General enquiries on this form should be made to: -...

General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: [email protected]

SID 5 Research Project Final Report

SID 5 (Rev. 3/06) of 30

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code PS2343

2. Project title

A review of the risks of endocrine disruptors to bird, fish and mammalian wildlife and implications for pesticide risk assessment

3. Contractororganisation(s)

CSLSand HuttonYorkYO41 1LZ

54. Total Defra project costs £ 26,601(agreed fixed price)

5. Project: start date................ 01 June 2008

end date................. 31/03/2009

SID 5 (Rev. 3/06) of 30

6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They

should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the

intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.This report reviews published information on the endocrine disrupting effects of plant protection product (PPP) active substances in fish- and mammalian-based systems in vitro, and in vivo in fish, mammals and birds. It also evaluates the suitability of in vivo toxicity testing methods for studying PPPs with endocrine activity.

1. In vitro test systemsFish-based test systems include those using fish tissue explants, sub-cellular organelles, normal cell lines or those transfected with genes of interest. The principles of the assays cover a large range of Mammalian based test systems include those of both human and other mammalian origin. The sensitivity of each assay system varies widely depending on the source/species making direct comparison difficult but it does allow the interaction of the pesticide with the system to be identified and the relative potency of differing pesticides in the same assay system to be compared. Information was found for 55 PPPs, including metabolites.There is an apparent lack of testing of PPPs on avian in vitro systems. The available in vitro data are rather patchy, with in many cases only a single representative chemical in a class and not all the possible mechanisms of endocrine disruption. For example, no in vitro data have been found for thyroid agonism or antagonism, androgenicity, or anti-estrogenicity. However, the available information shows that even this small selection of PPPs has a wide range of different endocrine MOAs, some of which are highly potent.

2. In vivo test systemsIn fish data are only available for a limited number (33) of substances, representing 8 herbicides, 5 fungicides and 20 insecticides (including metabolites). Although only relatively few PPPs have been examined for endocrine effects in vivo in fish, the available data provide examples of most of the known endocrine MOAs, some of which clearly cause adverse effects and occur at environmentally-relevant concentrations. It is important to point out that some of these effects (e.g. thyroid disruption; impaired olfactory responses to prostaglandin-triggered priming of milt maturation) were not predicted on the basis of in vitro tests, thus confirming the value of in vivo studies for describing endocrine disruption. In other cases, however, in vivo effects such as impaired steroidogenesis, estrogen agonism, and androgen antagonism were indeed detected in vitro. In about one third of the PPPs studied, the most sensitive endocrine-related effects in fish occur at lower concentrations than the most sensitive apical (whole organism) effects, but this is probably because the apical tests in common use (such as the early life stage test) are insensitive to endocrine modes of action (MOA).A far wider range of chemicals have been tested for endocrine disrupting activity in mammals but again the lack of a link between in vitro and in vivo results is evident due to the integrated response of the animal in manifesting effects. In approximately 10% of the PPPs studied, the most sensitive endocrine-related effects in mammals occur at lower concentrations than the most sensitive apical effects. This difference

SID 5 (Rev. 3/06) of 30

from fish is probably because the 2-generation study is in widespread use in mammalian toxicity testing.The searches for potential endocrine disrupting effects in birds identified a range of pesticides. In approximately 40% of the PPPs studied, the most sensitive endocrine-related effects in birds occurred at lower concentrations than the most sensitive apical effects. This is probably because the 1-generation study in widespread use in avian toxicity testing is insensitive to some types of endocrine mediated effects. In addition many of the studies related to precocial species such as the quail and mallard and their applicability to altricial species has not been robustly validated, e.g. the difference in importance of parental care.

3. Testing approachesFishA suite of apical fish toxicity tests with proven or expected sensitivity to EDCs is being developed by OECD, but is not expected to be fully validated for another 10 years. This includes a short-term screening test for endocrine activity (which should be finalised in 2009/10), two partial life cycle tests currently in validation, and a full life cycle test which has not yet begun validation. As it stands, this suite is probably sensitive to most, but not all, modes of endocrine action, and it will almost certainly require some further development. One possible strategy for deploying this suite of improved fish toxicity tests in a hazard and risk assessment programme for PPPs is proposed in this report. Finally, the way in which risk assessments should be carried out for PPPs which disrupt the endocrine systems of fish has not yet been fully considered. Some modifications to standard risk assessment procedures will probably be needed to recognise the unique properties of EDCs.MammalsFor birds and mammals the risk assessment for endocrine disruptors was reviewed in the EFSA 2008 Scientific Opinion of the Panel on Plant protection products and their Residues (PPR) on the Science behind the Guidance Document on Risk Assessment for birds and mammals. Under 91/414EEC reproductive and developmental effects studies are required in mammals and specifically identify multi-generation and developmental studies as required for authorisation. It is likely therefore that effects on mammalian reproduction and growth will be identified during the authorisation process. If screening assays are required to address specific effects the proposed tiered testing strategy under development by the US EPA provides an appropriate approach. BirdsEFSA recognises the distinction between birds and mammals in the recent Bird and Mammal guidance document opinion “For mammals, screening and testing methodology is available, that allows for adequate assessment of endocrine mediated effects on reproduction. This is much less the case for birds. Due to differences in mechanisms for sex differentiation, absence of endocrine mediated effects in in vivo mammalian studies cannot be considered sufficient to negate concerns for potential endocrine effects in birds when (in vitro-) screening tests have demonstrated for a substance to have a potential to influence endocrine processes. Therefore, at present a fully conclusive assessment for birds is not feasible.”

The EPA EDSTAC (Endocrine Disruptor Screening and Testing Advisory Committee) proposed that a tier 1 screening battery of in vitro and in vivo assays aimed at detecting effects based on estrogen, androgen and thyroid hormones would be sufficiently comprehensive to detect any activity likely to also occur in birds.At tier 2 the current one-generation reproduction test is not sufficient as it does not assess the ability of the F1 generation to reproduce. Therefore the recommended assay for birds is the 2-generation test. The proposed OECD 2-generation study in quail is in the pre-validation phase but it is likely to be many years before it is routinely used.

The clear omission in all approaches to date is the effects of endocrine disruptors on altricial species. All OECD test methods are based on precocial species (Japanese quail, northern bobwhite quail, mallard) and therefore studies do not take into account parental care, e.g. courtship, nest building and caring for young, which are important in a wide range of species. Only reproductive success is currently addressed in the proposed 2-generation OECD test design. It is therefore recommended that a study design capable of identifying endocrine-related effects in altricial species is developed so as to address this major gap in the ability to detect effects of pesticides.

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with

details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include:

SID 5 (Rev. 3/06) of 30

the scientific objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Transfer).

1. Review of ED effects and effect levels of pesticides on fish, birds and mammals (UK and overseas)1.0 Introduction

According to accepted definitions an endocrine disruptor (ED) is an exogenous substance or mixture that alters function of the endocrine system and consequently causes adverse health effects in an intact organism, or its progeny or (sub)populations. Endocrine disruptors may act by the following mechanisms: Direct damage to an endocrine organ; Direct altering of the function of an endocrine organ; Interaction with receptors; Altering hormone metabolism, either in endocrine organs or peripherally.A significant number of studies have been undertaken on the potential for endocrine disruption to occur in vertebrates (although few in wildlife species apart from fish) and the data reviewed in the context of EU including a significant review by Groshart and Okkerman (2000) for the establishment of a priority list and an updating for low production volume chemicals by Petersen et al in 2007. Several hundred existing substances are suspected by the European Union (EU) of having endocrine disrupting properties (Groshart and Okkerman, 2000; Petersen et al., 2007), although the supporting data are weak in most cases. The paucity of data is mainly due to the fact that there are, as yet, no internationally-standardised tests focussed directly at identifying for endocrine disrupting chemicals (EDCs). Some pesticides or their degradation products (e.g. vinclozolin) are nevertheless known to be EDCs (although they may not necessarily cause ED at environmental concentrations), but the available information is patchy, and possible endocrine disruption by proposed new actives is not routinely taken into account by pesticide regulators. Within the Groshart and Okkerman (2000) listing there are 18 groups of pesticides identified as having reported effects in invertebrate and/or vertebrate wildlife, mainly in aquatic laboratory-based studies. The Petersen et al (2007) report includes a number of pesticides e.g. trifluralin, omethoate, quinalphos. However, other reviews have questioned the exposure levels used in such studies and the evidence of effects on vertebrate wildlife in the environment is far more limited (Dawson 2000). Furthermore, reports of endocrine disruption in the surface waters of agricultural areas (e.g. Thomas et al., 2001; Matthiessen et al., 2006) are scattered, and not necessarily well-linked to pesticides.This review brings together the known facts about ED effects and effect levels of pesticides on fish, birds and mammals (UK and overseas) incorporating information from recent reviews, e.g. the recent EU reviews by Groshart and Okkerman, 2000; Petersen et al., 2007; as well as literature reviews for publications specifically relevant to pesticides (Ottinger et al 2005a, b).

2.0 An introduction to the endocrine systemsThe endocrine system is responsible for regulating most of the body’s essential functions, including reproduction, secondary sexual characteristics, moulting in birds, fluid balance, growth and response to stress. The endocrine system has been well reviewed e.g. Balthazart (1990), Hiller-Sturmhofel and Bartke 1998, and the basics are outlined below. The primary endocrine organs are the adrenal glands, pituitary, thyroid, parathyroid, ovaries/testes and pancreas. Control of the endocrine system is by the neurosecretory neurons within the hypothalamus and an overview of their action is shown in Figure 2.1. Several classes of hormones exist including steroids, amino acid derivatives, polypeptides and proteins. As a result of their differences in size and chemical properties their mechanisms of action also differ.

1. Steroid hormones have a molecular structure similar to cholesterol. The steroid sex hormones are synthesised from cholesterol through side chain hydrolysis by cytochrome P450 enzymes in families 11, 17, 19, 21 and 27 to form progesterone. Progesterone is further modified to form androstenedione and then to testosterone. Aromatase (another P450 enzyme) catalyses the conversion of androstenedione to estradiol and testosterone to estrone. Cytochrome P450 enzymes in families 1 and 4 also metabolise these steroids to more polar forms for excretion in urine and are inducible by a variety of chemicals. The major enzymes involved in steroid biosynthesis are shown in Figure 2.2. These molecules enter their target cell and interact with the receptors in the cytoplasm or cell nucleus. The hormone-receptor complexes then bind to regions of the DNA thereby regulating the activity of hormone-responsive genes.

