

1.0 Residual

Residual

Training

PCB126

Validation

0.0 -­‐1.0  

0.0 -1.0

PCB189

-­‐2.0 -­‐3.0   -­‐8.0  

-­‐7.0

-­‐6.0

PCB74

-2.0

-­‐5.0

-­‐4.0

-­‐3.0

-­‐2.0

-­‐1.0

0.0

-3.0

1.0

-7.0

-6.0

-5.0

-4.0

Exp. log REP

C.  Guinea  Pig   3

3

PCB126

-1.0

0.0

1.0

Training Validation

2

PCB126

1 Residual

1 Residual

-2.0

D. Human  

Training Validation

2

-3.0

Exp. log REP

0

0

-1

-1

-2

-2

1234678-HpCDF 123678-HxCDD

PCB74

-3

-7

-6

-5

-4

-3 Exp. log REP

-2

-1

0

1

-3

-7

-6

-5

-4

-3

-2

-1

0

1

Exp. log REP

Figure S6. The plot of residuals versus experimental log REP values in (A) rat, (B) mouse, (C) guinea pig and (D) human models. The names of the congeners with the largest residuals are shown.

166


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines - Supplemental Material A. Rat  

B. Mouse   1

R² = 0.91696 -5

-4

-3

-2

0

-1

1

R² = 0.91405 -5

0

-4

-3

-2

-1

Pred. log REP

Pred. log REP

-1 -2 -3

-3 -4 -5

-5

A. Guinea  Pig    

R² = 0.75757 -5

-4

-6

-6

1

-3

-2

-1

0

Pred. log REP

-1 -2 -3

B. Human   1 R² = 0.57029

0

-5

-4

-3

-2

0

0

-2 -3 -4

-5 -6

-1

-1

-4

log TEF

-7

log TEF

Pred. log REP

0

-2

-4

log TEF

0 -1

log TEF

-5

Figure S7. The correlation plot of the predicted values against log TEF values in (A) rat, (B) mouse, (C) guinea pig, and (D) human.

6 167


0

rat

-1

0

log REP EC20

log REP EC20

-2 -3 -4 -5

-2 -3 -4

-6

-5

-7

-6

-8

7.8

8

8.2

8.4

8.6

8.8

9

-7

9.2

Gap_AM1

rat

-1

0

10

20

30

40

50

60

70

PEOE_VSA_POS

0

rat

-1

log REP EC20

-2 -3 -4 -5 -6 -7 -8

4

4.2

4.4

4.6

4.8

5

5.2

5.4

5.6

5.8

Gap_B3LYP

Figure S8. Selected correlation plots between chemical descriptors and logREPBMR20TCDD.

168

80

90


REPs of DLCs in human, rat, mouse and guinea pig CALUX cell lines

6 169


Part

IV

Discussion, Conclusion & Annex

Wat de rups het einde noemt,

noemt de rest van de wereld een vlinder.

Lao-Tse


Chapter

7

Summary, General Discussion, Conclusions and Recommendations


174


Summary, Discussion and Conclusions

Summary

T

he toxic equivalency factor (TEF) approach is the most commonly used method for assessing the risk of complex mixtures of dioxin-like compounds (DLCs). Consequently, the actual value of a TEF is crucial for accurate risk assessment. When determining the current WHO-TEFs for DLCs, the highest priority was given to in vivo studies as these include both toxicokinetic and toxicodynamic aspects (Van den Berg et al., 2006). During the latest WHO expert meeting in 2005, a number of uncertainties concerning the current TEFs were also brought forward. One of these concerns was related to the question whether or not the current TEFs, primarily based on in vivo studies using oral dosage as the principal route of exposure, can be used for risk assessment based on a systemic concentration in e.g. human blood or tissues. Another important aspect is the possible species-specific difference in potency of DLCs and consequently the reliability of the current TEFs based on rodent studies for accurate human risk assessment.

The research described in this thesis was part of a large international EU project, called SYSTEQ (www.systeqproject.eu), that had above-mentioned concerns as main objectives. In chapters 2 and 3 of this thesis, relative potency on the basis of intake or systemic concentrations (intakeREPs and systemicREPs) were compared in female C57bl/6 mice and Sprague-Dawley rats based on the administered dose as well as on liver, adipose, or plasma concentrations. Studies were performed with 2,3,7,8-tetrachlorodibenzodioxin (TCDD), 1,2,3,7,8-pentachlorodibenzo-p-dioxin (PeCDD), 2,3,4,7,8-pentachlorodibenzofuran (4-PeCDF), 3,3’,4,4’,5-pentachlorobiphenyl (PCB 126), 2,3’,4,4’,5-pentachlorobiphenyl (PCB 118) and 2,3,3’,4,4’,5-hexachlorobiphenyl (PCB 156) and the non-dioxin-like 2,2’,4,4’,5,5’-hexachlorobiphenyl (PCB 153). Three days after oral exposure, intakeREPs and systemicREPs were calculated based on hepatic cytochrome P450 (CYP)1A1 associated ethoxyresorufin-O-deethylase (EROD) activity and Cyp1a1, 1a2, and 1b1 gene expression in the livers and peripheral blood lymphocytes (PBLs). Results from both studies show that systemicREPs can deviate significantly from intakeREPs (see Figure 2, Chapter 2 and Figure 1, Chapter 3). When based on plasma concentrations (the matrix often used for risk assessment) the combined mouse and rat data show that systemicREPs are generally within a half log range around the intakeREPs for all congeners tested, except 4-PeCDF (See Figure 3, chapter 3). For 4-PeCDF, the median plasma-based systemicREP of 0.3 was significantly higher compared to its intakeREP of 0.04. A relevant question related to these three-day single dosing studies, described in chapter 2 and 3, is their comparability with previously conducted studies using a chronic or subchronic dosing regimen. The latter types of studies have formed a major role in the 175

7


derivation of WHO-TEFs (Haws et al., 2006; Van den Berg et al., 2006). The answer to this question is addressed in chapter 4, where we compared tissue distribution across the tested DLCs from single dose studies with previous rodent studies using both single and subchronic dosing regimens. In addition, EC50 values of hepatic concentrationresponse relationships for CYP1A1 activity or its gene expression were evaluated. This comparison shows that the distribution patterns between liver and adipose tissue were comparable for the DLCs in our studies and in other studies using either a single dose or subchronic dosing. In addition, the CYP1A1 concentration-effect relationships for TCDD were found to be comparable between the different oral dosing regimens. Based on these observations it can be concluded that systemicREPs calculated using our threeday studies, as described in this thesis, are likely to also be applicable for (sub)chronic exposure situations.

In chapters 5 and 6 of this thesis, species-specific differences in REPs between human and rodents have been investigated for twenty DLCs in both primary cell systems (human peripheral blood lymphocytes (PBL) and mouse splenocytes), as well as in cell-lines (mouse, rat, guinea pig and human) that have chemical-activated luciferase expression (DR-CALUX速). In chapter 5, REPs were determined based on CYP1A1, 1B1 and aryl hydrocarbon receptor repressor (AhRR) gene expression as well as CYP1A1 activity in human PBLs. Results were compared with Cyp1a1 gene expression in mouse splenic cells. The results presented in this chapter show that the human PBL-derived REPs for 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin (1234678-HpCDD), 4-PeCDF, 1,2,3,4,7,8-hexachlorodibenzofuran (123478-HxCDF) and 1,2,3,4,7,8,9-heptachlorodibenzofuran (1234789-HpCDF) were significantly higher compared to REPs determined in mouse splenic cells. Similar differences were observed between human PBL-derived REPs and WHO-TEFs. In contrast, the REP for PCB 126 in human PBLs was significantly lower compared to its REP from mouse splenic cells and the present WHO-TEF. In chapter 6, the potency of twenty DLCs was determined using AhR-dependent luciferase reporter gene bioassays from rat, mouse and human hepatoma cells, and guinea pig intestinal adenocarcinoma cells. Furthermore, quantitative structure-activity relationship (QSAR) analysis was performed to predict dioxin-like activity of structurally similar, but untested compounds. These QSARs were also used to examine possible structural analogies that may influence the activity of a group of compounds. The results show, similar to chapter 5, higher human-derived REPs for 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF compared to those derived from the three rodent species. Again, the human cell-line derived REP for PCB 126 was significantly lower than the REPs observed in the rodent derived CALUX速 cell-lines. No REP for any of the other PCBs tested could be determined in the human CALUX速 cell-line due to their low dioxin-like activity. Principal Component Analysis 176


Summary, Discussion and Conclusions

(PCA) and corresponding loading plots indicate that REPs derived from rat, mouse and guinea pig CALUX® cell-lines are similar to each other and to the present WHO-TEFs, but not to the human-derived CALUX® REPs. In rodent CALUX® assays, differences in LUMO and HOMO energy (GAP), total positive van der Waals surface area (PEOE_VSA_POS), Balaban’s connectivity topological index (Balaban’s index), selected UV descriptors and shape index (Kier3) were the most significant descriptors. In the human CALUX® assay total positive and negative partial charges (PEOE_PC+ and PEOE_PC-) were identified as the most influential descriptors. These differences between the human and rodent CALUX® assays may indicate a different ligand-receptor interaction between humans and rodents. Taken together, the in vitro data from chapter 5 and 6 show clear congenerand species-specific differences between humans and rodents, with higher in vitro REPs for 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF and a distinct lower REP for PCB 126 in the human cell model compared to those from the three rodent species. General Discussion Without doubt, the toxicological or biological potency of a compound depends on both toxicokinetic and toxicodynamic factors. The toxicodynamic processes are in particular responsible for compound- and species-specific differences based on molecular interactions e.g. at tissue or receptor level in a target tissue. On the other hand, toxicokinetic factors like metabolism and body distribution govern the systemic concentrations in the body and can be species- and compound-specific. Within this thesis, both aspects have been investigated and will be discussed with respect to their consequences for relative effect potencies and human risk assessment of DLCs.

The impact of toxicokinetics on the relative potency of DLCs Differences between the intake and systemic REPs described in this thesis are, at least in part, the result of variation in distribution due to CYP1A2 protein binding in the liver. However, besides distribution also other toxicokinetic aspects like absorption, metabolism and elimination can potentially influence the relative potency of a DLC (DeVito et al., 1997).

The selected congeners for the in vivo studies described in this thesis behave similarly to TCDD with respect to absorption, metabolism, and elimination (Van den Berg et al., 1994). These congeners were chosen because of their contribution to the overall toxic equivalency (TEQ) in human food based on the current WHO-TEFs. However, if the systemicREP of other congeners would be significantly different from their intakeREP, due to differences in toxicokinetic aspects relative to TCDD it might be possible that 177

7


these other congeners also contribute significantly to systemic TEQs even when these are currently not significant contributors on the basis of present WHO-TEFs. It can be expected that toxicokinetic considerations will have the greatest effect on congeners that are very different in terms of absorption, distribution, metabolism or elimination compared to TCDD. Examples are 2,3,7,8-tetrachlorodibenzofuran (TCDF), 1,2,3,7,8-pentachlorodibenzofuran (1-PeCDF), 1,2,3,4,6,7,8,9-octachlorodibenzo-pdioxin (OCDD), 1,2,3,4,6,7,8,9-octachlorodibenzofuran (OCDF) and 4-PeCDF. TCDF and 1-PeCDF are much more rapidly metabolized and eliminated than TCDD, OCDD and OCDF are more poorly absorbed than TCDD and 4-PeCDF is sequestered in liver to a much higher degree than TCDD.

Absorption, Metabolism and Elimination Passage across the intestinal wall is predominantly limited by the molecular size and solubility of the congener. The higher the molecular weight of a DLC the more difficult it is to be absorbed from the gastrointestinal (GI) tract. Several rodent studies report an absorption rate of 70 – 90% for TCDD or TCDF, but only 2 – 15% for OCDD in rat and mice after oral administration (Van den Berg et al., 1994). In humans, also an absorption rate of around 90% was observed for TCDD. Interestingly, for OCDD and OCDF the absorption rate for humans was estimated to be 40 – 50%, which is considerably higher than that observed in rodents (Andreas Moser and McLachlan, 2001). Whether this is a real difference between humans and rodents or it is an artifact due to a more indirect method for estimating the absorption is uncertain. However, the results indicate that OCDD and OCDF in rodents and humans are more poorly absorbed from the GI tract compared to TCDD. Consequently, there will clearly be lower systemic concentrations of OCDD and OCDF when compared to TCDD for a similar oral dose. As a result of these GI absorption differences, systemicREPs for OCDD and OCDF could be significantly higher compared to intakeREPs. This phenomenon is outlined in more detail in Figure 1.

Once OCDD and OCDF are absorbed from the GI tract these DLCs are mainly retained in the liver (Birnbaum and Couture, 1988). This is usually attributed to strong CYP1A2 binding, but also likely occurs because transport across membranes into different compartments of the body becomes more difficult for the highly chlorinated DLCs. This was also shown in a human study by Wittsiepe et al, who found increasing blood:milk ratios of DLCs with increasing molecular weight (Wittsiepe et al., 2007). Differences between OCDF and TCDD were also observed in a (sub)chronic 13 week steady-state study in mice. This study showed that exposure to OCDF resulted in an approximately 80-fold higher systemicREPs based on skin concentrations compared to the associated intake REPs (DeVito et al., 1997). 178


Summary, Discussion and Conclusions

In contrast to OCDD and OCDF, TCDF and 1-PeCDF are absorbed from the GI tract in a similar degree as TCDD. However, once these congeners are absorbed, they are much more rapidly metabolized and eliminated than TCDD (Van den Berg et al., 1994). In rat and mouse studies, TCDF and 1-PeCDF have a whole body half-life of approximately 2 and 6 days, respectively, which is much shorter than that of TCDD, which is between 17 and 31 days (Van den Berg et al., 1994). These congener-specific differences result in relatively lower systemic concentrations of TCDF and 1-PeCDF compared to TCDD for a similar intake dose. Thus systemicREPs for TCDF and 1-PeCDF might be significantly higher compared to intakeREPs (outlined in Figure 1). The only available study where intake REPs and systemicREPs were compared for TCDF and 1-PeCDF is a 13-week steady state mice study (DeVito et al., 1997). This study shows that systemicREPs based on skin concentration are 5 and 15-fold higher compared to intakeREPs for TCDF and 1-PeCDF, respectively. Congener vs  TCDD  

Absorp.on

Higher absorp.on  

Lower absorp.on  

Lower systemic  REP  

Higher systemic  REP  

Metabolism /   Elimina.on    

Distribu.on

More   liver   sequestra.on  

Lower   liver-­‐based   systemic  REP  

Higher plasma-­‐based   systemic  REP  

Less   liver  sequestra.on  

Higher   liver-­‐based   systemic  REP  

Lower   plasma-­‐based   systemic  REP  

More   rapidly  

Less   rapidly  

Higher systemic  REP  

Lower systemic  REP  

Figure 1. Hypothetical impact scheme of toxicokinetic differences in absorption, distribution, metabolism and elimination between TCDD and another congener on the REP based on either a systemic concentration or administered dose.

Distribution As mentioned above and in chapters 2 and 3, the relative potency of a congener can also be altered by differences in their disposition compared to TCDD. At least in rodents, differences in disposition are caused by congener-specific differences in liver sequestration due to CYP1A2 protein binding. It is known that the binding affinity towards CYP1A2 differs between DLCs due to differences in structure and number of 179

7


chlorine atoms (Van den Berg et al., 1994). In general, 2,3,7,8-substituted PCDFs are sequestered to a greater extent than their dioxin analogues, while the highly chlorinated congeners are sequestered more than the less chlorinated congeners. Furthermore, PCBs, except PCB 126, do not sequester in the liver. At present it is unclear whether DLCs that are bound to CYP1A2 are still bioavailable and can activate the AhR to cause dioxin-like responses. For this reason, systemicREPs calculated on total hepatic tissue concentration instead of the â&#x20AC;&#x153;freeâ&#x20AC;? available concentrations may lead to either an overor under-estimation of the potency of a congener, depending on the relative degree of hepatic sequestration compared to TCDD. Data presented in chapters 2 and 3 clearly demonstrate that 4-PeCDF sequesters to a much higher degree than TCDD in rat and mouse liver. As a result, measured hepatic tissue concentrations are higher than those obtained for TCDD at similar intake dose levels. Consequently, systemicREPs based on hepatic concentrations and hepatic effects will be lower compared to intakeREPs. At the same time, plasma or adipose tissue concentrations are relatively lower. Thus, systemicREPs that are calculated based on extrahepatic tissue concentrations will be higher compared to intakeREPs (see Figure 1). In contrast to 4-PeCDF, the mono-ortho PCBs 118 and 156 do not sequester in the liver, but rather distribute based on the lipid content of various tissues. As a result, the highest concentrations for these congeners are found in adipose tissue. Consequently, hepaticbased REPs for PCB 118 and 156 are much higher compared to intakeREPs, and plasma or adipose tissue-based systemicREPs are lower.

In vitro derived REPs as predictor for in vivo systemic REPs Translating in vitro-derived REPs to an in vivo situation is a challenge for risk assessment, as toxicokinetic properties are often not taken into account. However, it is possible that in vitro studies better represent the actual potency of a congener determined at a target tissue and in vitro-derived REPs may better predict in vivo REPs based on a systemic concentration. In figure 2, an overview is given for the intake-, systemic- and in vitroderived REPs of PeCDD, 4-PeCDF and PCB 126 that have been obtained from the in vivo studies described in chapter 2 and 3 and the in vitro studies described in chapter 5 and 6. For PeCDD and 4-PeCDF, these data indeed indicate that the in vivo plasma-based systemic REPs are closer to the in vitro determined REPs. Such similarities are less distinct for PCB 126. However, it should be noted that in particular for the in vivo studies of PCB 126 described in chapter 2 and 3, significant differences were observed between BMR20TCDD and EC50 derived REPs, with up to 10-fold higher EC50 REPs (data not shown). These differences were less pronounced for the other congeners tested in vivo, when compared to the in vitro derived REPs. For PCB 118 and 156, a comparison between the in vivo and in vitro derived REPs was not possible, as dose-response curves from our in 180


Summary, Discussion and Conclusions

vivo studies were incomplete. Nevertheless, the data for PeCDD and 4-PeCDF provide indications, that in vitro REPs can be more predictive for systemicREPs than intakeREPs. PeCDD

4-PeCDF

rodent - in vivo - intake

rodent - in vivo - systemic

rodent - in vitro

0.0

01

1 0.0

0.1

1

10

0 10

1

10

0.0

01

1 0.0

0.1

1

10

0 10

PCB 126

rodent - in vivo - intake

rodent - in vivo - systemic

rodent - in vitro 1 00 0. 0

0.0

01

1 0.0

0.1

Figure 2. Comparison of in vivo derived REPs based on either intake dose or systemic plasma concentration with in vitro derived REPs in relation to the WHO-TEF Âą half log uncertainty range (black dotted line and grey area).