2. Hormones which are amino acid derivatives are primarily produced by the thyroid and the adrenal medulla. They enter the cell where they interact with receptor proteins already associated with specific DNA regions.

3. Polypeptide and protein hormones vary from three to several hundred amino acids. They are found primarily in the hypothalamus, pituitary and pancreas. They may be derived from inactive precursors, pro-hormones, and due to their chemical structure cannot enter cells but interact with receptors on the cell surface initiating biochemical changes in either the cell’s membrane or interior eventually modifying the cell’s activity or function.

SID 5 (Rev. 3/06) of 30

To maintain the body’s homeostasis and respond appropriately to the environment hormone production and secretion is tightly regulated through several hormones regulating each other. Thus the hypothalamus secretes releasing hormones (e.g. gonadotrophin releasing hormone (GnRH), thyrothrophin releasing hormone (TRH)) which are transported in the blood to the pituitary gland where they induce the production and secretion of pituitary hormones (e.g. luteinising hormone, follicle stimulating hormone, thyroid stimulating hormone) which are then transported to the target site (the thyroid, adrenal glands or gonads) resulting in the release of hormones (estrogens, progestagens, androgens). There is constant, usually negative, feedback from the target site to the hypothalamus and pituitary to ensure that when the appropriate levels of hormones on the blood are reached the hormone cascade system ceases. In addition, a short feedback loop occurs between the pituitary and the hypothalamus. The gonads (ovaries and testes) synthesise steroid sex hormones necessary for the development and function of the male and female reproductive organs and secondary sex characteristics as well as pregnancy, lactation etc., as well as affecting the metabolism of lipids and carbohydrates, the cardiovascular systems and bone growth and development. There are three types of sex hormones1. Estrogens exert feminising effects; the major estrogen is estradiol, with estrone and estrol produced to a lesser extent. The primary function is in the normal development and functioning of the female gonads. In mammals estrogens also promote the growth of the uterus and mammary glands and regulate the estrus cycle; in birds they also regulate calcium deposition and release from bones, brain development and the stimulation of the egg protein vitellogenin by the liver.2. Progestagens (the major form is progesterone) in mammals affect the uterus in preparation for, and during, pregnancy and in birds cause development of the brood patch and are responsible for nest attentiveness during incubation in altricial species.3. Androgens, the primary androgenic steroid is testosterone, exert masculising effects and stimulate the growth and development of the male gonad and sex organs as well as regulating sperm production.There are many similarities in the endocrine systems of vertebrates but also some significant differences. In birds the default sex, the phenotype to which the embryo will develop in the absence of sex-specific hormones, is male whereas in mammals it is female. Therefore birds require estradiol synthesis to occur to cause differentiation of the gonad into an ovary. Lack of estrogen, even when androgen is also absent, will result in formation of phenotypic males. In mammals the reverse is true, embryos will develop into phenotypic females unless sufficient levels of androgens are present to induce gonadal differentiation into testicular tissue. In mammals, fetoproteins in the developing foetus bind estrogen to protect it from maternal levels of estrogen; birds do not possess this protein and embryos are exposed in ovo to the level of estrogen present in the maternal circulation at the time of egg formation. Due to these differences, xenobiotic estrogens have different effects on birds and mammals during embryonic development. In genetically male birds the presence of excess estrogens results in the gonad beginning to resemble an ovary and may develop an oviduct, but the seminiferous tubules are also retained with a lower number of primordial germ cells and thus no or low spermatogenesis.

Production of thyroid hormones is closely regulated by the hypothalamus and pituitary through a negative feedback cycle with the corticosteroids, testosterone and estrogen depressing production of thyroid hormone. Thyroid hormones are responsible for controlling metabolic rate, regulating body temperature, interacting with growth hormone to determine body size and triggering differentiation/maturation of specific tissues. Follicles in the thyroid gland synthesise amino acid derived hormones producing a large globulin glycoprotein called thyroglobulin. The molecule contains iodinated tyrosine residues that are coupled to form iodinated thyronine of which the major circulating form (approx 90%) contains 4 iodides and is known as thyroxin, or tetraiodothyronine, (T4). T4 is converted to triiodothyroinine (T3) that is currently assumed to be the physiologically active thyroid hormone based on similarities in thyroid function between mammals and birds.

The adrenal glands (small structures on the top of the kidneys) produces primarily corticosteroids – glucocorticoids and mineralocorticoids. The primary glucocorticoid is cortisol (also called hydrocortisone) which helps in the control of carbohydrate, protein and lipid metabolism and also appears to assist in protecting the body against the adverse effects of stress. The primary mineralocorticoid is aldosterone that is important in the control of water and electrolyte balance in conserving sodium and excreting potassium.

3.0 Modes of Action of Pesticidal Endocrine Disrupting Chemicals in In Vitro SystemsFishTable 3.1 lists data for PPPs which have been reported to cause responses in fish-based in vitro test systems. These systems include those using fish tissue explants, sub-cellular organelles, normal cell lines or those transfected with genes of interest. The principles of the assays cover a large range of mechanisms including binding to, or displacement from, hormone receptors; alterations in hormone secretion; changes in gamete maturation rates; steroidogenic enzyme inhibition or induction; and displacement of sex steroids from transport proteins.

Figure 2.1 Overview of the major endocrine pathways (from Hiller-Sturmhofel and Bartke 1998)

SID 5 (Rev. 3/06) of 30

Figure 2.2 Major pathways of steroid hormone synthesis

SID 5 (Rev. 3/06) of 30

Information was found for just 24 PPPs and metabolites, including triazine herbicides, urea herbicides, organochlorine insecticides, organophosphate insecticides, conazole fungicides, pyrimidine fungicides, dithiocarbamate fungicides, and dicarboximide fungicides. Of these 24, 11 are fungicides, 3 are herbicides, and 10 are insecticides or insecticide metabolites.MammalsTable 3.2 lists data for PPPs which have been reported to cause responses in mammalian based in vitro test systems. These systems include those of both human and other mammalian origin and include cell explants, sub-cellular organelles, normal cell lines and those transfected with the genes of interest. The principles of the assays cover receptor binding, steroidogenic enzyme inhibition/ induction and gene expression. The sensitivity of each assay system varies widely depending on the source/species making direct comparison difficult but it does allow the interaction of the pesticide with the system to be identified and the relative potency of differing pesticides in the same assay system to be compared. Information was found for 55 PPPs, including metabolites, of these 12 are fungicides, 35 insecticides, 6 herbicides and 2 plant growth regulators.BirdsThere were no reports identified of effects of PPPs on avian systems in vitro

3.1 TriazinesFishThese are represented in Table 3.1 by just atrazine and prometon, and although the dataset is small, it suggests that triazines may have multiple potential modes of endocrine action in fish. The most potent effect is disruption of the non-genomic action of progestin in fish oocytes which causes blocked oocyte maturation at an atrazine concentration of 10-25 µM (micromolar). Prometon appears to cause increased steroidogenesis of estradiol (E2) at the higher concentration (NOEC) of 52 µM. The third mode of action concerns damage to the adrenocortical (interrenal) cells which prevents the cortisol response to adrenocortical hormone (ACTH) – i.e. damage to the

SID 5 (Rev. 3/06) of 30

stress response, but the IC50 for this effect is very high (>50,000 µM). Overall, insufficient in vitro data are available to draw broad conclusions about the possible endocrine disrupting action of triazines, but multiple modes of action (MOA) seem to occur.MammalsThis group of herbicides is represented in Table 3.2 by 3 compounds, atrazine (the greatest contributor), simazine and propazine. The triazines are reported to have low binding affinity to the human estrogen, androgen and progesterone receptors (Scippo et al 2004) and therefore effects are unlikely to be receptor related. These chemicals affect the synthesis and metabolism of the steroid hormones. Aromatase inhibition/inductionAll (including the atrazine metabolite diaminochlorotriazine) are reported to affect aromatase (CYP19) activity (catalyses the conversion of testosterone to estradiol and androstenedione to estrone) and reported effects range from inhibition to induction. Induction is often a response to inhibition of P450 based systems (of which CYP19 is one), e.g inhibition is reported at 0.01 μM and induction at 1-3 μM for atrazine. Atrazine (0.01 μM) also induces EROD activity (another P450 enzyme) which catalyses the metabolism of estrogen, this probably also occurs following exposure to other triazine herbicides based on their similar structure.

3.2 UreasFishThis group is represented solely by two studies of linuron. These suggest that it is able to displace an androgen from transfected fish androgen receptors with an IC50 of 42 µM, and can therefore be classed as an anti-androgen. However, insufficient in vitro data are available to confirm this observation.MammalsThis group of herbicides is represented by a single chemical linuron which is reported as a androgen receptor antagonist (binding constant 0.758 μM )acting by competitively antagonising the transcriptional activity of the androgen receptor induced by dihydrotestosterone.

3.3 OrganochlorinesFishThis group is represented by a large and varied group of compounds, including chlordecone, DDT and its metabolites, endosulfan, and methoxychlor and its metabolites. The reported modes of action are also varied, including:

antagonism of 17,20 beta, 21-trihydroxy-4-pregnen-3-one induced oocyte maturation (EC50 = 2 µM chlordecone; 100 µM o,p’-DDD); damage to the cortisol response to stress (LOEC = 50,000 µg/l o,p’-DDD; NOEC =50 µM o,p’-DDD; LOEC >100,000 µg/l p,p’-DDD; LOEC = 100,000 µg/l p,p’-DDT; EC50 = 17-38 µM endosulfan); anti-androgenicity (EC50 = 10 µM o,p’-DDD; EC50 = 35 µM p,p’-DDD; EC50 = 16 µM o,p’-DDE; EC50 = 40 µM p,p’-DDE; IC50 = 22 µM p,p’-DDE; EC50 = 35 µM p,p’-DDT; EC50 = 12 µM o,p’-DDT; 0.005-0.008 = relative binding affinity of methoxychlor and metabolites compared with testosterone); estrogenicity (IC50 = 190 µM o,p’-DDT; EC50 = 0.3-6.5 µM methoxychlor and metabolites) displacement of E2 from sex steroid binding protein (IC50 = 950 µM o,p’-DDT);

It is apparent that DDT and its metabolites have a very wide range of endocrine activities of which one of the most potent is anti-androgenicity, which is revealed by its ability to block the androgen receptor. Methoxychlor and its metabolites also have anti-androgenic action although very weakly, but much stronger estrogenic activity. The other type of activity with relatively high potency concerns chlordecone (kepone) which antagonises the induction of oocyte maturation by 17,20 beta, 21-trihydroxy-4-pregnen-3-one. It is of particular interest that two organochlorines (o,p’-DDT and methoxychlor) are simultaneously weak estrogen agonists and weak anti-androgens, and o,p’-DDT is also able to displace E2 from sex steroid binding protein. These multiple modes of action are characteristic of many EDCs, and complicate the interpretation of effects observed in vivo.MammalsA wide range of organochlorines (including metabolites of DDT) are reported to have endocrine related effects in in vitro systems. Many are reported to affect enzyme metabolising systems, particularly those that are P450 dependent. In vitro receptor binding studies on their own, of which there are many reported in Table 3.2, although providing information on relative affinity are difficult to interpret in relation to effects in vivo; binding to a receptor does not provide direct information as to whether the compound acts as an agonist or antagonist. More informative in vitro assays are those such as the transactivation gene responses.Estrogen receptor agonist: dieldrin (5 μM); endosulfan (1 μM)Aromatase inhibition: endosulfan (50 μM)Androgen receptor antagonist: dieldrin (20 μM); endosulfan (20 μM); p,p’DDE (IC50 20μM)

3.4 OrganophosphatesFishOnly one study could be found, showing that fairly high concentrations (EC50 = 233 µM) of the insecticide diazinon are able to damage the cortisol response of steroidogenic interrenal cells. It will by now have been noticed that although damage of the ability to mount a cortisol stress response tends to occur at rather high

SID 5 (Rev. 3/06) of 30

concentrations, it is a property shared by several structurally-unrelated compounds. Furthermore, it has been observed to occur in the field, although most reports do not involve pesticides (Pottinger, 2003).MammalsEffects reported following exposure to organophosphate insecticides relate to two main modes of action: effects on enzyme activity and effects at the receptor/ receptor activation. Inhibition of steroidogenesisTrichlorfon (0.2 μM) has been reported to decrease progesterone production through suppression of steroidogenesis via decreasing the expression of steroidogenic regulatory protein mRNA and inhibition of the P450 side chain cleavage enzyme. Estrogen receptor agonist. Chlorpyrifos (50 μM), parathion (0.1 μM )and tolclofos methyl (5 μM) are reported to act as estrogen receptor agonists but parathion (0.1 μM )also antagonised the effect of estradiol.Androgen receptor antagonistParathion (IC50 0.2 μM) and dichlorvos (20 μM ) have been shown to affect the transcriptional activity following interaction of DHT with the androgen receptor.

3.5 CarbamatesMammalsThe carbamate insecticides have been reported to have a range of modes of actionEstrogen agonists/antagonistsCarbaryl (0.1 μM ), oxamyl (0.1 μM) and methomyl (0.1 μM ) are reported to be a estrogen receptor agonists but an antagonists in the presence of estradiol and progesterone receptor agonists but antagonists in the presence of progesterone. Methiocarb (5 μM) acts as a estrogen receptor agonist.Androgen receptorsMethiocarb (5.8 μM) inhibits androgen induced activation of the androgen receptor by inhibiting binding and decreasing receptor transactivation.Aromatase inductionMethomyl (50 μM), pirimicarb (50 μM) and propamocarb (50 μM ) all increase aromatase activity in exposed human placental microsomes.

3.6 ConazolesFishThe mode of action in fish of this group of fungicides has been fairly intensively studied in vitro. All 8 conazole PPPs that have been tested in vitro are medium to strong aromatase inhibitors (IC50 = 0.01-70 µM (concentration which inhibits 50% of the activity of the enzyme)), implying that they are able to interfere with steroidogenesis, in particular, the synthesis of E2 from testosterone (T), which would be expected to have masculinising (or at least de-feminising) effects in vivo. Both ketoconazole and prochloraz are potently able to inhibit the synthesis of E2 in ovary explants in vitro (ketoconazole NOEC <0.006 µM; prochloraz IC50 = 1.6 µM), but interestingly, they are also able to inhibit T synthesis in the same system, implying inhibitory effects further back in the steroidogenic enzyme cascade. There is also limited evidence that ketoconazole can act as a weak anti-androgen, and that prochloraz can disrupt cell membrane receptivity to peptide hormone-like growth factors, but these effects are less sensitive than aromatase inhibition.MammalsThe conazole fungicides have well characterised effects on in vitro mammalian systems. They are reported to act as aromatase and sterol 14 α demethylase inhibitors (Zarn et al 2003). Aromatase inhibitionThe potency of the inhibition varies widely between chemicals with IC50’s for aromatase inhibition ranging from 0.1 μM for imazalil and prochloraz to 20 μM for penconazole. Androgen receptor antagonistProchloraz is also reported as an androgen receptor antagonist both in decreasing binding of androgens (IC50 60 μM) and inhibiting dihydrotestosterone gene expression (1-10μM ).

3.7 PyrethroidsMammalsA range of pyrethroid insecticides have been tested on in vitro mammalian systems. Estrogenic receptor agonistTrans-allethrin (10 μM), deltamethrin (5 μM), fenvalerate (1-10 μM), l-cyhalothrin (0.1 μM) and sumithrin (10 μM) induce pS2 and MCF7 cell proliferation. Cypermethrin (0.1 μM), permethrin (0.1 μM) and pyrethrin (0.1 μM) synergise the effects of estradiol exposure in cell proliferation assays. Androgen receptor antagonistCypermethrin (IC50 57 μM), deltamethrin (5.8 – 20 μM) and permethrin (IC50 57 μM ) demonstrate anti-androgenic properties by inhibiting the transcriptional induction by dihydrotestosterone Progesterone receptor antagonist

SID 5 (Rev. 3/06) of 30

Trans-allethrin (30 μM ) and fenvalerate (30 μM) have also been shown to act as a progesterone receptor antagonists and fenvalerate (5 μM) has also been shown to inhibit FSH stimulated progesterone production in ovarian luteinising-granulosa cells through effects on the calcium signalling system.

3.8 PyrimidinesFishFenarimol is the only pyrimidine fungicide to have been tested in fish-based systems in vitro. These experiments show that it is a moderately potent aromatase inhibitor (IC50 = 6-18 µM), and like the conazoles, is able to reduce both E2 and T synthesis in ovary explants (NOEC <1 and 10 µM, respectively).MammalsThe pyrimidine fungicide fenarimol has a number of modes of action; being reported to be estrogen receptor agonists (3-5 μM), androgen receptor antagonist (10 μM) and aromatase inhibitor (50 μM).

3.9 DithiocarbamatesFishOnly one in vitro fish-based study (with the fungicide mancozeb) has been conducted of this group, which suggests that a rather high concentration (EC50 = 312 µM) is able to damage the interrenal cortisol response to stress.

3.10 DicarboximidesFishOnly one carboximide PPP (the fungicide vinclozolin) has been studied in in vitro fish-based systems, but several investigations have shown that the parent compound is a moderately strong anti-androgen (lowest IC50 = 12 µM) with an ability to interfere with T synthesis in ovary explants (NOEC for reduced T = 0.04 µM). Its M2 metabolite is an even more potent anti-androgen (IC50 = 0.8-3.5 µM).MammalsAndrogen receptor antagonistThe best known anti-androgenic dicarboximide is vinclozolin. Vinclozolin itself is an anti-androgen (IC50 0.3 μM) but its primary metabolite (M2) is more potent (0.17 μM) competing for AR androgen binding and inhibiting DHT induced transcriptional activation by blocking AR binding to androgen response element DNA. Procymidone (3.16 μM) but not iprodion also acts as androgen receptor antagonist. Estrogenic effectsProcymidone (100 μM ) also has estrogenic effects not by direct effects at the estrogen receptor but via induction of mitogen activated protein kinase through oxygen free radical production.Aromatase inhibitionIprodion is reported to inhibit aromatase activity at 50 μM but as with chlorthalonil above this may be a cytotoxic effect. Vinclozolin (100 μM) induces aromatase activity proposed to be by the inhibition of phosphodiesterase activity.Other receptorsVinclozolin has also been reported to antagonise mineralcorticoid (IC503.2 μM) and progesterone (IC5018 μM) receptors

3.11 Chloronitrile fungicidesMammalsThe sole representative of this group of fungicides reported to have effects in vitro is chlorothalonil. This has been shown to be inhibit aromatase activity at 50 μM but is thought to be related to cytotoxic effects in vitro rather than effects at the enzyme level.

3.12 Chloracetamide herbicidesMammalsThe herbicide alachlor has been reported to bind to both the progesterone (298 μM) and estrogen (240 μM ) receptors but it is unclear whether such binding is agonistic or antagonistic.

3.13 Sulfonurea herbicidesMammalsThe herbicide tribenuron methyl (25 μM) is an estrogen receptor agonist in the MCF7 cell proliferation assay.

SID 5 (Rev. 3/06) of 30

3.14 Triazole plant growth regulatorMammalsPaclobutrazol (20 μM) a a triazole type plant growth retardant which blocks gibberellin biosynthesis has been demonstrated to competitively bind to the androgen receptor

3.15 Hydrazine plant growth regulatorMammalsDaminozid (50 μM) a plant growth regulator, has been shown to be a estrogen receptor agonist increasing the estrogen receptor transactivation gene response.

3.2 ConclusionsThe available in vitro data are clearly rather patchy, with in many cases only a single representative chemical in a class and not all the possible mechanisms of endocrine disruption. For example, no in vitro data have been found for thyroid agonism or antagonism, androgenicity, or anti-estrogenicity. However, the available information shows that even this small selection of PPPs has a wide range of different endocrine MOAs, some of which are highly potent. A significant number of these substances are still in use in the UK, although the organochlorines have been essentially phased out. The best characterised group with endocrine activity in vitro is the conazole fungicides, which consistently cause aromatase inhibition. Another group of fungicides (the pyrimidines) are also aromatase inhibitors. However, despite the clear structure-activity relationships of the conazoles, no other obvious structural alerts emerge from these in vitro data, although there are some structural similarities between the conazoles and the pyrimidine fenarimol (which is also an aromatase inhibitor). Indeed, groups like the organochlorines appear to have multiple MOAs, the structural causes of which are not easy to discern. Furthermore, the ability of widely differing PPPs to interfere with the way in which interrenal cells respond to ACTH (and hence to stress) by synthesising cortisol is not obviously related to chemical structure.