Species- and congener-specific differences in REPs It is generally assumed that REPs or TEFs based on rodent studies are appropriate for human risk assessment. Yet, it is well known that upon AHR activation a wide variety of species-specific toxic and biological effects can occur (Denison et al., 2011). Some of the species differences in AHR-mediated responses can clearly be attributed to genetic variations. Generally, the human AHR is considered to be relatively less responsive to DLCs than the AHR from rodents. However, also within species large differences exist. For example, DBA mouse strains are relatively resistant to TCDD toxicity in contrast with C57bl/6 mice (Connor and Aylward, 2006; Ema et al., 1994). Nonetheless, differences in AHR affinity between species are considered not to be uniquely responsible for speciesspecific differences in toxicity of DLCs. Several authors have suggested that congener181

7


specific REPs can vary across species due to intrinsic differences in e.g. efficacy. In particular, the species-differences in REP of the non-ortho substituted PCB 126 have been subject of much scientific debate (Nagayama et al., 1985; Silkworth et al., 2005; Sutter et al., 2010; Van Duursen et al., 2003; 2005; Zeiger et al., 2001). In addition, some PCDDs and PCDFs, such as 4-PeCDF, 123478-HxCDF and 1,2,3,6,7,8-hexachlorodibenzofuran (123678-HxCDF) also show species-specific differences in REPs (Nagayama et al., 1985; Sutter et al., 2010).

In chapters 5 and 6 of this thesis, species-specific differences in REPs between human and rodents have been investigated for twenty DLCs in primary cell systems (human peripheral blood lymphocytes (PBL) and mouse splenocytes) as well as in chemicalactivated luciferase expression (DR-CALUX速) cell lines (mouse, rat, guinea pig and human). The calculated REPs from human in vitro models for 1234678-HpCDD, 4-PeCDF, 123478-HxCDF, 1234789-HpCDF and PCB 126 deviate significantly from the rodent derived REPs. To illustrate these differences between human and rodent derived REPs more clearly, we combined the REPs from both chapters and compared these with the 2004 REP database (Haws et al., 2006), on which the current WHO-TEFs are based together with newly published human data. For PCB 126, the WHO-TEF of 0.1 is consistent with the median of REPs of 86 in vivo and 29 in vitro REPs obtained from 20 and 19 studies, respectively (Haws et al., 2006). These in vivo REPs comprise of 23 mouse- and 63 rat-based REPs, the latter mainly consisting of in vivo studies from the National toxicology program (NTP) using female Sprague-Dawley rats (National Toxicology Program, 2006c). Especially, REPs for PCB 126 from rat studies are consistently close to 0.1 (Haws et al., 2006). However, a wider distribution exists for the mouse-based in vivo REPs in this database. The median in vitro REP from rodents for PCB 126 is with 0.09, very similar to the median in vivo REP of 0.1.

Only 8 of the 29 in vitro REPs for PCB 126 within this 2004 REP database are derived from studies in human cells (Drenth et al., 1998; Pang et al., 1999; Van Duursen et al., 2003; Zeiger et al., 2001). Compared with the rodent data, the human in vitro REPs are clearly much lower with a median REP of 0.009 (range 0.0007-0.02) (Figure 3A). Since 2005, another five in vitro studies with human primary PBLs, hepatocytes, keratinocytes or human hepatoblastoma cells (HepG2) were conducted that determined the relative potency of PCB 126, with REPs ranging from 0.00009 to 0.11 (Carlson et al., 2009; Silkworth et al., 2005; Sutter et al., 2010; Van Duursen et al., 2005; Westerink et al., 2008). When we compare our data with data from the 2004 REP database and newly published literature, it is evident that both our rodent and human in vitro REPs are 182


Summary, Discussion and Conclusions

very comparable to those reported in literature (See Figure 3B and C). In vitro REPs from human cell systems are consistently at least one, but possibly up to two, orders of magnitude lower than the current WHO-TEF. All in vitro REPs

2004 REP database (Haws et al.)

115

all REPs rodent - vivo

86

rodent - vitro

19

rodent - vitro - literature

19

rodent - vitro - SYSTEQ

human - vitro

0. 0

A

human - vitro - SYSTEQ

8

01 01 01 0.01 00 0. 00 0. 0

0.1

1

10

0. 0

4

B

6

01 01 01 0.01 00 0. 00 0. 0

0.1

1

10

Human REPs all human in vitro data

26

vitro - SYSTEQ

vitro - literature

0. 0

6 19

C

01 01 0.01 01 00 0. 00 0. 0

0.1

1

10

Figure 3. Boxplot comparison of in vivo and in vitro derived REPs for PCB 126 based on rodent or human data from literature and this thesis. Graph A. represents the 2004 REP database that was used to calculate the WHO-TEF of 0.1 for PCB 126. REPs and selection criteria are published elsewhere (Haws et al., 2006). The WHO-TEF for PCB 126 is based on 115 REPs of which 86 are in vivo REPs (23 mouse, 63 rat) and 29 in vitro REPs (3 mouse, 8 human, 16 rat, 1 primate, 1 pig). Graph B. represents rodent and human in vitro derived REPs. Upper bar represents rodent in vitro REPs from the 2004 database (n=19), middle bar represents rodent in vitro REPs from this thesis (n=4) and the lower bar represents human in vitro REPs from this thesis (n=6). Graph C. represents all human REPs. Upper bar represents human in vitro REPs from the 2004 database (n=8), new literature (n=10) and this thesis (n=6), middle bar represents human in vitro REPs from this thesis (n=6) and the lower bar represents human in vitro REPs from literature (n=18).

For 4-PeCDF, the 2004 REP database contains 80 in vivo and 17 in vitro REPs obtained from 20 and 10 studies, respectively (Haws et al., 2006). The in vivo REPs comprise 22 mouse- and 58 rat-based REPs, the latter again mainly consisting of in vivo studies from the NTP using female Sprague-Dawley rats (National Toxicology Program, 2006d). The median rodent in vitro REP for 4-PeCDF is 0.7, which is notably higher compared to the 183

7


median in vivo REP of 0.2 (See Figure 4A). The rodent in vitro REPs from this thesis are in line with the REPs from the 2004 REP database (See Figure 4B). Since 2005, one new human in vitro study published REPs for 4-PeCDF using primary human hepatocytes (Budinsky et al., 2010). The median human in vitro REP from the combined literature data is 0.8, which is comparable to the human in vitro REPs from this thesis (See Figure 4C). Furthermore, the human in vitro REPs of this congener are in line with the rodent in vitro derived REPs. These data indicate that for 4-PeCDF there are no species-specific differences. However, the data do suggest a difference between in vivo and in vitro REPs, which is likely due to the toxicokinetics properties of 4-PeCDF as described earlier. 2004 REP database (Haws et al.) all REPs

97

rodent - vivo

80

rodent - vitro

17

All in vitro REPs

17

rodent - vitro - literature

rodent - vitro - SYSTEQ

human - vitro

A

01 0. 0

human - vitro - SYSTEQ

5 0.0

1

0.1

1

10

0 10

01 0. 0

4 6

B 0.0

1

0.1

1

10

0 10

Human REPs

15

all human in vitro data

6

vitro - SYSTEQ

vitro - literature 01 0. 0

9

C 0.0

1

0.1

1

10

0 10

Figure 4. Boxplot comparison of in vivo and in vitro derived REPs for 4-PeCDF based on rodent or human data from literature and this thesis. Graph A. represents the 2004 REP database that was used to calculate the WHO-TEF of 0.3 for 4-PeCDF. REPs and selection criteria are published elsewhere (Haws et al., 2006). The WHO-TEF for 4-PeCDF is based on 97 REPs of which 80 are in vivo REPs (2 guinea pig, 21 mouse, 57 rat) and 17 in vitro REPs (2 mouse, 5 human, 10 rat). Graph B. represents rodent and human in vitro represents rodent in vitro REPs from this thesis (n=4) and the lower bar represents human in vitro REPs from this thesis (n=6). Graph C. represents all human REPs. Upper bar represents human in vitro REPs from the 2004 database (n=5), new literature (n=4) and this thesis (n=6), middle bar represents human in vitro REPs from this thesis (n=6) and the lower bar represents human in vitro REPs from literature (n=9).

For 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF, the 2004 REP database contains respectively 12, 6 and 0 in vivo REPs and 5, 7 and 2 in vitro REPs (Haws et al.,

184


Summary, Discussion and Conclusions

2006). The 12 in vivo REPs for 1234678-HpCDD were all obtained from rat studies. Whereas for 123478-HxCDF, the in vivo REPs have been derived from 3 mouse- and 3 rat-based studies. The median rodent based in vitro REPs for 1234678-HpCDD and 123478-HxCDF are 0.03 and 0.3, which is distinctly higher than the median in vivo REP of 0.01 and 0.05, respectively (See Figure 5A and Figure 6A). Based on data described in this thesis, the median rodent and human in vitro REPs for 1234678-HpCDD are 0.08 and 1, respectively, and clearly higher compared to the 2004 REP database rodent in vitro REP of 0.03 (See Figure 5B and 5C). All in vitro REPs

2004 REP database (Haws et al.)

all REPs

19

rodent - vitro - SYSTEQ

4

5

rodent - vitro

0. 0

5

12

rodent - vivo

human - vitro

rodent - vitro - literature

A 01 0. 0

1 00

human - vitro - SYSTEQ

1 0.0

1

0.1

1

0. 0

1 00

5

B 01 0. 0

0.0

1

0.1

1

Human REPs all human in vitro data

6

vitro - SYSTEQ

5

vitro - literature

0. 0

1 00

1

C 01 0. 0

0.0

1

0.1

1

Figure 5. Boxplot comparison of in vivo and in vitro derived REPs for 1234678-HpCDD based on rodent or human data from literature and this thesis. Graph A. represents the 2004 REP database that was used to calculate the WHO-TEF of 0.01 for 1234678-HpCDD. REPs and selection criteria are published elsewhere (Haws et al., 2006). The WHO-TEF for 1234678-HpCDD is based on 19 REPs of which 12 are in vivo REPs (all rat) and 6 in vitro REPs (1 mouse, 1 human, 4 rat). Graph B. represents rodent and human in vitro derived REPs. Upper bar represents rodent in vitro REPs from the 2004 database (n=6), middle bar represents rodent in vitro REPs from this thesis (n=4) and the lower bar represents human in vitro REPs from this thesis (n=5). Graph C. represents all human REPs. Upper bar represents human in vitro REPs from the 2004 database together with the human REPs derived in this thesis (n=6), middle bar represents human in vitro REPs from this thesis (n=5) and the lower bar represents human in vitro REP from the 2004 database (n=1).

For 123478-HxCDF, a similar REP-range was seen for the rodent data from this thesis 185

7


in comparison with the 2004 REP database, with median rodent in vitro REP of 0.2 and 0.3, respectively (See Figure 6B). However, a significantly higher human REP range with a median of 1.2 was seen for 123478-HxCDF compared to the rodent REPs (See Figure 6B). For 1234789-HpCDF, a comparison with literature can only be made based on in vitro data, as no in vivo studies were reported in the 2004 REP database. The median rodent and human in vitro REPs from this thesis are both 0.1, which are clearly higher than the 2 in vitro REPs (0.02 and 0.04) from the 2004 REP database and its WHO-TEF of 0.01 (See Figure 7). Since 2005, no new studies have been published for 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF. As shown in chapters 5 and 6, all three congeners have higher human-derived REPs compared to rodent-derived REPs in the in vitro test systems. However, when the data from chapter 5 and 6 are combined this species difference only remains for 123478-HxCDF. Even so, these data show that rodent and/or human REPs are one to two orders of magnitude higher compared to their WHO-TEFs and suggest at least for 1234678-HpCDD and 123478-HxCDF a distinct difference between in vivo and in vitro REPs. All in vitro REPs

2004 REP database (Haws et al.) all REPs

13

rodent - vitro

01 0. 0

7

A 0.0

1

0.1

4

rodent - vitro - SYSTEQ

6

rodent - vivo

7

rodent - vitro - literature

1

10

human - vitro - SYSTEQ

01 0. 0

5

B 0.0

1

0.1

1

10

Figure 6. Boxplot comparison of in vivo and in vitro derived REPs for 123478-HxCDF based on rodent or human data from literature and this thesis. Graph A. represents the 2004 REP database that was used to calculate the WHO-TEF of 0.1 for 123478-HxCDF. REPs and selection criteria are published elsewhere (Haws et al., 2006). The WHO-TEF for 123478-HxCDF is based on 13 REPs of which 6 are in vivo REPs (3 mouse, 3 rat) and 7 in vitro REPs (all rat). Graph B. represents rodent and human in vitro derived REPs. Upper bar represents rodent in vitro REPs from the 2004 database (n=7), middle bar represents rodent in vitro REPs from this thesis (n=4) and the lower bar represents human in vitro REPs from this thesis (n=5).

186


Summary, Discussion and Conclusions All in vitro REPs

2

rodent - vitro - literature

4

rodent - vitro - SYSTEQ

5

human - vitro - SYSTEQ

0. 0

1 00

01 0. 0

0.0

1

0.1

1

Figure 7. Boxplot comparison of in vitro derived REPs for 1234789-HpCDF based on rodent or human data from literature and this thesis. The upper bar represents the 2004 REP database (Haws et al., 2006) that was used to calculate the WHO-TEF of 0.01 for 1234789-HpCDF. The WHO-TEF for 1234789-HpCDF is based on 2 in vitro REPs (1 mouse, 1 rat). The middle bar represents the rodent in vitro REPs from this thesis (n=4) and the lower bar represents the human in vitro REPs from this thesis (n=5).

Human risk assessment Setting correct TEFs for DLCs is of great toxicological relevance. Human exposure to DLCs always occurs as complex mixtures. The number of chlorine atoms and substitution pattern of a congener as well as aspects like metabolism and body distribution strongly determine the biological and toxicological activity of the congener. Congeners with a 2378 chlorine substitution pattern are of highest relevance for risk assessment, but even within this group significant differences in toxicological effects and accumulation can be observed. Thus, to establish possible risks for humans it is necessary to evaluate individual PCDDs and PCDFs, even among the group of 2378 substituted congeners. As human risk assessment is often performed by the measurement of congener concentrations in human blood, TEF values should be considered to ensure that â&#x20AC;&#x153;intakeâ&#x20AC;? TEFs are applicable for measurements of blood concentration. Results from chapter 2 and 3 show that plasma-based systemicREPs derived from in vivo rodent studies for PeCDD, PCB 126, 118 and 156 are within a half log of the intakeREPs. Only for 4-PeCDF, plasma-based systemicREPs were considerably higher than REPs based on oral dosage. Differences between intakeREPs and systemicREPs within our studies are primarily caused by differences in distribution between hepatic or extra-hepatic tissue. Theoretically, congener-specific hepatic sequestration in humans is possible, as CYP1A2 is one of the more prominent P450 enzymes present in the human liver (Bieche et al., 2007). This is also supported by some studies in humans with elevated exposure levels, where induction of CYP1A2 activity was found using the caffeine breath test (Abraham et al., 2002; Lambert et al., 2006). However, very minor hepatic sequestration occurs 187

7


within the human population exposed to environmentally background concentrations (Iida et al., 1999; Thoma et al., 1990; Watanabe et al., 2013; Weistrand and Norén, 1998).

In extra-hepatic tissue, there is relatively little CYP1A2 protein available (Bieche et al., 2007). This means that blood concentrations may better reflect the available free concentration of DLCs causing an AhR response, because potential sequestration of these compounds due to CYP1A2 plays less likely a role of importance. It can therefore be expected that for humans a similar shift for plasma-based systemicREPs compared to intake REPs will occur for those congeners tested in our rodent studies. In other words, the results described in this thesis do not warrant the development of separate “systemic” TEFs for those congeners tested in the EU-SYSTEQ project, with exception of 4-PeCDF. With respect to 4-PeCDF, both rodent- and human-derived in vitro REPs from this thesis show median REPs of 0.8 (0.2-1.3) and 1 (0.6-2.2), which are significantly higher than the present WHO-TEF. Our results are in agreement with findings from literature (Budinsky et al., 2010; Haws et al., 2006). Taken together, the in vivo and in vitro derived REPs of 4-PeCDF might suggest that separate TEFs for intake versus systemic approaches may be appropriate.

Absorption, metabolism and elimination also play an important role in differences between intakeREPs and systemicREPs. The congeners TCDF, 1-PeCDF, OCDF and OCDD are very different from TCDD with respect to toxicokinetics. This was shown in an earlier mouse study, in which these DLCs had higher skin-based systemicREPs compared to intakeREPs (DeVito et al. 2007). In addition, a recently published human study that calculated in vivo REPs based on two thyroid effect parameters showed that systemicREPs were 0.6 and 0.4 for TCDF and OCDF, respectively, thus significantly higher than their WHO-TEFs of 0.1 and 0.0003 (Trnovec et al., 2013). systemicREPs for 1-PeCDF and OCDD on thyroid hormone effects could not be determined in that study. If human and rodent in vitro-derived REPs of TCDF, 1-PeCDF, OCDF and OCDD presented in this thesis or from literature are compared with their WHO-TEFs, it is evident that these are similar for TCDF and OCDD (Budinsky et al., 2010; Haws et al., 2006; Sutter et al., 2010). In addition, a recent study showed that genomic-based REPs for TCDF in primary rat hepatocytes were generally 5-fold lower than its WHO-TEF based on both gene and pathway analysis (Rowlands et al., 2013). In contrast, for 1-PeCDF and OCDD, median in vitro based REPs were 0.1 and 0.002, respectively, which are higher compared to their WHO-TEFs (Haws et al., 2006). For human relevance, another important aspect is the presence in human blood and the quantitative contribution of each congener to the total amount of TEQs. Table 1 presents the concentrations of PCDDs, PCDFs and PCBs in human blood plasma as found in two 188


Summary, Discussion and Conclusions

different studies (Hsu et al., 2005; Rawn et al., 2012). The study of Rawn et al. represents a national baseline estimate of concentrations of PCDDs, PCDFs and PCBs in Canadians. In contrast, the study of Hsu et al. represents a congener profile of PCDDs, PCDFs and PCBs from Yu-cheng victims, fifteen years after exposure to toxic PCB-contaminated rice-bran oils. This table shows that the quantitative contributions to the total amount of TEQs for 1-PeCDF and OCDF are with 0.08% and 0.001%, respectively, very low in the general population. Therefore, a change in TEF value of 1-PeCDF and OCDF would not make a big difference for the total mixture toxicity expressed as TEQs. However, for TCDF and OCDD that contribute 0.4% and 0.5% to the total amount of TEQs in the general population, respectively, an increase in the TEF based on the systemicREP data can potentially be important for risk assessment.

It is of interest to note that after PeCDD, 123678-HxCDD is the second most important contributor to the total amount TEQs in the general population based on the current WHO-TEFs. Although 123678-HxCDD has always been prominently present, it appears to become a more important contributor over time (Ferriby et al., 2007; Kang et al., 1990). One of the reasons for this might be the longer half-life of 123678-HxCDD compared TCDD in humans. 123678-HxCDD appears to be one of the most slowly eliminated congeners in humans (Aylward et al., 2013; Flesch-Janys et al., 1996; Rohde et al., 1999). As body burdens decline over time, 123678-HxCDD becomes a more important contributor to the total amount of TEQs in humans and their environment. Human and rodent in vitro data from this thesis as well as from literature are in agreement with the WHO-TEF of 1 and 0.1 for PeCDD and 123678-HxCDD, respectively. In contrast, the median rodent and human in vitro REPs of 1234678-HpCDD is 0.1, which is one order of magnitude higher compared to the current WHO-TEF of 0.01. If the WHO-TEF for 123678-HpCDD would be adjusted from 0.01 to 0.1, this would most significantly increase its contribution to the total amount of TEQs (See Table 1) (Rawn et al. 2012 study), and this congener would become one of the major contributors to the total amount of TEQs for the general population. Of the furans, in particular 4-PeCDF is a significant contributor. If the WHO-TEF would be adjusted from 0.3 to 1, this would imply a significant increase in the total amount of TEQs. For 123478-HxCDF clearly human-specific differences in REPs were seen compared to rodents in this thesis. The median human in vitro-derived REP is at 1 significantly higher than the WHO-TEF of 0.1. Changing the TEF of this congener from 0.1 to 1 does not cause a significant change in contribution to the total amount of TEQs in the general population as reported by Rawn et al. (2012) (See Table 1). However, it might have a significant contribution in the accidental poisoning case studied by Hsu et al. (2005) (See table 1). Beside 123478-HxCDF, also 123678-HxCDF was found to have human189

7


Table 1: Concentrations, TEQ (pg/g lipid) and % contribution to total TEQ of PCDD/Fs and DLC PCBs in human blood from a general (Rawn et al., 2012) and an exposed population (Hsu et al. 2005). Rawn et al. 2012

IntakeTEFa

SYSTEQREP

mean Intake(pg/g lipid) TEQ

% of total Intake-TEQ

1

1

3,7

28,1

2378-TCDD

1

123478-HxCDD

0,1

12378-PeCDD

123678-HxCDD 123789-HxCDD

0,1 0,1

OCDD

12378-PeCDF

1234678-HpCDD 2378-TCDF

23478-PeCDF

123478-HxCDF 123678-HxCDF 123789-HxCDF 234678-HxCDF

1234678-HpCDF

PCB 169 PCB 105 PCB 114 PCB 118 PCB 123 PCB 156 PCB 157

0,3

27

2,7

2,3

20,5

0,26

2,0

0,0003

0,0003

180

0,054

0,4

0,03

0,03

0,35

0,0105

0,08

0,1

0,1

3

3,7

4,0

0,1

0,01

0,1

0,3

1

26

0,61

0,061

2,8 0,5

1,62

12,3

0,42

0,042

0,3

1,3

0,13

0,1

1

5,4

0,1 0,1

0,1

0,41

3,1

0,1

0,1

4,1

0,1

0,35

0,035

0,3

PCB 81

PCB 126

0,1

0,53

0,37

OCDF

PCB 77

0,1

0,53

3,7

0,01

1234789-HpCDF

1

0,1

1,3

0,0054

0,04

0,0003

0,0003

0,54

0,013

0,58

0,0003 0,003

10

0,001

0,0013

0,0003

10

0,000174 0,003 2

0,023

15,2

0,03

20

0,6

4,6

0,01

0,0001 0,1

0,03

0,01

0,99

0,0001

0,00003

0,00003

0,00003

0,00003

0,00003 0,00003 0,00003 0,00003

0,00003 0,00003 0,00003 0,00003

20

0,10

0,008

1100

0,033

0,3

6300

0,189

1,4

480 60

400 940

0,0144 0,0018 0,012

0,0282

0,1 0,01 0,1

0,2

PCB 167 0,00003 0,00003 990 0,0297 0,2 wTable 1: Concentrations, TEQ (pg/g lipid) and % contribution to total TEQ of PCDD/Fs and DLC 190


Summary, Discussion and Conclusions PCBs in human blood from a general (Rawn et al., 2012) and an exposed population (Hsu et al. 2005). Hsu et al. 2005

SYSTEQTEQ

% of total mean SYSTEQ-TEQ (pg/g lipid)

3,7

20,8

0,53 0,3 2,7

1,7

48,7

4,87

15,2 2,1

69,7 101

0,054

0,3

0,0105 5,4

0,06

30,4

0,42

2,4

755

0,41

2,3

0,035

0,2

0,3

0,7

0,013

0,054

0,3

0,00017

0,001

0,003 0,06

0,0

0,6

3,4

0,0144

0,1

0,033 0,189

0,1 0,0

0,3

0,2 1,1

0,0018

0,01

0,0282

0,2

0,012

0,0297

6,1

65,8

14,6

0,001

69,7

65,8

2,6

0,13

% of total Intake-TEQ

3,0

0,37 0,061

IntakeTEQ

0,1

0,2

56,4

10,1

% of total SYSTEQ-TEQ

69,7

2,4

5,7

65,8

0,4

4,87

0,9

10,1

2,3 0,2 0,4

1,51

0,5

0,1

5,64

15,1

0,2

1745

0,5235

0,05

0,5235

0,02

292

8,76

226,5

0,8

19,7

8,76

755

0,3

26,5

159,1

13,8

1591

55,8

117

11,7

1,0

11,7

0,4

58

5,8

0,5

5,8

0,2

151

102

1591 40,2 177

5,64

SYSTEQTEQ

10,2

4,02

0,3

10,2

4,02

0,4

0,1

1,54

0,2

0,1

1,77

15,4

0,06

1024

0,3072

0,0

0,3072

0,01

2940

294

25,6

8,82

0,3

8460

253,8

22,1

253,8

8,9

2620

0,0786

0,007

0,0786

154

1450

5720

1,77

0,9

0,5

0,145

0,1716

0,0

0,01

0,145

0,006 0,02

0,468

0,04

0,468

318000

9,54

0,8

9,54

2,994

0,3

0,005

0,1716

15600 99800

0,5

2,994

7

0,003 0,3 0,1 191


Table 1: Continued

PCB 167

PCB 189

Total PCDD-TEQ

Rawn et al. 2012 IntakeTEFa

0,00003

0,00003

Total PCDF-TEQ

Total non-ortho-PCBs-TEQ

Total mono-ortho-PCBs-TEQ

a

Total TEQs

Current WHO-TEF (Van den Berg et al., 2006)

SYSTEQREP

0,00003

0,00003

mean (pg/g lipid)

990

370

IntakeTEQ

0,0297

% of total Intake-TEQ

0,2

0,0111

0,1

2,3

17,7

7,9 2,6

0,32 13,2

60,1 19,8 2,4

100

specific differences in REP with a median REP of 1 compared to its WHO-TEF of 0.1, as determined in a human keratinocytes (Sutter et al., 2010). A higher TEF value for this congener has significant implications, as it has already a contribution of 3% in the general population (Rawn et al. 2012 study). In contrast with dioxins and furans, humans might be less responsive to PCBs compared to rodents. In this thesis, except for PCB 126, none of the other PCBs tested were capable to either induce a dioxin-like response, or a response high enough to calculate a REP for CYP1A1 gene expression or AhR-dependent luciferase response in human (primary) cells (See chapter 5 and 6). These differences in response between dioxins and furans on the one side and PCBs on the other may be due to differences in AhR binding mechanisms that are governed by the physico-chemical properties of either the PCDD, PCDF or PCB (Petkov et al., 2010). The very low or absent response by PCBs in human cells observed in our studies are in agreement with earlier studies using human primary cells or cell lines derived from the liver, breast, prostate, lymphocytes, or keratinocytes (Endo et al., 2003; Silkworth et al., 2005; Spink et al., 2002; Sutter et al., 2010; Van Duursen et al., 2003; 2005; Zeiger et al., 2001). For PCB 126, the combined data from chapters 5 and 6 in combination with data from literature gives a median human REP of 0.003. This is almost a 100 times lower than the WHO-TEF of 0.1 and far outside its suggested uncertainty range (Carlson et al., 2009; Haws et al., 2006; Silkworth et al., 2005; Sutter et al., 2010; Westerink et al., 2008). A toxicogenomic study where primary rat and human hepatocytes were exposed to PCB 126 indicated that only 5 of the 4000 orthologous genes tested were shared between the rodent species and humans (Carlson et al., 2009). PCB 126 is one of the major contributors to the total amount of TEQs in human blood when using the present WHO-TEFs. If the TEF for PCB 126 would be adjusted from 0.1 to 0.003, the contribution of PCB 126 to the total TEQ goes from 15 to 0.3 % and becomes negligible (Rawn et al. 2012 study). 192


Summary, Discussion and Conclusions

Table 1: Continued

Hsu et al. 2005

SYSTEQTEQ

% of total mean SYS-TEQ (pg/g lipid)

10,3

57,7

0,0297 0,0111 6,5

0,7

0,3

17,8

0,2 0,1

32800

36,8

3,7

1,8

100

urrent WHO-TEF (Van den Berg et al., 2006)

IntakeTEQ

% of total Intake-TEQ

SYSTEQTEQ

158,1

13,8

171,7

0,984

429,7

547,9

14,2

1150,0

0,09

37,4

47,6

1,2

100,0

0,984

2404,0

262,8

14,2

2852,7

% of total SYSTEQ-TEQ

0,03

6,0

84,3

9,2

0,5

100

specific differences in REP with a median REP of 1 compared to its WHO-TEF of 0.1, as determined in a human

If the WHO-TEFs would be adjusted for 1234678-HpCDD, 4-PeCDF, 123478-HxCDF and PCB 126 as described above, it results in a total increase of TEQs by PCDDs and PCDFs of 23% and 64%, respectively. In contrast, a total decrease of TEQs by non-ortho-PCBs of 75% can be expected if taking Rawn et al. (2012) data as an example for the general population. For the overall effect on TEQs it means that these changes in contribution of PCDDs, PCDFs and PCBs to the total amount of TEQs balance each other out. However, for a population that is exposed to a source with specific DLCs this impact on total TEQs might be completely different, depending on the specific mix of congeners, as can be seen in Table 1, for the Hsu et al. (2005) study. Furthermore, species-specific differences in REPs have also been observed for congeners that where not tested in this thesis, for example 123678-HxCDF (Sutter et al., 2010). It cannot be ruled out that these congeners may also have an impact on the total TEQs as well. Conclusions and recommendations The major objectives of this thesis were to establish if there is a need for specific development of systemic or human specific TEFs to improve human risk assessment. Taken all data together, it is evident that for some congeners the current WHO-TEF might under- or overestimate the risk in humans based on plasma concentrations due to either congener-specific toxicokinetics or species differences in response. For 4-PeCDF, up to one order of magnitude higher plasma-based systemicREPs were observed compared to intakeREPs. In addition, rodent and human in vitro-derived REPs were clearly higher when compared to the current WHO-TEF of 0.3. In contrast, for PCB 126, no differences between systemicREPs and intakeREPs were found. However, up to 193

7


two orders of magnitude lower REPs were calculated for PCB 126 in various human cell models, indicating a significantly lower human sensitivity to this DLC compared to rodents. These observations are in close agreement with many other studies that have been reported previously in literature. Based on these data and the prevalence of these two congeners in human biomonitoring data, it is recommended that the WHO-TEFs for 4-PeCDF and PCB 126 are re-evaluated.

In addition, rodent and human in vitro derived REPs for 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF were clearly higher compared to their WHO-TEFs. Findings from literature, although limited, are similar to the results presented in this thesis. Due to their potential significant contribution in total TEQs (if higher TEFs are adopted), further investigation for 1234678-HpCDD and 123478-HxCDF, with special focus on humanspecific REPs is recommended. This thesis shows that, despite the excessive scientific knowledge and huge amount of data that has been published since the development of the TEFs, still significant improvements can be achieved for human risk assessment.

194


Summary, Discussion and Conclusions

Main conclusions and recommendations: • Human in vitro derived REPs for PCB 126 from this thesis and literature are significantly lower than the present WHO-TEF and therefore warrant a re-evaluation of this TEF value for human risk assessment.

• Plasma-based systemicREPs for PeCDD, PCB 126, 118 and 156 are within a half log uncertainty around the intakeREPs and therefore do not warrant the development of separate “systemic” TEFs for these DLCs.

• Rodent and/or human in vitro derived REPs for 1234678-HpCDD, 123478-HxCDF and 1234789-HpCDF are significantly higher than WHOTEFs. These observations should be included in future WHO re-evaluations of TEFs.

• Distribution of dioxin-like compounds within our three-day single dose study is similar to former studies using subchronic dosing. As a result, derived systemicREPs from this thesis can also be of use for situations that include long-term exposures to these compounds. • Data for PeCDD and 4-PeCDF provide support that in vitro REPs can be an adequate replacement for estimation of in vivo plasma-based systemicREPs.

7 195


Annex References

Nederlandse samenvatting

Dankwoord

About the author


198


References

references Abel, J., Haarmann-Stemmann, T., 2010. An introduction to the molecular basics of aryl hydrocarbon receptor biology, Biol. Chem. 391, 1235-1248.

Abraham, K., Geusau, A., Tosun, Y., Helge, H., Bauer, S., Brockmöller, J., 2002. Severe 2,3,7,8 -tetrachlorodibenzo:p-dioxin (TCDD) intoxication: Insights into the measurement of hepatic cytochrome P450 1A2 induction, Clin. Pharmacol. Ther. 72, 163-174. Ahlborg, U.G., Becking, G.C., Birnbaum, L.S., Brouwer, A., Derks, H.J.G.M., Feeley, M., Golor, G., Hanberg, A., Larsen, J.C., Liem, A.K.D., Safe, S.H., Schlatter, C., Waern, F., Younes, M., Yrjanheikki, E., 1994. Toxic equivalency factors for dioxin-like PCBs: Report on a WHO-ECEH and IPCS consultation, December 1993, Chemosphere 28, 1049-1067.

Ahlborg, U.G., Hanberg, A., 1994. Toxic equivalency factors for dioxin-like PCBs, Environmental Science and Pollution Research 1, 67-68. Almenningen, A., Bastiansen, O., Fernholt, L., Cyvin, B.N., Cyvin, S.J., Samdal, S., 1985. Structure and barrier of internal rotation of biphenyl derivatives in the gaseous state. Part 1. The molecular structure and normal coordinate analysis of normal biphenyl and pedeuterated biphenyl, Journal of Molecular Structure 128, 59-76. Almlof, J., 1974. Ab initio calculations on the equilibrium geometry and rotation barriers of biphenyl. Chem.  Phys. 6, 135-139.

Andersson, H., Garscha, U., Brittebo, E., 2011. Effects of PCB126 and 17ß-oestradiol on endothelium-derived vasoactive factors in human endothelial cells, Toxicology 285, 46-56.

Andersson, P., Haglund, P., Rappe, C., Tysklind, M., 1996. Ultraviolet absorption characteristics and calculated semi-empirical parameters as chemical descriptors in multivariate modelling of polychlorinated biphenyls, J. Chemometrics 10, 171-185. Andersson, P.L., Van Der Burght, A.S.A.M., Van Den Berg, M., Tysklind, M., 2000. Multivariate modeling of polychlorinated biphenyl-induced CYP1A activity in hepatocytes from three different species: Ranking scales and species differences, Environmental Toxicology and Chemistry 19, 1454-1463.

Andreas Moser, G., McLachlan, M.S., 2001. The influence of dietary concentration on the absorption and excretion of persistent lipophilic organic pollutants in the human intestinal tract, Chemosphere 45, 201-211. Arulmozhiraja, S., Morita, M., 2004. Structure-Activity Relationships for the Toxicity of Polychlorinated Dibenzofurans: Approach through Density Functional Theory-Based Descriptors, Chem. Res. Toxicol. 17, 348-356. Aylward, L.L., Collins, J.J., Bodner, K.M., Wilken, M., Bodnar, C.M., 2013. Elimination rates of dioxin congeners in former chlorophenol workers from Midland, Michigan, Environ. Health Perspect. 121, 39-45.

Bandiera, S., Sawyer, T.W., Campbell, M.A., Toshio, F., Safe, S., 1983. Competitive binding to the cytosolic 2,3,7,8-tetrachlorodibenzo-p-dioxin receptor. Effects of structure on the affinities of substituted halogenated biphenyls-A QSAR analysis, Biochem. Pharmacol. 32, 3803-3813. 199

A


Bavithra, S., Selvakumar, K., Pratheepa Kumari, R., Krishnamoorthy, G., Venkataraman, P., Arunakaran, J., 2012. Polychlorinated Biphenyl (PCBs)-induced oxidative stress plays a critical role on cerebellar dopaminergic receptor expression: Ameliorative role of quercetin, Neurotoxicity Research 21, 149-159. Beischlag, T.V., Morales, J.L., Hollingshead, B.D., Perdew, G.H., 2008. The aryl hydrocarbon receptor complex and the control of gene expression, Crit. Rev. Eukaryot. Gene Expr. 18, 207-250.

Berge, G., Ă&#x2DC;vrebø, S., Eilertsen, E., Haugen, A., Mollerup, S., 2004. Analysis of resveratrol as a lung cancer chemopreventive agent in A/J mice exposed to benzo[a]pyrene, Br. J. Cancer 91, 13801383. Berghard, A., Gradin, K., Pongratz, I., Whitelaw, M., Poellinger, L., 1993. Cross-coupling of signal transduction pathways: The dioxin receptor mediates induction of cytochrome P-450IA1 expression via a protein kinase C- dependent mechanism, Molecular and Cellular Biology 13, 677-689.