4 In Vivo Effects of Pesticidal EDCs FishThe published data on the endocrine disrupting effects of PPPs on fish in vivo are summarised in Table 4.1. This table shows the most sensitive endocrine-related endpoint in a given study, and also shows the most sensitive apical endpoint in the same study if one was measured. This section will focus primarily on the endocrine endpoints, thus enlarging the information on modes of action available from in vitro studies. A comparison of endocrine with apical endpoints will be made in the next section.As with the in vitro data, data on endocrine disruption obtained with fish in vivo are only available for a limited number (33) of substances, representing 8 herbicides, 5 fungicides and 20 insecticides (including metabolites). The dataset includes 3 triazine herbicides, 1 phenoxy herbicide, 1 phosphonoglycine herbicide, 1 urea herbicide, 1 chloroacetanilide herbicide, 1 dinitroaniline herbicide, 3 conazole fungicides, 1 pyrimidine fungicide, 1 dicarboximide fungicide, 3 pyrethroid insecticides, 2 carbamate insecticides, 11 organochlorine insecticides (and metabolites), and 4 organophosphate insecticides. As such, it covers a wider range of PPP chemical groups than the in vitro dataset, but it still cannot be regarded as representative of all PPP groups in current use.MammalsOnline literature databases were searched for details of potential endocrine disrupting pesticides and the EPAs ECOTOX database was also searched. The results of searches for effects of pesticides on hormone levels, morphology and reproduction are shown in Table 4.2 where data was available for several studies on the same endpoint the most sensitive is reported. The searches identified pesticides, including metabolites, of a range of pesticides including conazole, dithiocarbamate, benzimidazole, phthalimide, pyrimidine and dicarboxiimide fungicides, carbamate, organophosphorus, organochlorine and pyrethroid insecticides, triazine, urea, diphenoxyacid herbicides and plant growth regulators.BirdsOnline literature databases were searched for details of potential endocrine disrupting pesticides and the EPAs ECOTOX database was also searched. The results of searches for effects of pesticides on hormone levels, morphology and reproduction are shown in Table 4.3 where data was available for several studies on the same endpoint the most sensitive is reported. Many of the endpoints were expressed as ppm or % in food which makes actual dose measurements difficult to compare as data on intake are not available. The searches identified identified pesticides, including metabolites, of a range of pesticides including conazole, dimethyldithiocarbamate, benzimidazole and phthalimide fungicides, carbamate, organophosphorus, organochlorine and pyrethroid insecticides and triazine and diphenoxyacid herbicides.

The in vivo actions of each of these groups are described below, and where possible compared with the in vitro data.

SID 5 (Rev. 3/06) of 30

4.1 TriazinesFish This group is represented by atrazine, simazine and prometon. It will be recalled that the in vitro data were insufficient to be certain about MOA, although multiple MOAs seemed to be occurring. The in vivo data suggest that atrazine is able to alter a range of reproductive endocrine variables, including vitellogenin (VTG) induction, gonadosomatic index (GSI), secondary sexual characteristics, and steroid hormone titres. At least some of these effects have been attributed to aromatase induction and generally altered steroidogenesis, although this has only been reported in vitro with prometon. Similar endpoints have not been studied in vivo with either simazine or prometon. However, the most noteworthy point about these effects is that they only occur at elevated concentrations in ambient water (21 d NOEC 50 µg/l in fathead minnows), and so may not be environmentally relevant. An atrazine/simazine mixture is also able to damage the interrenal response to stress in vivo (as it has also been shown to do in vitro), but again at very high ambient concentrations (4 d NOEC = 3500 µg/l).Of greater environmental interest is that both atrazine and simazine are able to interfere with the olfactory response to reproductive prostaglandins in male salmonids with 0.02-81 d NOECs in the range <0.04-1.1 µg/l. The final maturation of salmonid sperm is under the control of 17,20 beta-dihydroxy-4-pregnen-3-one, the release of which in males is controlled by prostaglandins in the urine of ripe females which are detected by the male’s olfactory epithelium. It appears that triazines (and other substances including organophosphates, carbamates and pyrethroids – see below) are able to damage the olfactory epithelium in such a way as to block the prostaglandin signal and thereby inhibit release of milt (sperm). These effects are rather potent and occur at environmentally-relevant concentrations, although no reports exist of similar observations in the field (perhaps due to the difficulty of conducting such studies under field conditions).Finally, no in vivo data have been reported which follow up the observations of disrupted progestin action which prevent oocyte maturation in vitro.MammalsThree triazine herbicides were reported as causing effects on hormone levels in mammals; atrazine affected prolactin (at 100-150 mg/kg), thyroxine (at 35 mg/kg), testosterone (at 50-100mg/kg) and luteinising hormone (at 30 mg/kg) levels in rat and luteinising hormone and estradiol levels (at 1 mg/kg) in the pig. Simazine affected cortisol, progesterone, prolactin (all at 300 mg/kg) and testosterone (at 25 mg/kg) levels in the rat. Metribuzin increased thyroxine levels in rats (10 mg/kg). The effects in vivo related to the serum hormone levels resulting from a proposed direct effect on the CNS causing disruption at the hypothalamus together with decreased serum estradiol resulting from the inhibition of aromatase identified in vitro. Metribuzin in the adult rat also affected the thyroid through interaction with the thyroxine activation/inactivation cycle resulting in increased thyoxine levels. These effects were also reported in the EFSA draft evaluation document for metribuzin where is was reported that metribuzin disturbs the pituitary gland-thyroid gland-liver axis and impairs the process of iodination of T4 to T3.

Changes in the morphology of treated animals can also be indicative of the endocrine effects observed. The triazine herbicides affected the weight of the pituitary gland (LH production), the adrenal gland (cortisol production), the thyroid (T3 and T4 production) and both male and female gonad (testosterone, progesterone and estradiol) and secondary sex glands with dose levels between 1 and 900 mg/kg).

Regarding the impact on reproduction, the timing of exposure of developing female offspring resulted in variation in observed effects. Exposure during gestation resulted in delayed vaginal opening and mammary gland development (4 days 100 mg/kg) whereas exposure via milk from the treated dam delayed mammary gland development only (36 days 100 mg/kg). This effect during suckling may be related to a impact on the levels of hormones or other growth factors expressed in the milk as it was demonstrated that neither atrazine nor its metabolites enter the milk of the exposed dam in more that trace amounts. Other effects included failed oestrus in female pigs (19 days 1mg/kg), reduced numbers of implantations and decreased reproductive success in the house mouse (8-14 days 46 mg/kg) and pseudopregnancy in rats (1 day 300 mg/kg).

Reproductive effects in male rats included delayed puberty and preputial separation (30 days 12.5 mg/kg), decreased sperm count and sperm motility (60 days 15mg/kg) increased lateral prostate weight (4 days 100 mg/kg) and decreased sperm production (60 days 15 mg/kg). Again the delayed preputial separation and increased in lateral prostate weight were only observed in rats exposed during gestation whereas increased lateral prostate weights were observed in male offspring suckling an atrazine treated dam.BirdsThe only triazine herbicide in which effects in birds were reported was atrazine. Exposure resulted in similar effects to those in mammals with effects on luteinising hormone (1000 ppm dietary) and estradiol (1000ppm dietary). Morphological changes after exposure include effects on oviduct weight (10ppm in diet). Reproductive effects included abnormal gonads in progeny (NOEL 3mg/chicken) and decreased 14-day hatchling weight (0.5 mg/kg injected into egg).

SID 5 (Rev. 3/06) of 30

4.2 Phenoxy acidsFishOnly a single in vivo study has been reported, in which 2,4-D produced plasma VTG induction in juvenile trout, with a 7 d NOEC of 16 µg/l. VTG induction in juveniles and males is a biomarker for exposure to estrogen, but there are no other data to suggest that either 2,4-D or its degradation product 2,4-dichlorophenol are estrogen agonists.MammalsThe single representative of the phenoxyacid herbicide class, which was not identified as reported as an endocrine disrupter in vitro in mammals, was 2,4 D which was shown to affect serum thyroxine levels in adult ewes (36 days 10 mg/kg) and rats (91 days 15 mg/kg). This was proposed to be linked to the ability of 2,4D to displace T4 binding in the serum and thus reduce the T4 levels. However both morphological and consequent reproductive effects were also reported with decreased gonad and secondary sex organ weights and abnormal sperm (8-91 days; 1-98 mg/kg resulting in decreased numbers implantations and progeny (8-23 days; 75-98 mg/kg) suggesting wider endocrine disrupting activity.BirdsExposure of chickens to 2,4D was reported to affect growth at 10mg/kg. which may be related to its effects on the thyroid.

4.3 PhosphonoglycinesFishThere are no in vitro data for this group, and the in vivo data concern just 2 studies in a single species (Rhamdia quelen). Both of these show that glyphosate at high concentrations (4-40 d NOEC = 2430-3600 µg/l) is able to interfere with cortisol titres, but in one case a reduced cortisol response to handling stress is reported, while in the other plasma cortisol was shown to increase in response to glyphosate in the absence of stress. The latter effect was accompanied by decreased plasma E2 in females, but the precise mode of action is unknown.MammalsGlyphosate was shown to affect the weight of the developing fetus and the number of embryos resorbed but the mode of action is unclear and may a direct toxic rather than endocrine mediated effect.BirdsGlyphosate was reported to affect testosterone levels in plasma of mallards (5mg/kg) and testis weight in relation to bodyweight (NOEL 100mg/kg).

4.4 UreasFishThe only in vivo study available was with linuron which clearly showed that a moderately low concentration (21 d NOEC = 15 µg/l) is able to reduce the spiggin response of sticklebacks to exogenous androgen exposure. Spiggin is the glue-like protein used by male sticklebacks to build their nest, and is not normally found in females, but it can be induced in females by exposure to androgens which act agonistically at the androgen receptor. Reduced spiggin induction in the presence of linuron therefore implies that it is an anti-androgen. This confirms the observations of in vitro androgen antagonism described above.MammalsTwo urea herbicides were identified as affecting endocrine activity: linuron and diuron. Both were reported to decrease serum testosterone levels in rats (4-30 days; 100-250 mg/kg) and this was linked to secondary sex gland effects (1—98 days; 30-100 mg/kg), abnormal development (10-65 days; 50-100 mg/kg) and decreased offspring weight (testosterone levels in males is related to weight gain) (65 days; 50 mg/kg). In females exposure resulted in effects on relative uterine weight (250 mg/kg 30 days). The reproductive consequences of exposure related to anti-androgen effects in males with decreased delayed preputial separation and areola/nipple retention (9-80 days; 40-50 mg/kg) and sperm counts (30 days 250 mg/kg) (linuron was shown in vitro to be a weak competitive androgen receptor antagonist and has structural similarities of linuron to flutamide, a potent androgen antagonist) but also reported effects on numbers of implantations and progeny at higher exposure levels (9-13 days; 75-125 mg/kg).BirdsThe only reported study with birds was the effect of linuron on the American goldfinch where exposure was reported to result in delayed pre-alternate molt progression by disrupting thyroid function (4 mg/kg for 4 months).

4.5 ThiocarbamatesMammalsSeveral pesticides and their metabolites were identified as causing effects in vivo. The alkylenebis-dithiocarbamates maneb and mancozeb and metam sodium (sodium methyldithiocarbamate) and the metabolite of the ethylene bisdithiocarbamate fungicides - ethylene thiourea (ETU)) showed impacts in vivo on the thyroid related hormones thyrotropin and thyroxine and on the cortisol levels regulated by the adrenal gland (<1-30 dats; 10-1000 mg/kg). Metam sodium was also shown to affect the regulation of luteinising hormone by the pituitary gland (<1 day 500 mg/kg).

SID 5 (Rev. 3/06) of 30

At the organ level the thiocarbamates and the metabolite ETU were shown the affect the weight of the gonads and secondary sex organs (14-15 days; 40-700 mg/kg), pituitary (14 days 40 mg/kg), adrenal gland (14 days 40 mg/kg) and thyroid (30 days 1500 mg/kg).The reproductive consequences of these effects were demonstrated by decreased numbers of implantations and litter size (8-14 days; 23-4000 mg/kg) and sperm counts (11 days; 800 mg/kg).BirdsThe dimethyldithiocarbamates ferbam, ziram and thiram and metam sodium (sodium methyldithiocarbamate) (0.43 mmol/kg food) were reported to affect growth in chickens and also at the same rates to affect normal development of the tibia in offspring. Such growth effects may be related to the same mechanism of effects on thyroid hormones identified for mammals. Ethylene thiourea was also reported to affect growth in birds. Although food intake rate will affect growth where intake is normal growth effects related to thyroid mediated changes cannot be excluded.

4.6 ChloroacetanilidesFishA single in vivo study, at a high concentration of metazachlor (20 d NOEC = 356 µg/l), produced an increase in the whole-body titre of VTG in juvenile zebrafish, suggesting that this substance can act as an estrogen agonist. No in vitro data are available to support this conclusion. This NOEC is probably not environmentally relevant given that metazachlor is not usually applied at rates in excess of 1.25 kg/ha.MammalsIn vitro the chloracetanilides were shown to bind the estrogen and progesterone receptors but the consequences of this were unclear. In vivo studies did not relate to reproductive effects but showed that administration of alachlor (120 day 126 mg/kg) to male rats resulted in increasing TSH levels, and thyroid neoplasia due to increased metabolism of T4 via hepatic enzyme conjugation.

4.7 DinitroanilinesFishTwo in vivo studies in Japanese medaka show that oryzalin can induce the protein choriogenin in males (part of the oocyte envelope, the induction of which in liver is under estrogenic control). It can also cause ovotestis and damaged spermatogenesis in males and ovarian hyperplasia in females. These data suggest that oryzalin can act as an estrogen agonist, but the 3 d NOEC in one study was very high (2200 µg/l), while in the other there was an unbounded 21 d LOEC of 400 µg/l. No in vitro data are available to support these observations.

There are no reported studies of trifluralin, either in vitro or in vivo, which attempt to investigate whether it can act as an estrogen agonist like oryzalin. The estrogenic properties of oryzalin are not easily predicted from its structure, but it has clear structural similarities with trifluralin which suggest that the latter may also have estrogenic activity at high concentrations.MammalsThere were no reports identified on the in vitro effects of the dinitroanilines and the single representative for which a report was found following in vivo exposure trifluralin, is shown to results in increased serum cortisol and estradiol and decreased serum LH in mature female sheep (36 days 17.5 mg/kg) suggesting it has estrogenic properties.BirdsInjection of pendimethalin (1.25%) into chicken eggs resulted in decreased embryo weight but there are no other studies reported with this class on which to base any conclusions on mode of action..

4.8 PhthalamidesMammalsCaptan is the sole representative of this class of fungicides. Many of the in vivo studies related to the levels of testosterone analogues in serum and a wide range of tissues. Exposure resulted in fetal abnormalities (13-14 days 100 mg/kg) and abnormal sperm (5 days 200 mg/kg) and reproductive effects included decreased reproductive success at 100 mg/kg (1-13 days).BirdsWeight of hatchlings and deformities were reported after captan was injected into eggs (5-36 mg/kg).

4.9 ConazolesFishIt will be recalled that in vitro data for several conazole fungicides show that they are all able to act as aromatase inhibitors. There is also limited in vitro data to suggest that they can interfere with earlier steps in the reproductive steroidogenesis cascade, and ketoconazole shows some anti-androgenic activity. The in vivo data for ketoconazole and prochloraz confirm the picture of perturbed steroidogenesis. With a 21 d NOEC of 6 µg/l and below, ketoconazole increases the level of steroidogenic acute regulatory protein (StAR) in fathead minnow testes and decreases ovarian T synthesis. Prochloraz reduces VTG titres in female fatheads (21 d NOEC = 30 µg/l), a clear sign of aromatase inhibition, and has also been shown to retard testicular development in rainbow trout (14-21 d NOEC = 10 µg/l).

SID 5 (Rev. 3/06) of 30

On the other hand, cyproconazole and prochloraz have both been shown to induce the VTG gene or VTG itself, with a 5 d NOEC of >4377 µg/l in zebrafish embryos, and a 60 d NOEC of <16 µg/l in adult male zebrafish, respectively. The MOA of this effect is unclear.MammalsSeven conazole fungicides were identified as causing endocrine effects and in vivo effects could be directly related to their in vitro activities. In vitro the conazoles were shown to inhibit aromatase activity and act as androgen receptor antagonists. In vivo these properties were manifested in decreased serum testosterone (5-19 days; 16-400 mg/kg) and thyroxine levels (3 days 300 mg/kg) and consequent decrease in androgen sensitive organ weights (10 days; 125 mg/kg) in rats.

Males exposed in utero resulted in an increase in late and very late resorptions of male embryos (epoxiconoazole14 day 50 mg/kg), increased anogenital distance and other feminised features (ketoconazole 14 day 50 mg/kg; prochloraz 4 day 125 mg/kg; tebuconazole 30 day 50 mg/kg, iprodione 29 day 100 mg/kg), reduced testosterone levels (myclobutanil 14 days 150 mg/kg; ketoconazole 14 day 50 mg/kg; prochloraz 10day 30 mg/kg), increased progesterone and 17 α hydroxyprogesterone in the testes (tebuconazole 30 days 50 mg/kg) and increased time to parturition (ketoconazole 10 day 25 mg/kg). In females virilisation was observed following exposure in utero, e.g. increased anogenital distance (epoxiconazole 30 day 15 mg/kg; tebuconazole 30 day 100 mg/kg). BirdsThe conazole fungicides are represented by just 3 pesticides and all studies were in quail. The effects of propiconazole (400 mg/kg/day) on androstenedione levels in microsomes prepared from treated birds can be related to its inhibitory effect on aromatase activity. Epoxiconazole (50ppm) reduced production of spermatids (similar to effects observed in mammals) and triadimefon (80 mg/kg/day) affected weight of treated birds which may be an endocrine effect related to feminisation of males.

4.10 BenzimidazolesMammalsEffects in vitro were not identified for benzimidazole fungicides but in vivo exposure resulted following in vivo exposure of rats in a range of effects depending on age. As with the conazoles the benzimidazoles appear to affect steroidogenesis with decreased estradiol and progesterone levels (5-9 days 25-100 mg/kg) and may act through the pituitary based on the decreased luteinizing hormone and follicle stimulating hormone levels observed (36-79 days; 45-1000 mg/kg). Carbendazim was also reported to act on the thyroid/parathyroid/adrenal glands with increased T3 levels (105 days; 300 mg/kg) and decreased adrenal gland weight (105 days 600 mg/kg). The benzimidazoles may also act as an androgen receptor antagonist resulting in decreased sperm production, testes and cauda epididymal weight in exposed immature male rats (5-250 days; 25-900 mg/kg). Exposure of immature females resulted in decreased ovarian, uterine, placenta and endometrial weight (5-14 days, 31-1000 mg/kg).

Reported reproductive effects in females included decreased numbers of implantations and progeny (6-67 days 31-400 mg/kg) and in males decreased sperm counts were reported(28-84 days; 45-900 mg/kg).BirdsCarbendazim injected into eggs has been reported to affected fetal weight but here were insufficient information to determine whether this was an endocrine or a toxic effect.

4.11 PyrimidinesFishTwo studies of the fungicide fenarimol have given rather high 21 d NOECs of <100 µg/l for an increase in plasma E2, and 177 µg/l for reduced plasma VTG, in female fathead minnows. Reduced VTG is consistent with the aromatase inhibition MOA reported from in vitro studies. However, it is harder to explain increased plasma E2, although fenarimol is suspected of being able to interfere with several enzymes in the steroidogenic chain.MammalsIn vitro fenarimol was shown to be an estrogen receptor agonist and androgen antagonist and an aromatase inhibitor. In vivo exposure resulted in changes in serum estradiol, increased luteinising hormone and FSH levels and lower levels of T3 (4 days; 200mg/kg). These changes may be linked to the observed increase oin uterine weight (4 days, 200 mg/kg) and effects on pregnancy rates (28 days; 35 mg/kg).

4.12 DicarboximidesFishAs will already have been noted from in vitro studies, vinclozolin and its two metabolites M1 and M2 are quite potent anti-androgens, and several in vivo experiments have given results which are consistent with this. When dosed in food to fish, it gives 30-98 d NOECs of <0.1 to >100 µg/mg food for effects such as reduced sperm count, reduced male GSI, and abnormal testicular structure. A similar picture results from exposures via the ambient water, with 4-90 d NOECs of <25 to <1000 µg/l for endpoints such as decreased titres of androgen-induced spiggin, abnormal spermatogenesis, increased male E2, decreased female GSI, increased ovarian T

SID 5 (Rev. 3/06) of 30

synthesis, and upregulation of male luteinising hormone receptor mRNA. Most of these effects occur, however, at concentrations which are probably above the environmentally relevant range.MammalsThe in vivo effects of procymidine and vinclozolin reflect the properties of these chemicals in vitro as androgen receptor antagonists and aromatase inducers. Exposure in vivo to procymidone resulted in increased serum LH and testosterone and isolated Leydig cells showed an enhanced capacity to produce testosterone (90 days 26.4 mg/kg). Exposure to vinclozolin resulted in increased TSH and decreased thyroxin levels (28 days; 12.5 mg/kg). Organ changes following exposure of males exposure in utero were decreased gonad and secondary sex organ weights (28-60 days;12.5-100 mg/kg) and in females increased uterine weight (4 days 200mg/kg).