Bieche, I., Narjoz, C., Asselah, T., Vacher, S., Marcellin, P., Lidereau, R., Beaune, P., de Waziers, I., 2007. Reverse transcriptase-PCR quantification of mRNA levels from cytochrome (CYP)1, CYP2 and CYP3 families in 22 different human tissues, Pharmacogenet Genomics 17, 731-742.

Bilau, M., Matthys, C., Baeyens, W., Bruckers, L., Backer, G.D., Hond, E.D., Keune, H., Koppen, G., Nelen, V., Schoeters, G., Van Larebeke, N., Willems, J.L., De Henauw, S., 2008. Dietary exposure to dioxin-like compounds in three age groups: Results from the Flemish environment and health study, Chemosphere 70, 584-592.

Birnbaum, L.S., 1994. The mechanism of dioxin toxicity: Relationship to risk assessment, Environ Health Perspect 102, 157-167. Birnbaum, L.S., Couture, L.A., 1988. Disposition of octachlorodibenzo-p-dioxin (OCDD) in male rats, Toxicol. Appl. Pharmacol. 93, 22-30.

Birnbaum, L.S., Tuomisto, J., 2000. Non-carcinogenic effects of TCDD in animals, Food Addit Contam 17, 275-288.

Black, M.B., Budinsky, R.A., Dombkowski, A., Cukovic, D., Lecluyse, E.L., Ferguson, S.S., Thomas, R.S., Rowlands, J.C., 2012. Cross-species comparisons of transcriptomic alterations in human and rat primary hepatocytes exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin, Toxicol. Sci. 127, 199215. Breivik, K., Sweetman, A., Pacyna, J.M., Jones, K.C., 2007. Towards a global historical emission inventory for selected PCB congeners - A mass balance approach. 3. An update, Sci. Total Environ. 377, 296-307. Brewster, D.W., Birnbaum, L.S., 1988. Disposition of 1,2,3,7,8-pentachlorodibenzofuran in the rat, Toxicol. Appl. Pharmacol. 95, 490-498.

Budinsky, R.A., LeCluyse, E.L., Ferguson, S.S., Rowlands, J.C., Simon, T., 2010. Human and rat primary hepatocyte CYP1A1 and 1A2 induction with 2,3,7,8-tetrachlorodibenzo-p-dioxin, 2,3,7,8-tetrachlorodibenzofuran, and 2,3,4,7,8-pentachlorodibenzofuran, Toxicol Sci. 118, 224-235.

Budinsky, R.A., Paustenbach, D., Fontaine, D., Landenberger, B., Starr, T.B., 2006. Recommended relative potency factors for 2,3,4,7,8-pentachlorodibenzofuran: The impact of different dose me200


trics, Toxicol Sci 91, 275-285.

References

Budinsky, R.A., Rowlands, J.C., Casteel, S., Fent, G., Cushing, C.A., Newsted, J., Giesy, J.P., Ruby, M.V., Aylward, L.L., 2008. A pilot study of oral bioavailability of dioxins and furans from contaminated soils: Impact of differential hepatic enzyme activity and species differences, Chemosphere 70, 1774-1786. Carlson, E.A., Mcculloch, C., Koganti, A., Goodwin, S.B., Sutter, T.R., Silkworth, J.B., 2009. Divergent transcriptomic responses to aryl hydrocarbon receptor agonists between rat and human primary hepatocytes, Toxicol. Sci. 112, 257-272.

Charnley, G., Kimbrough, R.D., 2006. Overview of exposure, toxicity, and risks to children from current levels of 2,3,7,8-tetrachlorodibenzo-p-dioxin and related compounds in the USA, Food and Chemical Toxicology 44, 601-615. Chen, C.Y., Hamm, J.T., Hass, J.R., Birnbaum, L.S., 2001. Disposition of polychlorinated dibenzo-p-dioxins, dibenzofurans, and non-ortho polychlorinated biphenyls in pregnant long evans rats and the transfer to offspring, Toxicol Appl Pharmacol 173, 65-88. Chen, P.H., Wong, C.K., Rappe, C., Nygren, M., 1985. Polychlorinated biphenyls, dibenzofurans and quaterphenyls in toxic rice-bran oil and in the blood and tissues of patients with PCB poisoning (Yu-Cheng) in Taiwan, Environ. Health Perspect. VOL, 59-65.

Cheney, B.V., Tolly, T., 1979. Electronic factors affecting receptor binding of dibenzo-p-dioxins and dibenzofurans. Int.  J.  Quantum Chem. 16, 87-110.

Chu, I., Villeneuve, D.C., Yagminas, A., Lecavalier, P., Poon, R., Feeley, M., Kennedy, S.W., Seegal, R.F., Häkansson, H., Ahlborg, U.G., Valli, V.E., 1994. Subchronic toxicity of 3,3′,4,4′,5-pentachlorobiphenyl in the rat: I. Clinical, biochemical, hematological, and histopathological changes, Toxicol Sci 22, 457-468.

Cleverly, D., Ferrario, J., Byrne, C., Riggs, K., Joseph, D., Hartford, P., 2007. A general indication of the contemporary background levels of PCDDs, PCDFs, and coplanar PCBs in the ambient air over rural and remote areas of the United States, Environmental Science and Technology 41, 15371544. Connor, K.T., Aylward, L.L., 2006. Human response to dioxin: Aryl hydrocarbon receptor (AhR) molecular structure, function, and dose-response data for enzyme induction indicate an impaired human AhR, J. Toxicol. Environ. Health Part B 9, 147-171. Consonni, D., Sindaco, R., Bertazzi, P.A., 2012. Blood levels of dioxins, furans, dioxin-like PCBs, and TEQs in general populations: A review, 1989-2010, Environ. Int. 44, 151-162.

Czuczwa, J.M., Hites, R.A., 1986. Airborne dioxins and dibenzofurans: Sources and fates, Environmental Science and Technology 20, 195-200. Czuczwa, J.M., Hites, R.A., 1984. Environmental fate of combustion-generated polychlorinated dioxins and furans, Environmental Science and Technology 18, 444-450.

De Mul, A., Bakker, M.I., Zeilmaker, M.J., Traag, W.A., Leeuwen, S.P.J.v., Hoogenboom, R.L.A.P., Boon, P.E., Klaveren, J.D.v., 2008. Dietary exposure to dioxins and dioxin-like PCBs in The Netherlands anno 2004, Regul Toxicol Pharmacol 51, 278-287. 201

A


Denison, M.S., Heath-Pagliuso, S., 1998. The Ah receptor: A regulator of the biochemical and toxicological actions of structurally diverse chemicals, Bull Environ Contam Toxicol 61, 557-568.

Denison, M.S., Nagy, S.R., 2003. Activation of the Aryl Hydrocarbon Receptor by Structurally Diverse Exogenous and Endogenous Chemicals, 43, 309-334.

Denison, M.S., Soshilov, A.A., He, G., Degroot, D.E., Zhao, B., 2011. Exactly the same but different: Promiscuity and diversity in the molecular mechanisms of action of the aryl hydrocarbon (dioxin) receptor, Toxicol. Sci. 124, 1-22.

Denison, M.S., Zhao, B., Baston, D.S., Clark, G.C., Murata, H., Han, D., 2004. Recombinant cell bioassay systems for the detection and relative quantitation of halogenated dioxins and related chemicals, Talanta 63, 1123-1133.

Devito, M.J., Birnbaum, L.S., 1995. The importance of pharmacokinetics in determining the relative potency of 2,3,7,8-tetrachlorodibenzo-p-dioxin and 2,3,7,8-tetrachlorodibenzofuran, Toxicol Sci 24, 145-148.

DeVito, M.J., Diliberto, J.J., Ross, D.G., Menache, M.G., Birnbaum, L.S., 1997. Dose-response relationships for polyhalogenated dioxins and dibenzofurans following subchronic treatment in mice: I. CYP1A1 and CYP1A2 enzyme activity in liver, lung, and skin, Toxicol. Appl. Pharmacol. 147, 267280. DeVito, M.J., Menache, M.G., Diliberto, J.J., Ross, D.G., Birnbaum, L.S., 2000. Dose-response relationships for induction of CYP1A1 and CYP1A2 enzyme activity in liver, lung, and skin in female mice following subchronic exposure to polychlorinated biphenyls, Toxicol Appl Pharmacol 167, 157-172.

Devito, M.J., Ross, D.G., Dupuy A.E., J., Ferrario, J., McDaniel, D., Birnbaum, L.S., 1998. Dose-response relationships for disposition and hepatic sequestration of polyhalogenated dibenzo-p-dioxins, dibenzofurans, and biphenyls following subchronic treatment in mice, Toxicol Sci 46, 223234. Diao, J., Li, Y., Shi, S., Sun, Y., Sun, Y., 2010. QSAR models for predicting toxicity of polychlorinated dibenzo-p-dioxins and dibenzofurans using quantum chemical descriptors, Bull. Environ. Contam. Toxicol. 85, 109-115.

Diliberto, J.J., Akubue, P.I., Luebke, R.W., Birnbaum, L.S., 1995. Dose-response relationships of tissue distribution and induction of CYP1A1 and CYP1A2 enzymatic activities following acute exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in mice, Toxicol Appl Pharmacol 130, 197-208. Diliberto, J.J., Burgin, D., Birnbaum, L.S., 1997. Role of CYP1A2 in hepatic sequestration of dioxin: Studies using CYP1A2 knock-out mice, Biochem. Biophys. Res. Commun. 236, 431-433.

Diliberto, J.J., Burgin, D.E., Birnbaum, L.S., 1999. Effects of CYP1A2 on disposition of 2,3,7,8-tetrachlorodibenzo-p-dioxin, 2,3,4,7,8-pentachlorodibenzofuran, and 2,2’,4,4’,5,5’-hexachlorobiphenyl in CYP1A2 knockout and parental (C57BL/6N and 129/Sv) strains of mice, Toxicol. Appl. Pharmacol. 159, 52-64. Diliberto, J.J., DeVito, M.J., Ross, D.G., Birnbaum, L.S., 2001. Subchronic Exposure of [3H]- 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in female B6C3F1 mice: relationship of steady-state levels to

202


disposition and metabolism, Toxicol. Sci. 61, 241-255.

References

Drahushuk, A.T., McGarrigle, B.P., Tai, H.-., Kitareewan, S., Goldstein, J.A., Olson, J.R., 1996. Validation of precision-cut liver slices in dynamic organ culture as an in vitro model for studying CYP1A1 and CYP1A2 induction, Toxicol. Appl. Pharmacol. 140, 393-403. Drenth, H.J., Bouwman, C.A., Seinen, W., Van Den Berg, M., 1998. Effects of some persistent halogenated environmental contaminants on aromatase (CYP19) activity in the human choriocarcinoma cell line JEG-3, Toxicology and Applied Pharmacology 148, 50-55. Dynes, J.D., Gottschalk, K.E., Pedersen, L.G., 1985. Interring dihedral angles in polychlorinated biphenyls from photoelectron spectroscopy. Can.  J.  Chem. 63, 1292-1299.

ECSCF, 2001. Opinion of the Scientific Committee on Food on the Risk Assessment of Dioxins and Dioxin-like PCBs in Food. CS/CNTM/DIOXIN/20 final. Brussels, Belgium: European Commission.  http://ec.europa.eu/food/fs/sc/scf/out90_en.pdf. EFSA, 2012. Update of the monitoring of levels of dioxins and PCBs in food and feed, EFSA Journal 10, 2832 [82 pp.]-Available online: www.efsa.europa.eu/en/efsajournal/pub/2832.htm#.

Ema, M., Ohe, N., Suzuki, M., Mimura, J., Sogawa, K., Ikawa, S., Fujii-Kuriyama, Y., 1994. Dioxin binding activities of polymorphic forms of mouse and human arylhydrocarbon receptors, J. Biol. Chem. 269, 27337-27343.

Endo, F., Monsees, T.K., Akaza, H., Schill, W.-., Pflieger-Bruss, S., 2003. Effects of single non-ortho, mono-ortho, and di-ortho chlorinated biphenyls on cell functions and proliferation of the human prostatic carcinoma cell line, LNCaP, Reprod.  Toxicol. 17, 229-236. doi: 10.1016/S08906238(02)00126-0.

Eriksson, L., Rosén, J., Johansson, E., Trygg, J., 2012. Orthogonal PLS (OPLS) modeling for improved analysis and interpretation in drug design, Molecular Informatics 31, 414-419. Eubig, P.A., Aguiar, A., Schantz, S.L., 2010. Lead and PCBs as risk factors for attention defcit/hyperactivity disorder, Environ. Health Perspect. 118, 1654-1667.

Ferriby, L.L., Knutsen, J.S., Harris, M., Unice, K.M., Scott, P., Nony, P., Haws, L.C., Paustenbach, D., 2007. Evaluation of PCDD/F and dioxin-like PCB serum concentration data from the 2001-2002 National Health and Nutrition Examination Survey of the United States population, Journal of Exposure Science and Environmental Epidemiology 17, 358-371. Fiedler, H., 1996. Sources of PCDD/PCDF and impact on the environment, Chemosphere 32, 55-64. Field, L.D., Skelton, B.W., Sternhell, S., White, A.H., Aust, S., 1985. Structural studies of some 2,6 (2’,6’)- halogen substituted biphenyl derivatives. J.  Chem. 38, 391-399.

Fisher, J.W., Campbell, J., Muralidhara, S., Bruckner, J.V., Ferguson, D., Mumtaz, M., Harmon, B., Hedge, J.M., Crofton, K.M., Kim, H., Almekinder, T.L., 2006. Effect of PCB 126 on hepatic metabolism of thyroxine and perturbations in the hypothalamic-pituitary-thyroid axis in the rat, Toxicological Sciences 90, 87-95.

Flaveny, C.A., Murray, I.A., Perdew, G.H., 2010. Differential gene regulation by the human and mouse aryl hydrocarbon receptor, Toxicol Sci 114, 217-225. 203

A


Flesch-Janys, D., Becher, H., Gurn, P., Jung, D., Konietzko, J., Manz, A., Papke, O., 1996. Elimination of polychlorinated dibenzo-p-dioxins and dibenzofurans in occupationally exposed persons, J. Toxicol. Environ. Health 47, 363-378. Fürst, P., 2006. Dioxins, polychlorinated biphenyls and other organohalogen compounds in human milk, Mol Nutr Food Res 50, 922-933.

Garrison, P.M., Tullis, K., Aarts, J.M.M.J.G., Brouwer, A., Giesy, J.P., Denison, M.S., 1996. Species-specific recombinant cell lines as bioassay systems for the detection of 2,3,7,8-tetrachlorodibenzo-p-dioxin-like chemicals, Toxicol Sci 30, 194-203.

Gaussian 09 Revision A.1, 2009. Gaussian, Inc., Wallingford CT, 2009; software available at; http://www.gaussian.com. Golbraikh, A., Tropsha, A., 2002. Beware of q2! J. Mol. Graph. Model. 20, 269-276.

Guida, M., Marra, M., Zullo, F., Guida, M., Trifuoggi, M., Biffali, E., Borra, M., De Mieri, G., D’Alessandro, R., De Felice, B., 2013. Association between exposure to dioxin-like polychlorinated biphenyls and miR-191 expression in human peripheral blood mononuclear cells, Mutation Research/Genetic Toxicology and Environmental Mutagenesis 753, 36-41.

Hafelinger, G., Regelman, C., 1985. Ab initio STO-3G optimization of planar, perpendicular and twisted molecular structures of biphenyl. J.  Comp.  Chem. 6, 368-376.

Hahn, M.E., Chandran, K., 1996. Uroporphyrin accumulation associated with cytochrome P4501A induction in fish hepatoma cells exposed to aryl hydrocarbon receptor agonists, including 2,3,7,8-tetrachlorodibenzo-p-dioxin and planar chlorobiphenyls, Arch. Biochem. Biophys. 329, 163-174. Hamm, J.T., Chen, C.Y., Birnbaum, L.S., 2003. A mixture of dioxins, furans, and non-ortho PCBs based upon consensus toxic equivalency factors produces dioxin-like reproductive effects, Toxicol Sci 74, 182-191.

Han, D., Nagy, S.R., Denison, M.S., 2004. Comparison of recombinant cell bioassays for the detection of Ah receptor agonists, BioFactors 20, 11-22.

Hanaoka, T., Yamano, Y., Pan, G., Hara, K., Ichiba, M., Zhang, J., Zhang, S., Liu, T., Li, L., Takahashi, K., Kagawa, J., Tsugane, S., 2002. Cytochrome P450 1B1 mRNA levels in peripheral blood cells and exposure to polycyclic aromatic hydrocarbons in Chinese coke oven workers, Sci. Total Environ. 296, 27-33.

Hankinson, O., 1995a. The aryl hydrocarbon receptor complex, Annu. Rev. Pharmacol. Toxicol. 35, 307-340.

Hankinson, O., 1995b. The aryl hydrocarbon receptor complex, Annual Review of Pharmacology and Toxicology 35, 307-340.

Harju, M., Hamers, T., Kamstra, J.H., Sonneveld, E., Boon, J.P., Tysklind, M., Andersson, P.L., 2007. Quantitative structure-activity relationship modeling on in vitro endocrine effects and metabolic stability involving 26 selected brominated flame retardants, Environmental Toxicology and Chemistry 26, 816-826. 204


References Haws, L.C., Su, S.H., Harris, M., DeVito, M.J., Walker, N.J., Farland, W.H., Finley, B., Birnbaum, L.S., 2006. Development of a refined database of mammalian relative potency estimates for dioxin-like compounds, Toxicol Sci 89, 4-30. Hays, S.M., Aylward, L.L., 2003. Dioxin risks in perspective: Past, present, and future, Regulatory Toxicology and Pharmacology 37, 202-217.

Hestermann, E.V., Stegeman, J.J., Hahn, M.E., 2000. Serum alters the uptake and relative potencies of halogenated aromatic hydrocarbons in cell culture bioassays, Toxicol. Sci. 53, 316-325. Hoogenboom, L.A.P., Van Eijkeren, J.C.H., Zeilmaker, M.J., Mengelers, M.J.B., Herbes, R., Immerzeel, J., Traag, W.A., 2007. A novel source for dioxins present in recycled fat from gelatin production, Chemosphere 68, 814-823.