Reproductive effects included feminisation of males including atrophied testes (6-14 days; 25-200 mg/kg) and in mature female rats a prolonged oestrus cycle (28 days 200mg/kg).BirdsExposure of Japanese quail to vinclozolin (25 mg/kg) was shown to be anti-androgenic with changes in GnRH levels in the preoptic area and median eminence of the brain of exposed birds.

4.13 PyrethroidsFishIn vivo fish data are available for 3 synthetic pyrethroids, bifenthrin, cypermethrin and permethrin, but there are no supporting in vitro studies. It has been shown that both bifenthrin and permethrin are able to induce VTG or its genes (2-10 d NOEC = 0.1 - <10 µg/l) in male fish, indicating that they have estrogenic action. This effect has not been demonstrated with cypermethrin, although a 45 d NOEC of <20 µg/l has been reported for reduced female GSI and plasma E2, and reduced male GSI, sperm motility and damaged spermatogenesis. These effects in males could also be due to estrogenic action, but the MOA of the effects in females in unknown.

Two studies have also shown that cypermethrin can impair the olfactory response and milt production of males resulting from exposure to female reproductive prostaglandins, with 4-5 d NOECs of <0.004-0.1 µg/l. This is clearly a very sensitive effect occurring at concentrations that are probably environmentally relevant, but as with similar effects for other PPPs, has not yet been observed in the field.MammalsEight pyrethroids have been reported to shown endocrine related effects in vivo. The androgen receptor antagonist properties identified in vitro were observed in male rats with effects on testosterone levels in plasma (42-123 days; 10-35 mg/kg) and also effects on TSH, T3 and T4 levels (15-21 days; 0.2-120 mg/kg). Exposure of male rats in utero resulted in decreased gonad and secondary sex organ weight (42-123 days; 10-70 mg/kg) and in females decreased uterine and vaginal weight at low doses (3 days 5 mg/kg) and increased at high doses (3 days 200 mg/kg) confirming the synergism of estradiol observed in vitro. Reproductive effects included decreased sexual behaviour (9-10 days 10 mg/kg), decreased numbers of progeny(17-60 days; 10-80 mg/kg) and effects on sperm counts and motility (30-84 days; 20-35 mg/kg).BirdsThree pyrethroids have been reported to show endocrine effects in birds. The majority were morphological effects and these were all related to growth. Injection of cypermethrin (25mg/kg) into eggs was reported to affect egg weight.

4.14 CarbamatesFishStudies of both carbaryl and carbofuran have shown them to cause reductions in the plasma thyroxine (T4) titres of catfish with 4-16 d NOEC values for carbaryl of <5000-12,000 µg/l, and 120-180 d NOECs of <4.5-5.0 µg/l for carbofuran. Reduced tri-iodothyronine (T3) is also caused by carbaryl, and the effects of carbofuran are accompanied by thyroid hypertrophy and hyperplasia. Higher concentrations of carbofuran additionally reduce E2 and VTG levels in pre-spawning or spawning female catfish. The modes of action underlying these effects are unknown, and no in vitro data are available.

Carbofuran is another PPP that interferes with the olfactory sensitivity of male salmonids to female prostaglandins, and their subsequent ability to release milt, with a reported 5 d NOEC of 1.1 µg/l. In company with other substances causing this effect, the latter occurs at rather low concentrations which are probably environmentally relevant. Precisely how carbofuran and other PPPs damage the olfactory epithelium is unknown.MammalsThree carbamates were represented in in vivo studies. Exposure increased progesterone, cortisol and estradiol levels and decrease testosterone levels (1 day; 1.5 mg/kg) with the mechanism though to be hypercholinergic stress which may be supported by the effects observed at higher levels (90 days; 17mg/kg) on FSH, LH and prolactin levels as effects were transient and subsided as the animals recovered from the signs of toxicity suggesting the former mechanism. Effects were also observed on thyroxine levels (36 days; 0.3 mg/kg).

SID 5 (Rev. 3/06) of 30

The hormone changes were supported by observations of effects on the testes and sperm (60-90 days; 0.2-50 mg/kg) and on both ovarian and placental weight (10-16 days; 0.7-200 mg/kg). Reproductive effects included decrease reproductive success and decreased numbers of progency (3-334 days; 0.1- 200 mg/kg)BirdsThere are a large number of reports of effects of carbamates on morphology and reproduction in birds but these all relate to 3 chemicals (aldicarb, carbaryl, carbofuran). Most morphological changes related to bodyweight but as these anticholinesterase pesticides can reduce food intake and therefore the link between growth rate and endocrine changes is unclear but may be thyroid related as above. However carbofuran also decreased the weight of adrenal glands (0.8 mg/kg) in house sparrow and reproductive effects were reported including impacts on fertility (carbaryl 200ppm), germ cell count (carbaryl >10mg/kg injected into eggs) and numbers of progeny (carbofuran 200ppm) suggesting hypercholinergic effects may also result in effects in birds.

4.15 OrganochlorinesFishIt will be recalled that organochlorines have a wide array of endocrine MOAs, depending on their precise structure, including androgen antagonism, damaged cortisol responses to stress, damaged oocyte maturation and estrogen agonism.

Estrogen agonismo,p’-DDT is an estrogen agonist in vitro, and this mechanism clearly also occurs in vivo, with effects such as ovotestis induction, male to female sex reversal, E2 and VTG induction, reduced male GSI and abnormal testicular development. The long-term NOECs for these effects range down to 0.14 µg/l for exposure via ambient water, and to <30 mg/kg body weight by i.p. injection. Similar effects appear to be caused by endosulfan, with long-term NOECs down to below 0.6 µg/l, and by methoxychlor and its metabolites, with 13-42 d NOECs down to 0.5 µg/l.Androgen antagonismEffects attributable to this MOA seem to be caused by p,p’-DDE, which is associated with abnormal secondary sexual characteristics, sex ratios, GSI, sperm counts and spermatogenesis, giving long-term NOECs as low as < 0.01 µg/mg food and >0.1 µg/l. However, the picture is somewhat confused because p,p’-DDE is also a weak estrogen.Damaged cortisol responseThis has only been observed in vivo with o,p’DDD (7-14 d NOEC <5 mg/kg body weight), although in vitro data suggests that it may also occur with o,p’-DDT and endosulfan.Antagonism of final oocyte maturationThis has been seen with o,p’-DDD in vitro, but not with any organochlorines in vivo.Altered steroidogenesisWhile not observed in vitro, this seems to occur with some organochlorines in vivo, although precise MOAs are uncertain. For example, p,p’-DDE causes altered expression of steroid hormone synthesis and metabolism genes (120 d NOEC < 5.3 µg/g food) in largemouth bass, and increased T4 in salmon parr (5 d NOEC <10 µg/l); dieldrin reduces male 11-ketotestosterone (11-KT) and female E2 titres in largemouth bass (120 d NOEC <0.04-<0.4 mg/kg feed); endosulfan increases plasma T4 and reduces T3 in catfish (4-16 d NOEC = 8-15 µg/l), but can also upregulate T4 in tilapia (21-35 d NOEC = 0.001 µg/g food); lindane can cause decreases in plasma T, 11-KT and E2 in catfish (28 d NOEC <100 µg/l); and finally mirex can reduce serum T3 and T4 in trout (56-84 d NOEC <5µg/g food).Thyroid effectsApart from the perturbations of T3 and T4 described above, a number of other thyroid-related effects of organochlorines have been observed. These include reduced thyroid stimulating hormone and reduced expression of thyroid receptors caused by DDE in salmon parr (5 d NOEC <10 µg/l); altered thyroid epithelial cell height in various species with long term NOECs for DDT down to <1 µg/l; thyroid hypertrophy and hyperplasia caused by endosulfan in tilapia (20 d NOEC = 1 µg/l) and mosquitofish (35 d NOEC <0.1 µg/l); and reduced thyroid epithelial cell height caused by endrin in goldfish (104 d NOEC <143 µg/kg food). Organochlorines have been implicated in thyroid abnormalities seen in some wild fish populations, for instance in the Great Lakes (Jobling and Tyler, 2003), but the MOAs of the effects observed in the laboratory are uncertain.

Overall, as was expected from the in vitro data, organochlorines display a very wide range of endocrine effects in vivo, and several appear to have multiple MOAs. Some effects clearly occur at environmentally-relevant concentrations, partly due to the well-known propensity of many organochlorines to bioconcentrate strongly in fish. Others do not appear sufficiently potent to cause environmental problems, at least not in areas where organochlorines are no longer in use.MammalsThere is a large literature on the endocrine effects of organochlorine insecticides in vivo. There are reviews of these data, e.g. Tiemann’s 2008 detailed review on the impact on the female reproductive tract through effects on steroidogenesis. This includes impacts on folliculogenesis, ovulation, fertilization and implantation in a wide range of species and demonstrates that organochlorine compounds can also substitute for estradiol in regulating the microanatomy of the female reproductive tract.

SID 5 (Rev. 3/06) of 30

The effects of organochlorines in vivo may be related to the in vitro observation of aromatase inhibition, anti-androgen and estrogenic activity of organochlorines. Exposure in vivo resulted in changes in estradiol, testosterone and luteinising hormone levels (15-36 days; 2.5-10 mg/kg) and also affected TSH, T3 and T4 (21-36 days; 0.66-2.5 mg/kg), considered to be due to the displacement of T4 binding in serum. These were related to changes in gonad and secondary sex gland weights in both males and females (4-110 days; 1-200 mg/kg) and the consequent reproductive effects included decreased sperm counts (15-140 days;1-5 mg/kg), decreased pregnancy rates and number of progeny (14-23 days; 1-6 mg/kg) and feminised males (4-196 days 1-100mg/kg).