Hoogenboom, R., Bovee, T., Portier, L., Bor, G., Van Der Weg, G., Onstenk, C., Traag, W., 2004. The German bakery waste incident; Use of a combined approach of screening and confirmation for dioxins in feed and food, Talanta 63, 1249-1253.

Hord, N.G., Perdew, G.H., 1994. Physicochemical and immunocytochemical analysis of the aryl hydrocarbon receptor nuclear translocator: Characterization of two monoclonal antibodies to the aryl hydrocarbon receptor nuclear translocator, Mol. Pharmacol. 46, 618-626.

Hsu, J.-., Guo, Y.-., Yang, S.-., Liao, P.-., 2005. Congener profiles of PCBs and PCDD/Fs in Yucheng victims fifteen years after exposure to toxic rice-bran oils and their implications for epidemiologic studies, Chemosphere 61, 1231-1243. Hu, S., Chen, C., Kuo, C., Lin, W., Lin, P., 2006. Increased cytochrome P4501B1 gene expression in peripheral leukocytes of municipal waste incinerator workers, Toxicol. Lett. 160, 112-120.

IARC, 2012. A review of human carcinogens. Part F: Chemical agents and related occupations. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans (2009: Lyon, France).  IARC monographs on the evaluation of carcinogenic risks to humans ; v. 100F. , http://monographs.iarc.fr/ENG/Monographs/vol100F/mono100F.pdf. Iida, T., Hirakawa, H., Matsueda, T., Nagayama, J., Nagata, T., 1999. Polychlorinated dibenzo-p-dioxins and related compounds: Correlations of levels in human tissues and in blood, Chemosphere 38, 2767-2774. JECFA, 2001. Joint FAO/WHO Expert Committee on Food Additives, Fifty-Seventh Meeting, 5–14 June 2001, Rome. Summary and<br />Conclusions.  Available: http://www.who.int/foodsafety/ chem/jecfa/summaries/en/summary_57.pdf. Kang, H.K., Watanabe, K.K., Breen, J., Remmers, J.A., Stanley, J., Flicker, M., 1990. Dioxins and Dibenzofurans in Adipose Tissue of U.S. Vietnam Veterans and Controls. NTIS PB-91167585. Washington, DC:U.S. Department of VeteransAffairs/U.S. Environmental Protection Agency. Kimbrough, R.D., Falk, H., Stehr, P., Fries, G., 1984. Health implications of 2,3,7,8-tetrachloro-dibenzodioxin (TCDD) contamination of residential soil, J. Toxicol. Environ. Health 14, 47-93.

Kobayashi, S., Saito, A., Ishii, Y., Tanaka, A., Tobinaga, S., 1991. Relationship between the biological potency of polychlorinated dibenzo-p-dioxins and their electronic states, Chemical and Pharmaceutical Bulletin 39, 2100-2105. 205

A


Lambert, G.H., Needham, L.L., Turner, W., Lai, T.J., Patterson Jr., D.G., Guo, Y.L., 2006. Induced CYP1A2 activity as a phenotypic biomarker in humans highly exposed to certain PCBs/PCDFs, Environmental Science and Technology 40, 6176-6180. Larsson, M., Kumar Mishra, B., Tysklind, M., Linusson, A., Andersson, P.L., 2013. On the use of electronic descriptors for QSAR modelling of PCDDs, PCDFs and dioxin-like PCBs£, SAR QSAR Environ. Res. 24, 727-745. Lauby-Secretan, B., Loomis, D., Grosse, Y., El Ghissassi, F., Bouvard, V., Benbrahim-Tallaa, L., Guha, N., Baan, R., Mattock, H., Straif, K., 2013. Carcinogenicity of polychlorinated biphenyls and polybrominated biphenyls, Lancet Oncol. 14, 287-288.

Li, F., Li, X., Liu, X., Zhang, L., You, L., Zhao, J., Wu, H., 2011. Docking and 3D-QSAR studies on the Ah receptor binding affinities of polychlorinated biphenyls (PCBs), dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs), Environ. Toxicol. Pharmacol. 32, 478-485.

Li, J., Zhang, L., Wu, Y., Liu, Y., Zhou, P., Wen, S., Liu, J., Zhao, Y., Li, X., 2009. A national survey of polychlorinated dioxins, furans (PCDD/Fs) and dioxin-like polychlorinated biphenyls (dl-PCBs) in human milk in China, Chemosphere 75, 1236-1242.

Liem, A.K.D., Fürst, P., Rappe, C., 2000. Exposure of populations to dioxins and related compounds, Food Addit. Contam. 17, 241-259. Llobet, J.M., Martí-Cid, R., Castell, V., Domingo, J.L., 2008. Significant decreasing trend in human dietary exposure to PCDD/PCDFs and PCBs in Catalonia, Spain, Toxicol. Lett. 178, 117-126.

Lonard, D.M., O’Malley, B.W., 2007. Nuclear Receptor Coregulators: Judges, Juries, and Executioners of Cellular Regulation, Mol. Cell 27, 691-700.

Loutfy, N., Fuerhacker, M., Tundo, P., Raccanelli, S., El Dien, A.G., Ahmed, M.T., 2006. Dietary intake of dioxins and dioxin-like PCBs, due to the consumption of dairy products, fish/seafood and meat from Ismailia city, Egypt, Sci. Total Environ. 370, 1-8.

McHale, C.M., Zhang, L., Hubbard, A.E., Zhao, X., Baccarelli, A., Pesatori, A.C., Smith, M.T., Landi, M.T., 2007. Microarray analysis of gene expression in peripheral blood mononuclear cells from dioxin-exposed human subjects, Toxicology 229, 101-113. McKinney, J.D., Singh, P., 1981. Structure-activity relationships in halogenated biphenyls: Unifying hypothesis for structural specificity, Chem. Biol. Interact. 33, 271-283. Mekenyan, O.G., Veith, G.D., Call, D.J., Ankley, G.T., 1996. A QSAR evaluation of Ah receptor binding of halogenated aromatic xenobiotics, Environ. Health Perspect. 104, 1302-1310.

Michalek, J.E., Wolfe, W.H., Miner, J.C., 1990. Health status of Air Force veterans occupationally exposed to herbicides in Vietnam. II. Mortality, J. Am. Med. Assoc. 264, 1832-1836.

Micka, J., Milatovich, A., Menon, A., Grabowski, G.A., Puga, A., Nebert, D.W., 1997. Human Ah receptor (AHR) gene: Localization to 7p15 and suggestive correlation of polymorphism with CYP1A1 inducibility, Pharmacogenetics 7, 95-101.

Mocarelli, P., Pocchiari, F., Nelson, N., 1988. Preliminary report: 2,3,7,8-tetrachlorodibenzo-p-dioxin exposure to humans Seveso , Italy. Morb Mortal Wkly Rep 48, 733–736.

206


References

MOE 2006.08., 2008. Chemical Computing Group, Quebec, Canada. Software available at: http:// www.chemcomp.com.

Moriguchi, T., Motohashi, H., Hosoya, T., Nakajima, O., Takahashi, S., Ohsako, S., Aoki, Y., Nishimura, N., Tohyama, C., Fujii-Kuriyama, Y., Yamamoto, M., 2003. Distinct response to dioxin in an arylhydrocarbon receptor (AHR)-humanized mouse, Proceedings of the National Academy of Sciences of the United States of America 100, 5652-5657. Muzio, G., Maggiora, M., Paiuzzi, E., Oraldi, M., Canuto, R.A., 2012. Aldehyde dehydrogenases and cell proliferation, Free Radical Biology and Medicine 52, 735-746.

Nagayama, J., Kiyohara, C., Masuda, Y., Kuratsune, M., 1985. Inducing Potency of Aryl Hydrocarbon Hydroxylase Activity in Human Lymphoblastoid Cells and Mice by Polychlorinated Dibenzofura Congeners, Environ Health Perspect 59, 107-112.

National Toxicology Program, 2010. Toxicology and carcinogenesis studies of 2,3’,4,4’,5-pentachlorobiphenyl (PCB 118) (CAS No. 31508-00-6) in female harlan Sprague-Dawley rats (gavage studies), Natl. Toxicol. Program. Tech. Rep. Ser. (559), 1-174.

National Toxicology Program, 2006a. NTP technical report on the toxicology and carcinogenesis studies of 2,2’,4,4’,5,5’-hexachlorobiphenyl (PCB 153) (CAS No. 35065-27-1) in female Harlan Sprague-Dawley rats (Gavage studies), Natl. Toxicol. Program. Tech. Rep. Ser. (529), 4-168. National Toxicology Program, 2006b. NTP technical report on the toxicology and carcinogenesis studies of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (CAS No. 1746-01-6) in female Harlan Sprague-Dawley rats (Gavage Studies), Natl. Toxicol. Program. Tech. Rep. Ser. (521), 4-232. National Toxicology Program, 2006c. NTP toxicology and carcinogenesis studies of 3,3’,4,4’,5-pentachlorobiphenyl (PCB 126) (CAS No. 57465-28-8) in female Harlan Sprague-Dawley rats (Gavage Studies), Natl. Toxicol. Program. Tech. Rep. Ser. (520), 4-246.

National Toxicology Program, 2006d. Toxicology and carcinogenesis studies of 2,3,4,7,8-pentachlorodibenzofuran (PeCDF) (Cas No. 57117-31-4) in female Harlan Sprague-Dawley rats (gavage studies), Natl. Toxicol. Program. Tech. Rep. Ser. (525), 1-198. Nguyen, L.P., Bradfield, C.A., 2008. The search for endogenous activators of the aryl hydrocarbon receptor, Chem. Res. Toxicol. 21, 102-116.

Niu, J., Huang, L., Chen, J., Yu, G., Schramm, K.-., 2005. Quantitative structure-property relationships on photolysis of PCDD/Fs adsorbed to spruce (Picea abies (L.) Karst.) needle surfaces under sunlight irradiation, Chemosphere 58, 917-924.

Novotna, A., Pavek, P., Dvorak, Z., 2011. Novel stably transfected gene reporter human hepatoma cell line for assessment of aryl hydrocarbon receptor transcriptional activity: Construction and characterization, Environmental Science and Technology 45, 10133-10139.

OECD, Quantitative Structure-Activity Relationships Project [(Q)SARs]. Available online: http:// www.oecd.org/document/23/0,3746,en_2649_34377_33957015_1_1_1_1,00.html.

Okey, A.B., Riddick, D.S., Harper, P.A., 1994. The Ah receptor: Mediator of the toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and related compounds, Toxicology Letters 70, 1-22. 207

A


Olafsson, P.G., Bryan, A.M., Stone, W., 1988. Polychlorinated biphenyls and polychlorinated dibenzofurans in the tissues of patients with Yusho or Yu-Chen: Total toxicity, Bull. Environ. Contam. Toxicol. 41, 63-70.

Pang, S., Cao, J.Q., Katz, B.H., Sutter, T.R., Spink, D.C., 1999. Inductive and inhibitory effects of non-ortho-substituted polychlorinated biphenyls on estrogen metabolism and human cytochromes P450 1A1 and 1B1, Biochem. Pharmacol. 58, 29-38. Parvez, S., Evans, A.M., Lorber, M., Hawkins, B.S., Swartout, J.C., Teuschler, L.K., Rice, G.E., 2013. A sensitivity analysis using alternative toxic equivalency factors to estimate U.S. dietary exposures to dioxin-like compounds, Regulatory Toxicology and Pharmacology 67, 278-284.

Peters, A.K., Leonards, P.E., Zhao, B., Bergman, A., Denison, M.S., Van den Berg, M., 2006. Determination of in vitro relative potency (REP) values for mono-ortho polychlorinated biphenyls after purification with active charcoal, Toxicology Letters 165, 230-241. Petkov, P.I., Rowlands, J.C., Budinsky, R., Zhao, B., Denison, M.S., Mekenyan, O., 2010. Mechanism-based common reactivity pattern (COREPA) modelling of aryl hydrocarbon receptor binding affinity, SAR QSAR Environ. Res. 21, 187-214. Poland, A., Teitelbaum, P., Glover, E., Kende, A., 1989. Stimulation of in vivo hepatic uptake and in vitro hepatic binding of [125I]2-iodo-3,7,8-trichlorodibenzo-p-dioxin by the administration of agonists for the Ah receptor, Mol. Pharmacol. 36, 121-127.

Pollenz, R.S., Sattler, C.A., Poland, A., 1994. The aryl hydrocarbon receptor and aryl hydrocarbon receptor nuclear translocator protein show distinct subcellular localizations in Hepa 1c1c7 cells by immunfluorescence microscopy, Mol. Pharmacol. 45, 428-438.

Rawn, D.F.K., Ryan, J.J., Sadler, A.R., Sun, W.-., Haines, D., Macey, K., Van Oostdam, J., 2012. PCDD/F and PCB concentrations in sera from the Canadian Health Measures Survey (CHMS) from 2007 to 2009, Environ. Int. 47, 48-55. Ritz, C., 2010. Toward a unified approach to dose-response modeling in ecotoxicology, Environmental Toxicology and Chemistry 29, 220-229. Rohde, S., Moser, G.A., P채pke, O., McLachlan, M.S., 1999. Clearance of PCDD/Fs via the gastrointestinal tract in occupationally exposed persons, Chemosphere 38, 3397-3410.

Rowlands, J.C., Budinsky, R., Gollapudi, B., Black, M.B., Wolfinger, R.D., Cukovic, D., Dombkowski, A., Thompson, C.M., Urban, J.D., Thomas, R.S., 2013. A genomics-based analysis of relative potencies of dioxin-like compounds in primary rat hepatocytes, Toxicological Sciences 136, 595-604.

Safe, S., 1993. Toxicology, structure-function relationship, and human and environmental health impacts of polychlorinated biphenyls: Progress and problems, Environ. Health Perspect. 100, 259268.

Safe, S., 1990. Polychlorinated biphenyls (PCBs), dibenzo-p-dioxins (PCDDs), dibenzofurans (PCDFs), and related compounds: Environmental and mechanistic considerations which support the development of toxic equivalency factors (TEFs), Crit. Rev. Toxicol. 21, 51-88. Safe, S., Bandiera, S., Sawyer, T., Robertson, L., Safe, L., Parkinson, A., Thomas, P.E., Ryan, D.E., Reik, L.M., Levin, W., 1985a. PCBs: structure-function relationships and mechanism of action. Environ.

208


References Health Perspect. 60, 47-56.

Safe, S., Bandiera, S., Sawyer, T., Zmudzka, B., Mason, G., Romkes, M., Denomme, M.A., Sparling, J., Okey, A.B., Fujita, T., 1985b. Effects of structure on binding to the 2,3,7,8-TCDD receptor protein and AHH induction--halogenated biphenyls. Environ. Health Perspect. 61, 21-33. Safe, S.H., 1994a. Polychlorinated biphenyls (PCBs): Environmental impact, biochemical and toxic responses, and implications for risk assessment, Crit Rev Toxicol 24, 87-149. Safe, S.H., 1994b. Polychlorinated biphenyls (PCBs): Environmental impact, biochemical and toxic responses, and implications for risk assessment, Crit. Rev. Toxicol. 24, 87-149.

Safe, S.H., 1986. Comparative toxicology and mechanism of action of polychlorinated dibenzo-p-dioxins and dibenzofurans, Annual Review of Pharmacology and Toxicology VOL. 26, 371399. Santostefano, M.J., Ross, D.G., Savas, U., Jefcoate, C.R., Birnbaum, L.S., 1997. Differential time-course and dose-response relationships of TCDD-induced CYP1B1, CYP1A1, and CYP1A2 proteins in rats, Biochem Biophys Res Commun 233, 20-24.

Sawyer, T., Safe, S., 1982. PCB isomers and congeners: Induction of aryl hydrocarbon hydroxylase and ethoxyresorufin O-deethylase enzyme activities in rat hepatoma cells, Toxicol. Lett. 13, 87-93. Schantz, S.L., Gasior, D.M., Polverejan, E., McCaffrey, R.J., Sweeney, A.M., Humphrey, H.E.B., Gardiner, J.C., 2001. Impairments of memory and learning in older adults exposed to polychlorinated biphenyls via consumption of Great Lakes Fish, Environ. Health Perspect. 109, 605-611.

Schecter, A., Constable, J.D., Bangert, J.V., Tong, H., Arghestani, S., Monson, S., Gross, M., 1989. Elevated body burdens of 2,3,7,8-tetrachlorodibenzodioxin in adipose tissue of United States Vietnam veterans, Chemosphere 18, 431-438.

Schecter, A., Kassis, I., Päpke, O., 1998. Partitioning of dioxins, dibenzofurans, and coplanar PCBs in blood, milk, adipose tissue, placenta and cord blood from five American women, Chemosphere 37, 1817-1823. Schecter, A., Fürst, P., Fürst, C., Päpke, O., Ball, M., Dai, L.C., Quynh, H.T., Phoung, N.T.N., Beim, A., Vlasov, B., Chongchet, V., Constable, J.D., Charles, K., 1991. Dioxins, dibenzofurans and selected chlorinated organic compounds in human milk and blood from Cambodia, Germany, Thailand, the U.S.A., the U.S.S.R., and Vietnam, Chemosphere 23, 1903-1912.

Schrenk, D., Stuven, T., Gohl, G., Viebahn, R., Bock, K.W., 1995. Induction of CYP1A and glutathione S-transferase activities by 2,3,7,8-tetrachlorodibenzo-p-dioxin in human hepatocyte cultures, Carcinogenesis 16, 943-946. Schulz, V.J., Smit, J.J., Huijgen, V., Bol-Schoenmakers, M., Van Roest, M., Kruijssen, L.J.W., Fiechter, D., Hassing, I., Bleumink, R., Safe, S., Van Duursen, M.B.M., Van den berg, M., Pieters, R.H.H., 2012. Non-dioxin-like AhR ligands in a mouse peanut allergy model, Toxicol Sci 128, 92-102. Scigress Version 2.2.0., 2008. Fujitsu Limited, Tokyo, Japan. Software available at; http://www. fqs.pl/chemistry_materials_life_science/products/scigress.

Sebaugh, J.L., 2011. Guidelines for accurate EC50/IC50 estimation, Pharmaceutical Statistics 10, 209

A


128-134.

Sewall, C.H., Lucier, G.W., 1995. Receptor-mediated events and the evaluation of the Environmental Protection Agency (EPA) of dioxin risks, Mutation Research - Fundamental and Molecular Mechanisms of Mutagenesis 333, 111-122.