The chlorinated disinfectant pentachlorphenol demonstrates similar endocrine disrupting activities to the insecticides. In sheep it has been shown to cause hypothyroidism in males and females exposed from conception (196-469 day LOEC 1 mg /kg) to displace T4 from binding in serum (adult 36 day LOEC 2 mg/kg), alter morphology of the thyroid gland (252 day LOEC 1 mg/kg). It has also been reported to increase the severity of oviductal epithelial cysts (36 days 2 mg/kg) and result in resorption of embryos (14-23 days; 30-80 mg/kg).BirdsMethoxychlor in the quail was demonstrated to change GnRH concentrations within the brain after injection into developing eggs (50ppm). Many of the morphological changes related to bodyweight effects which may be food intake related and/or an effect on thyroid hormone levels described above. However a range of estrogenic effects expected on exposure to organochlorines were reported. Methoxychlor (1000nmol/g) significantly increased oviduct mass in the zebra finch, dicofol/kelthane affected eggshell thickness/strength in American kestrel, ringed turtle dove, mallard and Eastern screech owl, clutch production in mallard and ringed turtle dove and progeny weight in mallard.

4.16 OrganophosphatesFishThe sparse in vitro data suggest that organophosphates may be able to damage the cortisol response to stress. This has indeed been observed in vivo, but only with catfish exposed to high concentrations of methyl parathion (4 d NOEC = <800 µg/l). The other organophosphates appear to cause a range of different endocrine effects whose MOAs are not clear. Diazinon, for example, causes ovarian necrosis and reduced plasma E2 (at high concentrations that might not constitute direct endocrine disruption), but is a potent disrupter of the ability of male salmonids to detect female reproductive prostaglandins and thereby to trigger milt expression (5 d NOEC <0.3 µg/l). Fenitrothion causes both thyroid follicular atrophy at high concentrations in snakeheads (120 d NOEC <1500 µg/l), but is also a fairly potent androgen antagonist in sticklebacks (21 d NOEC <15 µg/l). Finally, malathion at high concentrations causes thyroid hyperplasia and hypertrophy in snakeheads (180 d NOEC <2000-4000 µg/l) and reduced plasma T4 in catfish (16 d NOEC <3500-7000 µg/l), but at low concentrations it reduces VTG titres in female Japanese medaka (21 d NOEC = 2.8 µg/l).

The MOAs of these effects are not known, but it seems that while some may be caused by systemic toxicity at relatively high concentrations that are not in any case environmentally relevant, others (e.g. damaged milt priming and reduced VTG titres) are true examples of endocrine disruption occurring at much lower concentrations which are in the environmental range.MammalsA wide range of organophosphate pesticides have been reported to result in endocrine related effects in vivo. These anticholinesterase pesticides may be expected to affect the hypothalamus and pituitary through hypercholinergic stress in the same manner as the carbamates as well as through the androgen receptor antagonist properties identified in vitro. This can be identified in the in vivo studies through changes in the levels of follicle stimulating hormone (14-21 days; 5-20 mg/kg), ACTH (1 day; 5mg/kg), LH (<1 day; 50 mg/kg), TSH (21 days 0.06 mg/kg) as well as a wide range of hormones expressed in the gonads, thyroid and adrenal glands (<1- 91 days;0.06-500 mg/kg).

Impacts of the changing hormone levels are shown through changes in the weights of endocrine glands, e.g. thyroid (16-28 days; 0.16-28 mg/kg), adrenal gland (16-85 days; 1-44mg/kg), ovaries and female secondary sex organs (8-20 days; 1-500 mg/kg) and testes and male secondary sex organs (7-90 days; 4-50 mg/kg). The reproductive impacts of exposure included decreased sperm counts (28-130 days; 0.5-15 mg/kg), effects of fertility (2-273 days; 1-5000 mg/kg).BirdsA wide range of studies were reported on the effects of exposure to organophosphorus pesticides. Effects on hormone levels included impacts on T3 and T4 and corticosterone levels in plasma following exposure to dimethoate (2-10 mg/kg bodywt), methyl parathion (2.25 mg/kg bodywt) and malathion (75 mg/kg bodywt) and included a study on the American kestrel as well as chickens. Morphological changes included a wide range of effects of growth and development (although many of the effects on bodyweight may have been due to anorexic effects) but also seminiferous tubule development (methyl parathion 0.05mg/kg/day) and thymus weight (malathion 230mg/kg). Effects on reproductive measures included deformities in young (methyl parathion, diazinon) and on the development of the shell (diazinon, malathion, methyl parathion) but many of these effects are likely to be due to parental toxicity rather than endocrine effects.

SID 5 (Rev. 3/06) of 30

4.2 ConclusionIn summary, although only relatively few PPPs have been examined for endocrine effects in vivo in fish, the available data provide examples of most of the known endocrine MOAs, some of which clearly cause adverse effects and occur at concentrations which may be environmentally relevant. It is important to point out that some of these effects (e.g. thyroid disruption; impaired olfactory responses to prostaglandin-triggered priming of milt maturation) were not predicted on the basis of in vitro tests, thus confirming the value of in vivo studies for describing endocrine disruption. In other cases, however, in vivo effects such as impaired steroidogenesis, estrogen agonism, and androgen antagonism were indeed detected in vitro.

A far wider range of chemicals have been tested for endocrine disrupting activity in mammals but again the lack of a link between in vitro and in vivo results is evident due to the integrated response of the animal in manifesting effects.

However, the most important question is not whether a PPP is an endocrine disrupter, but whether the adverse effects which result from this disruption occur at concentrations below the NOECs derived from conventional toxicity tests. In other words, are current risk assessments of PPPs underestimating the risks in some cases? This issue is tackled in the next section.

5. Comparison of Endocrine and Apical (Whole Organism) Effect Levels of Pesticides FishFor each PPP with known endocrine disrupting action in vivo in fish, Table 5.1 lists the most sensitive known apical (i.e. whole-organism) effect and effect level, usually expressed as a chronic NOEC. In the rare cases where the results of a fish full life cycle (FFLC) test are available, these are also listed (irrespective of the reported effect level) because FFLCs are the only fish-based toxicity tests in general regulatory use that are expected to be sensitive to most of the effects of EDCs (see the next section for a fuller discussion of this issue). Table 5.1 also lists the most sensitive endocrine endpoint for each PPP, taken from Table 4.1, to enable a direct assessment of relative sensitivities.

Of the 33 substances listed in Table 5.1, endocrine and apical endpoints cannot be compared in 4 cases (due to absence of data or to incompatible exposure units). Of the remaining 29 substances, the most sensitive reported endocrine-related endpoints were more sensitive than the most sensitive apical endpoints in 10 cases (34.5%). In another 12 cases (41.4%), there was no obvious difference between endocrine and apical endpoints (operationally defined as a difference of less than a factor of 5). In the remaining 7 cases (24.1%), the reported endocrine endpoints were less sensitive than the apical data. These comparisons are not very precise because some data are reported as < or > values, but they give a reasonably reliable summary which suggests that apical data for about one third of the endocrine-active PPPs for which data exist may underestimate the potential extent of impact. These comparisons are summarised in Table 5.2.

However, before drawing firm conclusions, it must be stated that the database is incomplete in several cases, either due to a small range of apical studies, or endocrine studies, or both. It is nevertheless worth examining the comparisons in more detail in an attempt to draw wider lessons.

The first point to make is that there is no obvious relationship between high relative sensitivity of endocrine endpoints and the type of PPP under consideration. Where members of more than one chemical group are represented, they seem to be scattered more or less randomly across the columns in Table 5.2, and the same seems to apply to the different endocrine modes of action.

Considering the endocrine endpoints themselves, it is noteworthy that in 4 cases (atrazine, carbofuran, cypermethrin and simazine), the high relative sensitivity compared with apical effects is due to the damaged ability in male salmonids to detect and respond to reproductive prostaglandins emanating from females. In the other case in which this endpoint is flagged up (diazinon) there is no difference in sensitivity between it and the apical endpoint (damaged homing and anti-predator behaviour). This effect on pheromonal signalling does not seem to be closely related to chemical structure, but it is potentially of great importance because male salmonids cannot release mature sperm if they are unable to detect the relevant prostaglandin. It is probable that effects of this type are not fully, or at all, expressed in standard fish reproduction experiments, and indeed it is noteworthy that the 255 d NOEC for atrazine derived from a fathead minnow full life cycle test is 250 µg/l, a factor of >6000 times greater than the effect on olfaction and milt maturation in salmonids. It is doubtful whether this large difference can simply be explained by inter-species differences in sensitivity. The effect on olfaction of carbofuran is also more sensitive (x70) than the results (growth and fecundity) from a short-term reproduction experiment. These effects on pheromonal signalling have not been studied with the other PPPs considered in this review, but it would clearly be interesting to do so.

Considering the other endocrine endpoints with high relative sensitivity compared to apical effects, there appears to be a range of biochemical factors which are responsive at low PPP exposure concentrations, including abnormal steroid hormone titres (dieldrin), altered vitellogenin titres (metazachlor, permethrin), altered induced

SID 5 (Rev. 3/06) of 30

spiggin titres (fenitrothion), abnormal histologies of endocrine tissues (endosulfan), and abnormal secondary sexual characteristics (prometon). It might be expected that any apical effects resulting from these alterations in endocrine biomarkers (and in others, such as altered gonado-somatic index, biased sex ratio of offspring, induction of ovotestis, damaged spermatogenesis and reduced sperm count etc.) would be expressed in fish full life cycle (FFLC) experiments as changes in growth, reproductive output or offspring quality, but it should be noted that such experiments are not routinely requested by regulatory authorities or conducted by notifiers, and are largely missing from the published data. It might also be expected that partial life cycle experiments such as the fish short-term reproduction test (FSTRT) or the fish sexual development test (FSDT) would be at least partly responsive to biochemical changes of the types listed above, but these are also not routinely conducted in support of PPP registration, although some FSTRT test data appear in Table 5.1.

To date, there have been few overt attempts to assess whether changes in biomarkers of endocrine disruption are well-correlated with adverse apical changes, although one might expect such correlations to exist. An exception to this concerns induction of the yolk precursor protein vitellogenin (VTG) in response to exposure to oestrogens and oestrogen mimics. Several papers have shown this endpoint to be well correlated with adverse reproductive and other apical effects (both in the laboratory and the field), and that NOECs for VTG induction tend to occur at similar concentrations to NOECs for those apical effects (Folmar et al., 2001; Miller et al., 2007; Kidd et al., 2007; Thorpe et al., 2007). It has also been shown that the induction of ovotestis in male roach (Rutilus rutilus) via oestrogen exposure in the field is well-correlated with reductions in gamete quality measured by in vitro fertilisation experiments (Jobling et al., 2002).

In summary, the relatively small available dataset of comparative sensitivities (endocrine vs. apical) suggests that some PPPs are able to exert endocrine-related changes whose apical effects have not (yet) been detected in standard chronic toxicity tests with fish. It seems likely that this is due to the inability of those standard tests to express the full gamut of possible adverse effects. The following section therefore considers the availability of fish-based tests with sensitivity to EDCs, and discusses possible strategies for their deployment.