Silkworth, J.B., Koganti, A., Illouz, K., Possolo, A., Zhao, M., Hamilton, S.B., 2005. Comparison of TCDD and PCB CYP1A induction sensitivities in fresh hepatocytes from human donors, sprague-dawley rats, and rhesus monkeys and HepG2 cells, Toxicological Sciences 87, 508-519. Sorg, O., Zennegg, M., Schmid, P., Fedosyuk, R., Valikhnovskyi, R., Gaide, O., Kniazevych, V., Saurat, J.-., 2009. 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) poisoning in Victor Yushchenko: identification and measurement of TCDD metabolites, The Lancet 374, 1179-1185.

Spink, B.C., Pang, S., Pentecost, B.T., Spink, D.C., 2002. Induction of cytochrome P450 1B1 in MDAMB-231 human breast cancer cells by non-ortho-substituted polychlorinated biphenyls, Toxicology in Vitro 16, 695-704.

Sutter, C.H., Bodreddigari, S., Sutter, T.R., Carlson, E.A., Silkworth, J.B., 2010. Analysis of the CYP1a1 mRNA dose-response in human keratinocytes indicates that relative potencies of dioxins, furans, and PCBs are species and congener specific, Toxicol. Sci. 118, 704-715. Tard, A., Gallotti, S., Leblanc, J.-., Volatier, J.-., 2007. Dioxins, furans and dioxin-like PCBs: Occurrence in food and dietary intake in France, Food Addit. Contam. 24, 1007-1017. Thoma, H., Mücke, W., Kauert, G., 1990. Comparison of the polychlorinated dibenzo-p-dioxin and dibenzofuran in human tissue and human liver, Chemosphere 20, 433-442.

Toyoshiba, H., Walker, N.J., Bailer, A.J., Portier, C.J., 2004. Evaluation of toxic equivalency factors for induction of cytochromes P450 CYP1A1 and CYP1A2 enzyme activity by dioxin-like compounds, Toxicology and Applied Pharmacology 194, 156-168. Trnovec, T., Jusko, T.A., Šovcíková, E., Lancz, K., Chovancová, J., Patayová, H., Palkovicová, L., Drobná, B., Langer, P., Van den Berg, M., Dedik, L., Wimmerová, S., 2013. Relative effect potency estimates of dioxin-like activity for dioxins, furans, and dioxin-like PCBs in adults based on two thyroid outcomes, Environ. Health Perspect. 121, 886-892.

Trygg, J., Wold, S., 2002. Orthogonal projections to latent structures (O-PLS), J. Chemometrics 16, 119-128. Tsuzuki, S., Tanabe, K., Nagawa, Y., Nakanishi, H., Osawa, E., 1988. Calculations of structures of biphenyl and alkylbiphenyls by molecular mechanics, J. Mol. Struct. 178, 277-285.

Tuppurainen, K., Ruuskanen, J., 2000. Electronic eigenvalue (EEVA): A new QSAR/QSPR descriptor for electronic substituent effects based on molecular orbital energies. A QSAR approach to the Ah receptor binding affinity of polychlorinated biphenyls (PCBs), dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs), Chemosphere 41, 843-848. UKCOT, 2001. Statement on the Tolerable Daily Intake for Dioxins and Dioxin-like Polychlorinated Biphenyls. http://cot.food.gov.uk/pdfs/cot-diox-full.pdf.

USEPA, 2012. EPA’s Reanalysis of Key Issues Related to Dioxin Toxicity and Response to NAS Com210


References ments, Volume 1 (CAS No. 1746-01-6). In Support of Summary Information on the Integrated Risk Information System (IRIS). Washington, DC.   EPA/600/R-10/038F. , http://www.epa.gov/iris/ supdocs/1024index.html.

van Birgelen, A.P., DeVito, M.J., Akins, J.M., Ross, D.G., Diliberto, J.J., Birnbaum, L.S., 1996. Relative potencies of polychlorinated dibenzo-p-dioxins, dibenzofurans, and biphenyls derived from hepatic porphyrin accumulation in mice, Toxicol. Appl. Pharmacol. 138, 98-109.

Van Birgelen, A.P.J.M., Fase, K.M., Van der Kolk, J., Poiger, H., Brouwer, A., Seinen, W., Van den Berg, M., 1996. Synergistic effect 2,2’,4,4’,5,5’-hexachlorobiphenyl and 2,3,7,8-tetrachlorodibenzo-p-dioxin on hepatic porphyrin levels in the rat, Environ Health Perspect 104, 550-557.

Van Birgelen, A.P.J.M., Smit, E.A., Kampen, I.M., Groeneveld, C.M., Fase, K.M., Van Der Kolk, J., Poiger, H., Van Den Berg, M., Koeman, J.H., Brouwer, A., 1995a. Subchronic effects of 2,3,7,8-TCDD or PCBs on thyroid hormone metabolism: Use in risk assessment, European Journal of Pharmacology - Environmental Toxicology and Pharmacology Section 293, 77-85. Van Birgelen, A.P.J.M., Van der Kolk, J., Fase, K.M., Bol, I., Poiger, H., Brouwer, A., Van den Berg, M., 1995b. Subchronic dose-response study of 2,3,7,8-tetrachlorodibenzo-p-dioxin in female Sprague-Dawley rats, Toxicology and Applied Pharmacology 132, 1-13.

Van Birgelen, A.P.J.M., Van Der Kolk, J., Fase, K.M., Bol, I., Poiger, H., Brouwer, A., Van Den Berg, M., 1994. Toxic potency of 3,3’4,4’,5-pentachlorobiphenyl relative to and in combination with 2,3,7,8-tetrachlorodibenzo-p-dioxin in a subchronic feeding study in the rat, Toxicology and Applied Pharmacology 127, 209-221.

Van Caneghem, J., Block, C., Vandecasteele, C., 2014. Destruction and formation of dioxin-like PCBs in dedicated full scale waste incinerators, Chemosphere 94, 42-47. Van den Berg, M., Birnbaum, L., Bosveld, A.T.C., Brunström, B., Cook, P., Feeley, M., Giesy, J.P., Hanberg, A., Hasegawa, R., Kennedy, S.W., Kubiak, T., Larsen, J.C., Van Leeuwen, F.X.R., Liem, A.K.D., Nolt, C., Peterson, R.E., Poellinger, L., Safe, S., Schrenk, D., Tillitt, D., Tysklind, M., Younes, M., Wærn, F., Zacharewski, T., 1998. Toxic equivalency factors (TEFs) for PCBs, PCDDs, PCDFs for humans and wildlife, Environ Health Perspect106, 775-792. Van Den Berg, M., Birnbaum, L., Bosveld, A.T.C., Brunström, B., Cook, P., Feeley, M., Giesy, J.P., Hanberg, A., Hasegawa, R., Kennedy, S.W., Kubiak, T., Larsen, J.C., Van Leeuwen, F.X.R., Liem, A.K.D., Nolt, C., Peterson, R.E., Poellinger, L., Safe, S., Schrenk, D., Tillitt, D., Tysklind, M., Younes, M., Wærn, F., Zacharewski, T., 1998. Toxic equivalency factors (TEFs) for PCBs, PCDDs, PCDFs for humans and wildlife, Environ Health Perspect 106, 775-792. Van den Berg, M., Birnbaum, L.S., Denison, M., De Vito, M., Farland, W., Feeley, M., Fiedler, H., Hakansson, H., Hanberg, A., Haws, L., Rose, M., Safe, S., Schrenk, D., Tohyama, C., Tritscher, A., Tuomisto, J., Tysklind, M., Walker, N., Peterson, R.E., 2006. The 2005 World Health Organization reevaluation of human and mammalian toxic equivalency factors for dioxins and dioxin-like compounds, Toxicol Sci 93, 223-241.

Van den Berg, M., De Jongh, J., Poiger, H., Olson, J.R., 1994. The toxicokinetics and metabolism of polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) and their relevance for toxicity, Crit Rev Toxicol 24, 1-74. Van Der Burght, A.S.A.M., Clijsters, P.J., Horbach, G.J., Andersson, P.L., Tysklind, M., Van Den Berg, 211

A


M., 1999. Structure-dependent induction of CYP1A by polychlorinated biphenyls in hepatocytes of cynomolgus monkeys (Macaca fascicularis), Toxicol. Appl. Pharmacol. 155, 13-23. Van der Burght, A.S.A.M., Tysklind, M., Andersson, P.L., Horbach, G.J., Van den Berg, M., 2000. Structure dependent induction of CYP1A by polychlorinated biphenyls in hepatocytes of male castrated pigs, Chemosphere 41, 1697-1708. Van Duursen, M.B.M., Sanderson, J.T., van de Berg, M., 2005. Cytochrome P450 1A1 and 1B1 in human blood lymphocytes are not suitable as biomarkers of exposure to dioxin-like compounds: Polymorphisms and interindividual variation in expression and inducibility, Toxicological Sciences 85, 703-712.

Van Duursen, M.B.M., Sanderson, J.T., Van der Bruggen, M., Van der Linden, J., Van den Berg, M., 2003. Effects of several dioxin-like compounds on estrogen metabolism in the malignant MCF-7 and nontumorigenic MCF-10A human mammary epithelial cell lines, Toxicology and Applied Pharmacology 190, 241-250. van Ede, K.I., Andersson, P.L., Gaisch, K.P.J., van den Berg, M., van Duursen, M.B.M., 2013a. Comparison of intake and systemic relative effect potencies of dioxin-like compounds in female mice after a single oral dose, Environ. Health Perspect. 121, 847-853. van Ede, K.I., Andersson, P.L., Gaisch, K.P.J., van den Berg, M., van Duursen, M.B.M., 2014. Comparison of intake and systemic relative effect potencies of dioxin-like compounds in female rats after a single oral dose, Arch. Toxicol. 88 (3), 637-646.

van Ede, K.I., Aylward, L.L., Andersson, P.L., van den Berg, M., van Duursen, M.B.M., 2013c. Tissue distribution of dioxin-like compounds: Potential impacts on systemic relative potency estimates, Toxicol. Lett. 220, 294-302. Van Leeuwen, F.X.R., Feeley, M., Schrenk, D., Larsen, J.C., Farland, W., Younes, M., 2000. Dioxins: WHO’s tolerable daily intake (TDI) revisited, Chemosphere 40, 1095-1101. VanDen Heuvel, J.P., Clark, G.C., Tritscher, A.M., Lucier, G.W., 1994. Accumulation of polychlorinated dibenzo-p-dioxins and dibenzofurans in liver of control laboratory rats, Fundamental and Applied Toxicology 23, 465-469. Villeneuve, D.L., Blankenship, A.L., Giesy, J.P., 2000. Derivation and application of relative potency estimates based on in vitro bioassay results, Environ. Toxicol. Chem. 19, 2835-2843.

Vondracek, J., Svihalkova-Sindlerova, L., Pencikova, K., Krcmar, P., Andrysik, Z., Chramostova, K., Marvanova, S., Valovicova, Z., Kozubik, A., Gabelova, A., Machala, M., 2006. 7H-Dibenzo[c,g]carbazole and 5,9-dimethyldibenzo[c,g]carbazole exert multiple toxic events contributing to tumor promotion in rat liver epithelial ‘stem-like’ cells, Mutat. Res. 596, 43-56.

Voorman, R., Aust, S.D., 1987. Specific binding of polyhalogenated aromatic hydrocarbon inducers of cytochrome P-450d to the cytochrome and inhibition of its estradiol 2-hydroxylase activity, Toxicol. Appl. Pharmacol. 90, 69-78. Watanabe, M.X., Kunisue, T., Ueda, N., Nose, M., Tanabe, S., Iwata, H., 2013. Toxicokinetics of dioxins and other organochlorine compounds in Japanese people: Association with hepatic CYP1A2 expression levels, Environ. Int. 53, 53-61. 212


References Weistrand, C., Norén, K., 1998. Polychlorinated naphthalenes and other organochlorine contaminants in human adipose and liver tissue, Journal of Toxicology and Environmental Health - Part A 53, 293-311.

Wenner, M.M., Wilson, T.E., Davis, S.L., Stachenfeld, N.S., 2011. Pharmacological curve fitting to analyze cutaneous adrenergic responses, J. Appl. Physiol. 111, 1703-1709. Westerink, W.M.A., Stevenson, J.C.R., Schoonen, W.G.E.J., 2008. Pharmacologic profiling of human and rat cytochrome P450 1A1 and 1A2 induction and competition, Arch. Toxicol. 82, 909-921.

White, S.S., Birnbaum, L.S., 2009. An overview of the effects of dioxins and dioxin-like compounds on vertebrates, as documented in human and ecological epidemiology, J Environ Sci Health C Environ Carcinog Ecotoxicol Rev 27, 197-211.

Whitlock Jr., J.P., Cooper, H.L., Gelboin, H.V., 1972. Aryl hydrocarbon (benzopyrene) hydroxylase is stimulated in human lymphocytes by mitogens and benz[a]anthracene, Science 177, 618-619.

Wiebel, F.J., Wegenke, M., Kiefer, F., 1996. Bioassay for determining 2,3,7,8-tetrachlorodibenzo-p-dioxin equivalents (TEs) in human hepatoma HepG2 cells, Toxicol. Lett. 88, 335-338.

Wittsiepe, J., Fürst, P., Schrey, P., Lemm, F., Kraft, M., Eberwein, G., Winneke, G., Wilhelm, M., 2007. PCDD/F and dioxin-like PCB in human blood and milk from German mothers, Chemosphere 67, S286-S294. Wolfe, W.H., Michalek, J.E., Miner, J.C., Rahe, A., Silva, J., Thomas, W.F., Grubbs, W.D., Lustik, M.B., Karrison, T.G., Roegner, R.H., Williams, D.E., 1990. Health status of Air Force veterans occupationally exposed to herbicides in Vietnam. I. Physical health, J. Am. Med. Assoc. 264, 1824-1831. Xu, L., Li, A.P., Kaminski, D.L., Ruh, M.F., 2000. 2,3,7,8 Tetrachlorodibenzo-p-dioxin induction of cytochrome P4501A in cultured rat and human hepatocytes, Chem. -Biol. Interact. 124, 173-189.

Yoshimura, H., Kuroki, J., Koga, N., Kuroki, H., Masuda, Y., Fukasaku, N., Hasegawa, M., 1984. High accumulation of 2,3,4,7,8-pentachlorodibenzofuran to hepatic microsomes of rats, J. Pharmacobio-dyn. 7, 414-419.

Zeiger, M., Haag, R., Höckel, J., Schrenk, D., Schmitz, H.J., 2001. Inducing effects of dioxin-like polychlorinated biphenyls on CYP1A in the human hepatoblastoma cell line HepG2, the rat hepatoma cell line H4IIE, and rat primary hepatocytes: Comparison of relative potencies, Toxicological Sciences 63, 65-73.

213

A


214


Nederlandse samenvatting

Nederlandse samenvatting Inleiding Dioxinen en dioxine-achtige stoffen Dioxinen en dioxine-achtige stoffen zijn organische verbindingen, waarvan er enkele zeer giftig zijn. Ze lossen gemakkelijk op in vet en zijn moeilijk afbreekbaar. Door deze eigenschappen accumuleren deze stoffen in zowel het milieu (sediment en bodem) als in de vetweefsels van mens en dier. De naam â&#x20AC;&#x153;dioxinenâ&#x20AC;? wordt vaak gebruikt om te verwijzen naar een groep verbindingen met een nauw aan elkaar verwante chemische basisstructuur. Deze verbindingen worden aangeduid als polychloordibenzo-para-dioxinen (PCDDs, dioxinen), polychloordibenzofuranen (PCDFs, dibenzofuranen) en sommige dioxine-achtige polychloorbifenylen (PCBs). Tetrachloordibenzo-para-dioxine (TCDD) is de meest giftige dioxine. Bron Dioxinen en dibenzofuranen worden niet met opzet geproduceerd maar zijn ongewenste bijproducten van chemische en thermische processen zoals bijvoorbeeld de productie van sommige chloorhoudende bestrijdingsmiddelen, de productie van papier maar ook bij (afval)verbrandingsprocessen. Zo zijn dioxinen aangetoond in de uitstoot van afvalverbrandingsinstallaties, in sigarettenrook en in de as van barbecues en open haarden. Daarnaast kunnen dioxinen en furanen ook worden gevormd tijdens natuurlijke verbrandingsprocessen zoals bosbranden en vulkaanuitbarstingen. In tegenstelling tot dioxinen en dibenzofuranen zijn PCBs jarenlang geproduceerd en toegepast als onder andere vlamvertragers, isolatievloeistof in transformatoren en condensatoren en in verf en lijm. De productie van PCBs is sinds de jaren tachtig van de vorige eeuw verboden.

Blootstelling Dioxinen, dibenzofuranen en PCBs zijn wereldwijd verspreid en kunnen bijna in het gehele ecosysteem (lucht, water, dieren, mens) worden gevonden. Voor de mens is het consumeren van zuivel-, vlees- en visproducten de belangrijkste bron van blootstelling aan dioxinen, dibenzofuranen en PCBs. Door strenge emissie-reducerende maatregelen en het controleren van diervoeders en levensmiddelen op de aanwezigheid van deze stoffen, is de blootstelling voor de mens met ongeveer 90% verminderd in vergelijking met de jaren zeventig van de vorige eeuw. Desondanks is in sommige landen voor bepaalde bevolkingsgroepen de blootstelling nog steeds te hoog. Met name zuigelingen vormen een gevoelige groep die soms via moedermelk blootgesteld wordt aan concentraties tot wel honderd keer boven de toelaatbare dosis. 215

A


Belangrijkste effecten en werkingsmechanisme Een kortdurende blootstelling aan een hoge concentratie dioxine-achtige stoffen kan bij de mens leiden tot huidaandoeningen, zoals chlooracne (een zware acne-achtige aandoening) en verstoorde leverfuncties. Langdurige blootstelling leidt tot aantasting van het immuunsysteem, zenuwstelsel en de hormoonhuishouding, inclusief effecten op de voorplanting en ontwikkeling. Verder is op basis van dierstudies en gegevens over humane blootstelling aan dioxine-achtige stoffen gebleken dat in ieder geval TCDD kankerverwekkend is bij hoge doseringen. Het wordt aangenomen dat binding aan en activatie van de aryl hydrocarbon receptor (Ah-receptor) een belangrijke eerste stap is in de toxiciteit van dioxinen. Een receptor is een eiwit waaraan een specifiek molecuul kan binden waarna een cellulair response op gang gebracht wordt. De Ah-receptor bevindt zich in bijna iedere cel (lever, long, darm, etc.) van ons lichaam. Als dioxinen of dioxine-achtige stoffen binden aan de Ah-receptor worden onder andere bepaalde enzymen geactiveerd. De meest belangrijke enzymen zijn cytochroom P450 CYP1A1, 1A2 en 1B1, welke betrokken zijn bij de afbraak van lichaamsvreemde stoffen. Hoe hoger de blootstelling aan dioxine-achtige stoffen, hoe meer enzymen er worden geactiveerd. Om die reden worden deze enzymen vaak gebruikt als biomarker (parameter) om dioxineblootstelling aan te tonen.

Risicoschatting Vanwege de persistentie in het milieu en de aangetoonde toxiciteit van dioxinen bij lage concentraties is risicoschatting voor deze groep stoffen nog steeds erg belangrijk. Echter, risicoschatting is ook lastig doordat mensen en dieren worden blootgesteld aan een complex mengsel van verschillende dioxine-achtige stoffen met verschillende toxische activiteit. Aangezien de toxiciteit van dioxineverbindingen via hetzelfde werkingsmechanisme gaat (namelijk via de Ah-receptor) wordt algemeen aangenomen dat de toxiciteit van een mengsel additief is. Dit heeft geleid tot de ontwikkeling van het toxische equivalenten concept (TEQ). Voor dit concept is aan elke dioxine-achtige verbinding (congeneer) een specifieke toxische equivalentiefactor (TEF) toegewezen, die gerelateerd is aan de meest toxische congeneer, TCDD. Bij deze methode heeft TCDD een TEF van 1. Daarnaast zijn er nog 28 andere congeneren, die een TEF gelijk of lager dan 1 toegekend hebben gekregen (zie tabel 1, hoofdstuk 1). Iedere TEF is gebaseerd op zoveel mogelijk relatieve potenties (REPs) die zijn verkregen uit vaak verschillende experimenten. Een relatieve potentie is de verhouding van een effect-concentratie tussen een congeneer en TCDD. Met andere woorden, een REP geeft aan hoe potent (giftig) de stof is ten opzichte van TCDD in dat betreffende experiment. Deze REPs kunnen worden berekend uit dierstudies of celmodellen voor toxische eindpunten (zoals bijvoorbeeld gewichtsafname of het ontwikkelen van tumoren) of voor biochemische eindpunten (zoals activatie van bepaalde enzymen zoals CYP1A1, 1B1 en 1A2). Om de totale toxici216


Nederlandse samenvatting

teit (TEQ) in bijvoorbeeld voedsel te bepalen wordt de concentratie van elke congeneer vermenigvuldigd met zijn TEF en vervolgens bij elkaar worden opgeteld. Gedurende de afgelopen decennia heeft de Wereldgezondheidsorganisatie (WHO) de TEF waarden voor dioxine-achtige stoffen geharmoniseerd, zodat het bepalen van risico’s in levensmiddelen, diervoeders en bij bevolkingsonderzoek overal op de wereld op een zelfde wijze kan plaatsvinden.

Onzekerheden in het TEF-concept Ondanks de enorme hoeveelheid wetenschappelijke gegevens die zijn gepubliceerd sinds de ontwikkeling van het TEF-concept, zijn er nog steeds een aantal onzekerheden binnen de TEF-methodologie. Zo wordt het TEF-concept onder andere ook toegepast bij risicoschatting van een bevolkingsgroep op basis van concentraties gemeten in het bloed. Op dit moment is niet met zekerheid te zeggen of dit wetenschappelijk verantwoord is. Dit heeft te maken met het feit dat de huidige TEFs zijn afgeleid van REPs die zijn berekend in dierstudies waarbij de orale dosis (wat de dieren toegediend hebben gekregen) is gekoppeld aan het effect (zoals activatie van bijvoorbeeld een lever enzym). Hierdoor zouden deze TEFs mogelijk alleen toepasbaar zijn voor de risicoschatting waarbij de blootstelling via het dieet plaatsvindt en dus niet op basis van een bloedconcentratie. Immers de kinetiek, ofwel de opname via de darmen, distributie in het lichaam, afbraak (metabolisatie) en uitscheiding via urine en faeces, kan per dioxine-achtige stof verschillend zijn. Aan het begin van dit promotie-onderzoek was het niet bekend of een TEF waarde gebaseerd op orale inname ook gebruikt kan worden voor de risicoschatting op basis van een concentratie gemeten in het bloed, waardoor mogelijk het risico foutief wordt ingeschat. Het ontwikkelen van zogenaamde “systemische” TEFs, oftewel TEFs die zijn gebaseerd op REPs waarbij een systemische (bloed, lever of vetweefsel) concentratie is gekoppeld aan een effect, zou de risicoschatting mogelijk kunnen verbeteren en betrouwbaarder maken. Een ander belangrijke onzekerheid in het huidige TEF-concept komt voort uit het feit dat de TEFs voornamelijk zijn gebaseerd op studies met knaagdieren. Deze TEFs worden nu algemeen toegepast bij de risicobeoordeling voor mensen, terwijl voldoende wetenschappelijke validatie hiervoor ontbreekt. Tijdens een expert meeting in 2005 van de Wereldgezondheidsorganisatie, zijn deze onzekerheden uitvoerig besproken en benadrukt. Hierbij is geconcludeerd dat er meer wetenschappelijke studies noodzakelijk zijn om vast te stellen of er daadwerkelijk aparte “systemische” TEFs ontwikkeld moeten worden alsook specifieke “humane” TEFs. Om deze onzekerheden binnen het TEF-concept beter in kaart te kunnen brengen is 217

A


het EU-project SYSTEQ geïnitieerd. Het werk dat beschreven wordt in dit proefschrift is onderdeel van dit EU-project. Doel van dit proefschrift

1. Het bestuderen van mogelijke verschillen in relatieve potenties uitgerekend op basis van een orale (intake) dosis en op basis van een systemische (bloed, lever, vetweefsel) concentratie. 2. Het bestuderen van mogelijke verschillen in relatieve potenties bij knaagdieren en die van de mens. Resultaten In het onderzoek beschreven in hoofdstuk 2 en 3 van dit proefschrift hebben we relatieve potenties op basis van een orale dosis (intakeREPs) vergeleken met REPs op basis van een systemische lever, bloed of vetweefsel concentratie (systemicREPs) voor zowel de muis als de rat. Voor deze studies is gekozen voor vrouwelijke C57bl/6 muizen en Sprague Dawley ratten omdat beide soorten in het verleden veelvuldig zijn gebruikt voor onderzoek aan dioxine-achtige stoffen. De studies zijn uitgevoerd met 2,3,7,8-tetrachloordibenzodioxine (TCDD), 1,2,3,7,8-pentachlorodibenzo-p-dioxine (PeCDD), 2,3,4,7,8-pentachlorodibenzofuran (4-PeCDF), 3,3’,4,4’,5-pentachlorobiphenyl (PCB 126) 2,3’,4,4’,5-pentachlorobiphenyl (PCB 118) en 2,3,3’,4,4’,5-hexachlorobiphenyl (PCB 156) en de niet-dioxineachtige 2,2’,4,4’,5,5’-hexachlorobiphenyl (PCB 153). Deze congeneren vertegenwoordigen ongeveer 90% van de dioxine-achtige activiteit in ons dieet. Drie dagen na orale blootstelling zijn de intakeREPs en systemicREPs berekend op basis van Cyp1a1 activiteit en Cyp1a1, 1a2 en 1b1 genexpressie in de lever en witte bloedcellen. Uit de resultaten van beide studies blijkt dat systemicREPs aanzienlijk kunnen afwijken van intake REPs (zie figuur 2, hoofdstuk 2 en figuur 1, hoofdstuk 3). Zo zijn de systemicREPs voor PCB 118 en 156 gebaseerd op de lever concentratie in de muis tot wel 10 keer hoger dan de REP gebaseerd op de orale dosis. Echter, voor humane risicoschatting is de systemic REP op basis van plasma-concentratie het meest interessant. Als we alleen kijken naar de systemicREPs op basis van plasma-concentratie dan zien we dat, behalve 4-PeCDF, geen van de andere congeneren een verschuiving laat zien in relatieve potentie ten opzichte van de intakeREP. Alleen voor 4-PeCDF is de plasma-gebaseerde systemicREP duidelijk hoger in vergelijking met zijn intakeREP (zie figuur 3, hoofdstuk 3).

Een relevante vraag met betrekking tot onze drie-dagen studies, waarbij de dieren éénmalig en kortstondig zijn blootgesteld, is of deze uitkomsten vergelijkbaar zijn met 218


Nederlandse samenvatting

andere studies met een chronisch (langer dan 90 dagen) of subchronisch (tot 90 dagen) doseringsschema. Dit is belangrijk om te weten omdat chronisch en subchronische studies een belangrijke rol hebben gespeeld bij het opstellen van de huidige WHO TEFs voor humane risicoschatting. Daarnaast vindt humane blootstelling aan dioxine-achtige stoffen in de praktijk meestal chronisch plaats en niet acuut. In hoofdstuk 4 proberen we een antwoord te geven op deze vraag. Omdat er in de literatuur nog maar heel weinig data beschikbaar zijn over systemicREPs was het niet mogelijk om deze data uitgebreid met elkaar te vergelijken. Er zijn echter wel studies bekend waarbij de concentratie in de lever en het vetweefsel gemeten is na een éénmalige of (sub)chronische blootstelling aan dioxine-achtige stoffen. Deze data zijn bruikbaar omdat het wat zegt over de distributie (verdeling in het lichaam) van dioxine-achtige stoffen. Van uit kwantitatief oogpunt vindt distributie van dioxine-achtige stoffen vooral plaats naar de lever en het vetweefsel, maar dit kan sterk verschillen tussen dioxine-achtige stoffen. Zo verzamelen dioxinen en dibenzofuranen zich met name in de lever, maar PCBs juist meer in het vetweefsel. Onderling, tussen de dioxinen, dibenzofuranen en PCBs zelf, zit ook verschil in distributie. De ratio tussen de concentratie gemeten in de lever en het vetweefsel geeft hier veel informatie over. Deze lever:vetweefsel concentratie ratio hebben wij vergeleken tussen muis- en ratstudies, waarbij de dieren éénmalig of subchronisch werden blootgesteld. Hieruit blijkt dat het distributiepatroon tussen de lever en het vetweefsel voor de verschillende congeneren in onze studies overeenkomen met andere studies die een enkelvoudig of subchronisch doseringsschema gebruikten. Verder hebben we in hoofdstuk 4 voor TCDD de effect-leverconcentraties voor CYP1A1 activiteit vergeleken voor de verschillende orale doseringen. Ook deze concentraties kwamen overeen tussen beide doseringsschema’s. Op basis van deze waarnemingen kan worden geconcludeerd dat systemicREPs die berekend zijn uit onze drie dagen studies hoogstwaarschijnlijk ook representatief zijn voor situaties waarin sprake is van (sub-)chronische blootstelling.

In het onderzoek beschreven in de hoofdstukken 5 en 6 van dit proefschrift is gekeken naar verschillen tussen knaagdieren en de mens in de relatieve potentie van dioxine-achtige stoffen. In het onderzoek beschreven in hoofdstuk 5 zijn hiervoor witte bloedcellen van zowel de mens als de muis gebruikt. Deze witte bloedcellen komen van bloeddonors (mens) of uit de milt (muis). Nadat ze zijn geïsoleerd uit het bloed of de milt hebben we ze in het laboratorium blootgesteld aan 20 verschillende dioxine-achtige stoffen. Net als bij de dierstudies, gaan witte bloedcellen die worden blootgesteld aan dioxinen en dioxine-achtige stoffen, meer CYP1A1 en CYP1B1 enzymen aanmaken. Uit deze experimenten zijn REPs berekend en deze laten zien dat van de 20 stoffen, 1234678-HpCDD, 4-PeCDF, 123478-HxCDF en 1234789-HpCDF aanzienlijk potenter (meer giftig) zijn in de witte bloedcellen van de mens vergeleken met de muis. In tegenstelling, de REP voor PCB 126 was aanzienlijk lager voor de humane witte bloedcellen

219

A


vergeleken met de muis. In het onderzoek beschreven in hoofdstuk 6 hebben we gekeken naar verschil in relatieve potentie voor dioxine-achtige stoffen tussen de rat, muis, cavia en mens. Hiervoor zijn vier genetisch gemodificeerde cellijnen gebruikt. De cellijnen zijn genetisch gemodificeerd met een gen uit een vuurvlieg. Hierbij leidt binding aan de Ah-receptor in deze cellijnen tot een meetbare lichtproductie. Hoe meer dioxinen, hoe meer licht de cel maakt. Ook hier hebben we 20 verschillende dioxine-achtige stoffen getest. De resultaten in dit onderzoek laten opnieuw zien dat van de 20 stoffen, 1234678-HpCDD, 123478-HxCDF en 1234789-HpCDF potenter zijn in de humane cellijn vergeleken met de drie knaagdiercellijnen. En ook in deze studie was PCB 126 aanzienlijk minder potent in de humane cellijn vergeleken met de drie knaagdiercellijnen. Samengevat laten de resultaten uit hoofdstuk 5 en 6 zien dat er voor sommige dioxine-achtige stoffen verschil in relatieve potentie is tussen de mens en knaagdieren, waarbij 1234678-HpCDD, 123478-HxCDF en 1234789-HpCDF meer en PCB 126 minder potent voor de mens lijken te zijn. Humane risicoschatting Het mag duidelijk zijn dat voor een goede risicoschatting voor de mens het belangrijk is om te rekenen met correcte en wetenschappelijk verantwoorde TEFs. Risicoschatting voor een bevolkingsgroep wordt onder andere uitgevoerd op basis van een concentratie gemeten in het bloed. Omdat de huidige TEFs zijn afgeleid van dierstudies waarbij de orale dosis is gekoppeld aan een effect, is niet met zekerheid te stellen dat deze TEFs ook gebruikt kunnen worden bij humane risicoschatting op basis van een bloedconcentratie. In dit proefschrift laten we in hoofdstuk 2 en 3 zien dat van de zes dioxine-achtige stoffen die getest zijn, alleen de relatieve potentie voor 4-PeCDF op basis van een plasmaconcentratie duidelijk hoger ligt in vergelijking met de relatieve potentie berekend op basis van de orale dosis. Het is aannemelijk dat deze verschuiving, die we zien bij deze dierstudies, ook bij de mens zijn te verwachten. Dit geeft aan dat voor 4-PeCDF het ontwikkelen van een aparte “systemische” TEF wenselijk is. Voor de overige vijf dioxine-achtige stoffen lijkt dit niet noodzakelijk. Voor 1234678-HpCDD, 123478-HxCDF en 1234789-HpCDF zien we in hoofdstuk 5 en 6 duidelijke verschillen in relatieve potenties tussen de mens en knaagdieren. Gebaseerd op de huidige TEFs vertegenwoordigen de congeneren 1234678-HpCDD, 123478-HxCDF en 1234789-HpCDF bij de algemene bevolking maar een heel klein percentage (< 3%) van de dioxine-achtige activiteit gemeten in het bloed. Als er specifieke “humane” TEFs ontwikkeld zouden worden voor deze congeneren dan zien we dat vooral de bijdrage van 1234678-HpCDD behoorlijk stijgt en met 15% één van de belangrijkere 220


Nederlandse samenvatting

congeneren wordt van de totale dioxine-achtige activiteit gemeten in het bloed. Voor 123478-HxCDF en 1234789-HpCDF is de bijdrage aan de totale dioxine-achtige activiteit in het bloed ook na aanpassing van de TEF-waarde nog steeds klein voor de algemene bevolking. Een aangepaste TEF zou wel gevolgen kunnen hebben voor de risicoschatting van een specifieke bevolkingsgroep, die is blootgesteld aan dibenzofuranen. In de literatuur is nog weinig bekend over de relatieve potenties van deze congeneren in de mens. Het is daarom belangrijk dat hier meer onderzoek naar wordt gedaan en deze nieuwe bevindingen bij een her-evaluatie van de TEFs worden meegenomen in de besluitvorming. In tegenstelling tot de dioxinen en dibenzofuranen lijkt de mens minder gevoelig te zijn voor PCBs dan knaagdieren. In hoofdstuk 5 en 6 van dit proefschrift is te lezen dat, afgezien van PCB 126, geen van de andere PCBs het CYP1A1 enzym zodanig konden induceren dat er relatieve potenties berekend konden worden. Dit heeft waarschijnlijk te maken met het feit dat dioxinen en dibenzofuranen anders binden aan de Ah-receptor in vergelijking met PCBs. De gecombineerde data van hoofdstuk 5 en 6 laten voor PCB 126 een relatieve potentie zien die 100 keer lager ligt dan zijn huidige TEF van 0.1. Ook in de literatuur is PCB 126 al vaak getest in humane lever-, borst-, prostaat- en huidcellen. Resultaten uit deze studies laten, net als de resultaten in dit proefschrift zien, dat de relatieve potentie voor PCB 126 tot 100 keer onder de huidige TEF ligt. PCB 126 is, gebaseerd op de huidige TEFs, met ongeveer 15% bijdrage, één van de meest belangrijke congeneren voor de totale dioxine-achtige activiteit gemeten in het bloed van de algemene bevolking. Als de TEF met een factor 100 naar beneden zou worden bijgesteld voor humane risicoschatting, betekent dit dat de bijdrage van PCB 126 ineens geheel te verwaarlozen is. Op basis van de data uit dit proefschrift en uit de literatuur lijkt de mens dus duidelijk minder gevoelig voor PCB 126. Het is daarom sterk aan te bevelen dat de TEF voor PCB 126 wordt her-geëvalueerd voor humane risicoschatting. Conclusie De belangrijkste doelstelling van dit proefschrift was het in kaart brengen of het wenselijk is om aparte “systemisch” dan wel “humaan” specifieke TEFs te ontwikkelen ter verbetering van de humane risicoschatting voor dioxine-achtige stoffen. Als we alle data van dit proefschrift samenvoegen, is duidelijk dat voor sommige congeneren de huidige TEF een onder- of overschatting van het risico kunnen geven voor humane risicoschatting op basis van een concentratie gemeten in het bloed. Voor 4-PeCDF zien we hogere relatieve potenties op basis van plasmaconcentratie vergeleken met de orale dosis. Het is aan te bevelen om een aparte systemische TEF voor 4-PeCDF te ontwikkelen. Hu221

A


mane relatieve potenties voor PCB 126 zoals berekend in dit proefschrift en zoals beschreven in de wetenschappelijke literatuur zijn duidelijk lager dan de huidige TEF en her-evaluatie van deze TEF voor humane risicoschatting is zeer wenselijk. Verder zien we duidelijk hogere humane relatieve potenties voor 1234678-HpCDD, 123478-HxCDF en 1234789-HpCDF in vergelijking met de huidige TEFs. Vanwege hun mogelijke bijdrage in de totale dioxine-achtige activiteit is verder onderzoek voor deze congeneren eveneens aan te raden. Het opnieuw her-evalueren van de TEFs is ook momenteel nog steeds relevant. Organisaties als de WHO en de Gezondheidsraad gaan er, ondanks de aanzienlijke afname van deze dioxine-achtige stoffen in het milieu, nog steeds vanuit dat de huidige blootstellingsniveaus subtiele effecten teweeg kunnen brengen bij onder andere de fetus, zuigeling en kleine kinderen. Dit proefschrift kan een belangrijke bijdrage leveren voor een eventuele her-evaluatie van de TEFs.

222


Nederlandse samenvatting

223

A


DANKWOORD Tja, en dan ben je zomaar aangekomen bij het schrijven van het dankwoord van je proefschrift. Wie had dat bijna 30 jaar geleden gedacht, toen ik als klein meisje mijn “carrière” begon op een basisschool voor moeilijk lerende kinderen. Het is een lange maar hele mooie weg geweest naar dit voor mij toch wel hoogtepunt. Ik wil graag iedereen die op welke manier dan ook een bijdrage heeft geleverd hiervoor bedanken. Zonder al jullie hulp, inzet en steun, zowel voor als tijdens mijn promotietraject, was dit niet gelukt! Een aantal mensen wil ik graag in het bijzonder noemen.

Allereerst mijn promotor en co-promotor Martin en Majorie. Bedankt voor jullie vertrouwen in mij, jullie begeleiding, kritische discussies, enthousiasme en gedrevenheid. Ik had me geen betere begeleiders kunnen wensen. Martin, jij bent binnen de dioxinenwereld een fenomeen. Dit viel vooral op tijdens congressen zoals DIOXIN, als Majorie en ik je weer eens moesten redden uit een belaging van “fans”. Jouw directe betrokkenheid bij mijn project was dan ook groot. Vooral tijdens het schrijven van mijn manuscripten had jij altijd goede adviezen en wist je de verloren “rode draad” weer feilloos boven te brengen. Heel erg bedankt voor al je tips, kennis en… vrolijkheid! Majorie, ik zie ons nog zitten aan het begin van het project met alle dierstudies voor ons… “Where to start?” Zonder jouw goede begeleiding en inschattingen hadden deze studies nooit zo’n succes geworden. Maar ook je kritische blik en commentaren op mijn manuscripten waren altijd erg leerzaam. Je bent een kei in het recht breien van kromme zinnen met het wel bekende “Tadaaaa” effect! Daarnaast werkt jouw enthousiasme heel aanstekelijk en ben ik je dankbaar voor alle goede gesprekken. Ook heb ik erg genoten van al onze reisjes naar congressen en EU annual-meetings, waarbij we naast het officiële gebeuren erg goed waren in het vinden van de dansvloer. Majorie, proost, houdoe en bedankt ;)! Konrad, zonder al jouw harde werken had dit boekje toch behoorlijk leeg geweest. Ik denk dat er niemand zo nauwkeurig werkt als jij. Of ik nu je labjournaal opensloeg of een kleurrijk excel-bestand, ik werd er altijd even blij van. Dank je wel voor al je inzet, je precisie, je gezelligheid EN je sushi! Ik wens je alle goeds. Esmée, stickerkoningin, ik was laatst nog eens aan het nagaan hoeveel vials, bloedbuizen, greinerbuizen en petrischaaltjes je gestickerd moet hebben… Ik ben maar gestopt bij 10.000… Oprecht respect. Ook als student heb je op het SYSTEQ project gewerkt en ik vraag me nog steeds af: “Wel zon maar geen licht”… Esmée (en Evelien)… ik ben er nog steeds niet uit! Suzan, jij was mijn dropverslaafde Speedy Gonzales. Ongelooflijk hoe snel jij ZONDER fouten experimenten kon uitvoeren. Esmée en Suzan, heel erg bedankt voor jullie harde werken, ik wens jullie alle succes toe in de toekomst. Wouter en Maarke, dank jullie wel voor jullie hulp tijdens de dierstudies!

224


Dankwoord

Lieve GDL-mensen, zonder jullie hadden de muisjes en ratjes letterlijk op de tafel gedanst. Helma, Anja, Sabine, Ron, Janine, Tamara en Kiki, dank jullie wel voor al jullie hulp, gezelligheid, roddels en meezing muziek!

I would also like to thank Mirek, Jan, Tomas, Dieter, Sylke, Christiane, Hans, Patrik, Malin, Mehdi, Sture and Lorenz, for the close collaboration within the EU SYSTEQ project. I will definitely miss our collaboration and annual meetings. Sture, a BIG thank you for analyzing all the liver, adipose tissue and plasma samples. I don’t even want to imagine how much work that must have been! Mehdi, fingers crossed for the CALUX paper! Sylke and Christiane, we managed to prepare over 500 RNA and 1500 cDNA samples in two weeks time and still having fun, thank you for the great time in the lab, at the annual meetings and at the DIOXIN conference in Brussels!

Naast alle SYSTEQ mensen wil ik graag de ETX-groep bedanken. Sandra, duizendpoot op het lab. Dank je wel voor al je hulp en dat ik altijd bij je terecht kon met vragen of gewoon even kletsen. Maarke, roomy, ex-roomy en nu weer roomy. Bedankt voor al je gezelligheid. Beetje jammer dat ik je nooit mee kon krijgen in het aanbrengen van alle foute kerstversiering in onze kamers… maar verder… :) ! Heel veel succes met het afronden van je boekje en natuurlijk met je nieuwe baan. Ik zal je missen als roomy! Kamila, thank you for all the good talks and great plums! Before you know, we will be opening your bottle of AhR-wine. I wish you lots of luck with finishing your thesis and finding a job. Fiona, samen met jou naar de eerste hulp met Konrad en daar vervolgens (Konrad half stoned) Madagascar kijken, zal ik niet snel vergeten. Fijn dat je nu bij NTX aan de slag bent en we je nog niet hoeven missen! Irene and Deborah, it seems like such a long time ago! Thank you for giving me a good start at IRAS. Furthermore, I would like to thank all (former)PhD-students, colleagues, roommates and students of IRAS for all the “gezelligheid” in the lab, coffeecorner, during lunch, parties, lab-outings and conferences! Because of you, there was this great balance between hard work and pleasure. Thank you all!

Lieve Lesa, ik leerde je via Majorie kennen tijdens het DIOXIN congres in 2010. Wie had toen kunnen bedenken dat we nu 4 jaar later bijna dagelijks zouden whatsappen! We hebben samen een heel mooi artikel geschreven en reisjes gemaakt in Australië en de USA. Ik ben je heel dankbaar voor alle wetenschappelijke discussies die we via Skype of e-mail hebben gehad maar misschien nog veel meer voor alle support en steun die je gegeven hebt. A BIG THANK YOU! 225

A


Anita, ook wel bekend onder de naam Dutchie… Of ik je nu spreek aan de telefoon (ik geloof dat het record op 2,5 uur staat) of real-life, zeker is dat ik aan het einde van de avond altijd buikpijn van het lachen heb :). Ik heb genoten van onze reis door de US… Na een lange wandeling in de schemering in het dal uitkomen terwijl de kamper boven aan de berg staat, is denk ik wel kenmerkend voor hoe de vakantie verliep. Samen met Harm hebben we ook de AiO-dagen van de NVT georganiseerd. Super leuke tijd. Harm en Anita, bedankt!

Veronica, ik mis je nog steeds op het IRAS, even kletsen, samen zeil-uitjes organiseren of gewoon even van de zon genieten op de patio. Bedankt voor je steun en gezelligheid! Peter en Guus, treinvriendjes… dankzij jullie (en de Latte Macchiato, chocolade croissant en horoscoop) begint de dag altijd goed!

Voor ik op het IRAS begon als AiO was ik werkzaam als analist bij het RIKILT. Ik ben heel veel dank verschuldigd aan Ron. Jij hebt mij de mogelijkheid gegeven me te ontwikkelen in het onderzoek en me enthousiast gemaakt om te promoveren. Zonder jou had ik hier niet gestaan, dat weet ik zeker. Dank je wel voor alles! Naast Ron wil ik graag alle andere collega’s waar ik op het RIKILT mee samengewerkt heb bedanken voor alle gezellige jaren! Liza, sinds ik weg ben bij het RIKILT zijn mijn gadgets aankopen drastisch verminderd maar ik mis alle dagelijkse updates nog steeds! Dank je wel voor alle gezellige etentjes en je vriendschap! Jeroentje, we zien elkaar niet zo heel vaak meer maar als we elkaar zien is het altijd als vanouds gezellig! Ben jij niet aan de beurt met koken? Marleen, bedankt voor alle fijne jaren dat ik als buuf naast je mocht zitten. We hebben wat gelachen samen! Lieve Maurice, dank je wel voor alles. Onze vriendschap is me heel dierbaar. Ik wens je al het geluk van de wereld toe. Win, Marja en Daniella, dank jullie wel voor alle mooie momenten, steun, gezelligheid en een bitje Limburgs kalle :).

Verder wil ik al mijn vrienden en familie bedanken voor alle broodnodige ontspanning (zowel niet-sport als sport gerelateerd). Ik weet dat ik jullie flink verwaarloosd heb maar het schijnt dat het eind in zicht is… Ik maak het goed, dat beloof ik! To all friends, thank you, danke, obrigada, efchariesto and merci for all the great and relaxing moments together and for being there for me! Kristel, Jeroen, Kees, Daphne, Edith en Noor, dank jullie wel voor jullie steun, gezellige etentjes en interesse in mijn onderzoek. Hoewel we nu ver bij elkaar vandaan wonen en we elkaar niet meer zo regelmatig zien, betekenen jullie heel veel voor me. Daan… door onze reis in Zuid-Korea zie ik nu overal vogels vliegen! Dank je wel voor deze mooie reis en onze vriendschap. 226


Dankwoord

Antonio e Teresa, obrigada por todo seu suporte e amor.

Elsa, sol da minha vida, wat hebben we veel meegemaakt samen. Dank je wel voor al je steun, dat ik bij je langs kan komen met een lach en een traan. Voor alle heerlijke Portugese etentjes en voor alle gezelligheid. Wat zou ik zonder je moeten! Ik ben heel blij dat je mijn paranimf wilt zijn! Dank je wel voor wie je bent!

Lieve Oma Liesje, wat ben ik trots op u! Wie heeft er nu een oma van 93 die nog zo bij de tijd is als u. Met wie je het nieuws kan bespreken, wat er in de wereld gaande is, als ook alle gewone dingen in het dagelijks leven. Nu met uw nieuwe aanwinst, de iPad, kunnen we ook gaan Skypen, hoe cool is dat :) ! Ik ben heel blij dat u erbij bent op mijn promotiedag. Je bent de liefste oma van de wereld, dikke knuffel. Lieve Opa Wim, het is me gelukt :)! Lief broertje, we zijn heel verschillend maar toch zo gelijk. Je bent me heel dierbaar. Dank je wel dat je er altijd voor me bent. Nu ook als paranimf. Marieke, met jou heb ik er de beste en liefste zus bij gekregen die ik me had kunnen wensen. Dank je wel! Maite en Bente, Karin is uitgespeeld met haar ratjes en muisjesâ&#x20AC;Ś zullen we naar de dierentuin gaan? Lieve Henrique, your support during this last year is priceless. Thank you for always being there for me, for your love and for your care! Now it is time for having fun! Muito obrigada por tudo, gosto MUITO de ti! Lieve mam en pap, het valt met geen pen te beschrijven wat jullie voor mij betekenen. Jullie hebben altijd in mij gelooft en voor me gevochten. Zonder al jullie inspanningen op school en buiten school, mamaâ&#x20AC;&#x2122;s creativiteit en jullie onvoorwaardelijke steun had ik dit nooit bereikt. Ik ben jullie heel dankbaar voor het geven van een warm en liefdevol thuis waarbinnen ik mij heb kunnen ontplooien tot wie ik nu ben. Dank jullie wel voor alles! Karin

227

A


Curriculum Vitae Karin Irene van Ede was born in Amsterdam, the Netherlands on August 13, 1979. In 1996 she graduated from secondary school at the Groenstrook in Aalsmeer (IVBO) and started the short vocational education in laboratory techniques (KMLO) at the ROC in Leiden. After completing the second year of KMLO she switched to the 4-year vocational education in laboratory techniques (MLO) with a focus on analytical chemistry. During her last year of education she completed two internships at Kraft-Foods in Munich, Germany. During her first internship she worked at the chromatography department where she performed quantitative sugar analysis in food products. After this, she continued as an intern at the R&D department were she characterized and investigated the effect of aroma components to the overall flavor of coffee. After completing MLO in 2001, she continued with a bachelor of applied science in food toxicology at the Hogeschool Larenstein in Velp. During her bachelor Karin did her internship at The Netherlands Organization for Applied Scientific Research (TNO) in Zeist where she investigated the transport of folic acid and 5-methyltetrahydrofolate across Caco-2 cells under supervision of Dr. Miriam Verwei. Upon completion of her bachelor degree in food toxicology in 2004, Karin started working as a research technician within the department of toxicology and effect monitoring at RIKILT - Institute of Food Safety in Wageningen. In addition to screening feed and food products for dioxins and dioxinlike compounds using the DR-CALUX bioassay, she investigated and identified, under supervision of Dr. Ron Hoogenboom, natural arylhydrocarbon receptor (AhR) agonists in citrus fruit using the HPLC in combination with the DR-CALUX bioassay. In 2007 she presented this work with an oral presentation during the 3rd International Symposium on Recent Advances in Food Analysis (RAFA) in Prague, Czech Republic. In addition, in the following year, she and her co-authors also published this work in a scientific journal. Following this work, in 2009 she had the opportunity to undertake a PhD research program at the Institute for Risk Assessment Sciences (IRAS) at Utrecht University. Under supervision of Dr. Majorie van Duursen and Prof. Dr. Martin van den Berg she investigated the uncertainties in the risk assessment of dioxin-like compounds. The results from this research are presented in this thesis. Karin is currently working as a Post-Doctoral Employee within the toxicology department of IRAS, Utrecht University.

228


About the author

List of Publications van Ede KI, Andersson PL, Gaisch KPJ, van den Berg M, and van Duursen MBM (2014) Comparison of intake and systemic relative effect potencies of dioxin-like compounds in female rats after a single oral dose. Arch Toxicol 88:637-646. van Ede KI, Gaisch KP, van den Berg M, and van Duursen MB (2014) Differential relative effect potencies of some dioxin-like compounds in human peripheral blood lymphocytes and murine splenic cells. Toxicol Lett 226:43-52.

van Ede KI, Andersson PL, Gaisch KPJ, van den Berg M, and van Duursen MBM (2013) Comparison of intake and systemic relative effect potencies of dioxin-like compounds in female mice after a single oral dose. Environ Health Perspect 121:847-853.

van Ede KI, Aylward LL, Andersson PL, van den Berg M, and van Duursen MBM (2013) Tissue distribution of dioxin-like compounds: Potential impacts on systemic relative potency estimates. Toxicol Lett 220:294-302.

van Ede KI, Li A, Antunes-Fernandes E, Mulder P, Peijnenburg A, and Hoogenboom R (2008) Bioassay directed identification of natural aryl hydrocarbon-receptor agonists in marmalade. Anal Chim Acta 617:238-245. van Ede KI, Stelloo S, van den Berg M, van Duursen MBM (2010) TCDD induces biomarkers for endometriosis in rat endometrium and human ECC-1 cells. Organohalogen Compounds 72: 1054-1057.

Hoogenboom R, van Ede KI, Portier L, Bor G, Bovee T, and Traag W (2008) The use of the DR CALUX速 assay for identification of novel risks. Organohalogen Compounds 70: 760-763.

Ghorbanzadeh M, van Ede KI, Larson M, Duursen MBM, Poellinger L, L端cke S, Machala M, Pencikova K, Vondracek J, van den Ber M, Denison MS, Ringsted T, Andersson PL. In vitro and in silico derived relative effect potencies of Ah-receptor mediated effects by PCDD/Fs and PCBs in human, rat, mouse and guinea pig CALUX cell lines. Manuscript in preparation.

229

A


Overview of completed training activities Postgraduate Education in Toxicology (PET) program • Toxicological Risk Assessment (2009) • Food Toxicology and Food safety (2009) • Mutagenesis and Carcinogenesis (2010) • Epidemiology (2010) • Medical, Forensic and Regulatory Toxicology (2010) • Molecular Toxicology (2010) • General Toxicology (2011) • Organ Toxicology (2011) • Environmental Toxicology (2011) • Pathobiology (2011) • Immunotoxicology (2012)

General Courses • Laboratory animal science (Art. 9), Utrecht University (2009) • Radiation expert 5B, Van Hall Larenstein (2010)

Meetings • 30th International Symposium on Halogenated Persistent Organic Pollutants, DIOXIN (2010), San Antonio, USA (oral presentation) • 31st International Symposium on Halogenated Persistent Organic Pollutants, DIOXIN (2011), Brussels, Belgium (oral presentation) • 7th Düsseldorf Symposium on Immunotoxicology Biology of the Arylhydrocarbon Receptor, AHR (2011), Düsseldorf, Germany • 51st annual meeting of the American Society of Toxicology, SOT (2012), San Francisco, USA (poster) • 32nd International Symposium on Halogenated Persistent Organic Pollutants, DIOXIN (2012), Cairns, Australia (oral presentation) • 33rd International Symposium on Halogenated Persistent Organic Pollutants, DIOXIN (2013), Daegu, Republic of Korea (oral presentation) • Annual meeting of the Dutch Society of Toxicology and PhD student symposia (2009, 2010, 2011, 2012, 2013).

Organizational Member of the organizing committee of the Dutch Society of Toxicology PhD student symposia, 2011. 230


If children canâ&#x20AC;&#x2122;t learn the way we teach,

then we have to teach the way they learn . Robert Buck


2014

Proefschrift v Ede  
Proefschrift v Ede  
Advertisement