MammalsUnlike fish, reproduction studies are routinely undertaken in mammals with, in some cases, data available from 2-generation studies in rats. The available data are listed in Table 5.3 and show the most sensitive endocrine endpoint and the reproduction test, single or multi-generation NOEC based on the most sensitive endpoint listed in the EFSA document or EPA ECOTOX database. Of the substances listed in Table 5.3, endocrine and apical endpoints the data can be compared for 29 substances, in 20 cases (70%), there was no obvious difference between endocrine and apical endpoints (operationally defined as a difference of less than a factor of 5 due to differences such as normal variations between laboratories). The most sensitive reported endocrine-related endpoints were more sensitive than the most sensitive apical endpoints in 3 cases (10%). In the remaining 6 cases (20%), the reported endocrine endpoints were less sensitive than the apical data. This suggests that the reproduction and multi-generation studies undertaken in the rat are likely to identify the majority of cases in which endocrine disruption occurs and only in 10% of cases apical data may underestimate the potential extent of impact.

The wide range of substances and studies identified in the review of the database and literature does, however, provide the opportunity to compare the relative sensitivity of the various endpoints. The distribution of NOEL for hormone, morphological and reproductive endpoints were compared across all substances and are shown in Figure 5.1. There were 172 hormone level, 191 morphological and 186 reproductive data points available. The figure shows the relative sensitivity of hormone changes compared to morphological or reproductive effects.

This suggests that, as with the data for fish, the endocrine based endpoints are more sensitive than standard reproductive test endpoints with the difference in sensitivity at around 5-10 fold. However there are differences between the classes of pesticides. When there were sufficient datapoints to compare sensitivity these were plotted using the same approach and are shown in Figure 5.1. For the organophosphorus and carbamate pesticides, which are though to act through anticholinergic activity in the hypothalamus and pituitary the difference in sensitivity is small. However for the phthalimide fungicide captan which has many studies hormones are shown to be far more sensitive than morphological or reproductive effects.

BirdsAs data extracted were often based on different measures of exposure the most sensitive endpoints in the hormone, morphology and reproduction categories identified above were listed together with the reproduction test NOEC where these were available (Table 5.4). Reproduction studies are routinely undertaken in birds when exposure may occur during the breeding period but many of the pesticides were withdrawn before such data were published via web based databases. The available reproduction NOEC (apical endpoint) used in risk assessment were extracted from AFSSA and EFSA databases and listed in Table 5.4 (the EPA Ecotox data were already included under the previous listed reproduction endpoints in Table 4.3). Of the substances listed in Table 5.4, endocrine and apical reproductive endpoints can be compared for 29 substances, However, in most cases the units of exposure and/or species tested differ making direct comparison

SID 5 (Rev. 3/06) of 30

difficult. Of the 26 cases where some direct comparison was possible; in 9 cases (35%) the reproduction endpoints were less sensitive, in 13 cases (50%) they were similar (within 5-fold) and in the remaining 4 cases they were more sensitive (15%). Of the studies only 18 included a reproduction endpoint with quail or mallard and of these in 8 cases (44%) the reproduction endpoints were less sensitive, in 7 cases (39%) they were similar (within 5-fold) and in the remaining 3 cases they were more sensitive (17%). This suggests that the data from one generation quail and mallard studies are likely to underestimate the cases in which endocrine disruption occurs and 44% of cases apical data may underestimate the potential extent of impact. As almost all the data was collected from precocial species (quail, mallard, chicken) with only few datapoints from altricial species (owl, kestrel, goldfinch) there is no basis on which to make assumptions about ability to extrapolate data between precocial and altricial species.

Figure 5.1 Comparison of the distribution of the NOEL for hormone, morphological and reproductive endpoints in mammalian in vivo studies for all pesticides (A); organophosphorus insecticides (B) and pthalimide fungicides (C)A

B C

6.0 Identification of test methodology for evaluating the ED activity of pesticides

FishThe non-acute fish toxicity test protocols that have been internationally standardised by the Organisation for Economic Cooperation and Development (OECD) are available on the OECD website (http://www.oecd.org/document/62/0,3343,en_2649_34377_2348862_1_1_1_1,00.html). These guidance documents on fish testing are the basis of most regulatory work concerning PPPs and fish, and they are listed below:-

Guidance Document 204: Fish, prolonged toxicity test – 14 day studyGuidance Document 210: Fish, early-life stage toxicity testGuidance Document 212: Fish, short-term toxicity test on embryo and sac-fry stages.Guidance Document 215: Fish, juvenile growth test

In particular, the early life stage test (which exposes fish from the stage of fertilised eggs until the larvae are free-feeding) is the basis for most so-called chronic or long-term testing with fish in the European Union and elsewhere. However, none of these tests includes the periods in the fish life cycle of either sexual development or reproduction, so they are clearly unsuitable for assessing the hazards of EDCs.

In recognition of this, the OECD has spent more than 10 years developing standard fish-based tests which are known to be sensitive to some EDCs. The first of these to be ready as an advanced draft is the Test Guideline on

SID 5 (Rev. 3/06) of 30

0.0

20.0

40.0

60.0

80.0

100.0

120.0

0.1 1 10 100 1000 10000

mg/kg

% s

tudi

es hormones

morph

repro

0.0

20.0

40.0

60.0

80.0

100.0

120.0

0.1 1 10 100 1000 10000

hormones

morph

repro

0.0

20.0

40.0

60.0

80.0

100.0

120.0

0.1 1 10 100 1000 10000

mg/kg

% stud

ies hormones

morph

repro

http://www.oecd.org/document/62/0,3343,en_2649_34377_2348862_1_1_1_1,00.html

the Fish Screening Assay for Endocrine Active Substances, which is currently available for comment on the OECD website. This is a 3 week assay with adult fish designed to detect EDCs, primarily through alterations in vitellogenin titres, but not to provide data that can be used directly in risk assessment. It has received extensive validation with all three standard OECD species (fathead minnow, medaka and zebrafish), and is known to be sensitive to strong and weak oestrogens, strong and weak androgens, and aromatase inhibitors. It is not very sensitive to anti-androgens, and its sensitivity to anti-oestrogens and thyroid-active compounds is unknown. As its name implies, this test can be used to screen suspected EDCs, but if a compound tests positive, it would then need more advanced testing to provide data on impacts likely to damage populations (e.g. reduced growth or damaged reproductive success) which could be used in a risk assessment programme. The 3 week screening assay has just been published by OECD (OECD Technical Guideline TG 230, 2009), and another fish-based screening assay that includes measurements of reproductive success including fecundity has also been published (OECD TG 229, 2009).

Two advanced fish tests with sensitivity to EDCs that could be used in risk assessment programmes are both partial life cycle (PLC) tests, and have begun validation through OECD. The first of these is the Fish Sexual Development Test (FSDT) which exposes fish from the egg stage to the point of sexual differentiation (Holbech et al., 2006; OECD, 2008b). In essence, it is an extended early life stage test, and the primary endpoint of value for risk assessment is biased sex ratio. Early validation efforts (mainly in European laboratories) suggest that it may be as sensitive as the FFLC test to some EDCs, but clearly not to those whose primary MOA operates during reproduction. It is likely to be several years before this test has been validated to the point where a guidance document can be published.

The other PLC test which is therefore being validated by OECD is the Fish Short Term Reproduction Test (FSTRT) (Ankley et al., 2001; USEPA, 2007). This starts with reproductively active F0 adults and continues to the stage of F1 larvae or juveniles. It can include several endpoints, but one of the most useful for risk assessment is fecundity. This endpoint is highly variable between individuals, thus necessitating the use of large numbers of replicates in order to achieve satisfactory discriminatory power. Like the FSDT, the FSTRT has already received some validation (mainly in United States laboratories) which indicates that it is sensitive to certain EDCs, but not always as sensitive as FFLC tests (Crane and Matthiessen, 2007). Furthermore, the ratio between FSTRT and FFLC effect levels for EDCs seems to be very variable, thus potentially precluding the use of a fixed and realistic assessment factor to extrapolate from a PLC test result to a likely full life cycle NOEC (Crane and Matthiessen, 2007). As with the FSDT, it is likely that an FSTRT test guideline will not be published by OECD for several years.

OECD is also considering the validation of a fish full life cycle (FFLC) test (OECD, 2008a) with a view to eventual drafting of an international guideline. This type of test usually exposes fish from the fertilised egg stage of the F0 generation to the stage of free-feeding larvae or juveniles of the F1 generation, although it can alternatively start with adults. It is tacitly assumed that such a test will represent all possible modes of action of EDCs (and other compounds), although it is conceivable that strongly bioaccumulative materials may be more potent in 2-generation or multi-generation fish tests due to the maternal transfer of residues via the eggs (no data are yet available on this point). It also seems possible that the very sensitive phenomenon of damaged pheromonal priming of milt release in salmonids may not be given full expression in the FFLC (at least not with the current OECD standard fish species – fathead minnow Pimephales promelas; Japanese medaka Oryzias latipes; zebrafish Danio rerio). However, OECD has not yet decided to begin validation of the FFLC, so it is likely to be at least a decade before this expensive and difficult test has been properly standardised. The United States Environmental Protection Agency (USEPA) has published a protocol for an FFLC test with fathead minnows (Benoit, 1981), and another for the saltwater sheepshead minnow (Hansen et al., 1978), but although these have not been properly validated to modern standards (either with EDCs or other chemicals), the former has been used intermittently in the EU and elsewhere in support of some PPP registrations. The Japanese Ministry of the Environment has also published a life cycle test guideline for medaka (Anon., 2002), again without full validation.

In summary, it is likely to be 10 years before a reasonably comprehensive suite of EDC-sensitive fish toxicity tests has been internationally validated and standardised. However, putting this practical difficulty to one side, it is nevertheless possible to suggest some ideas for how such a suite of tests could be deployed for the hazard and risk assessment of PPPs. The simplest (but highly expensive and ethically undesirable) approach would be to require all PPPs with suspected endocrine action to be subjected to an FFLC test – considered by many to be the ‘gold standard’ of fish toxicity testing. While bearing in mind the possibility that even FFLC tests (as currently envisaged) may not be of maximum sensitivity to some EDCs, this would probably detect most endocrine-active PPPs and provide sound data for use in risk assessment. However, it would clearly be a huge drain on resources, and its routine use for all suspected EDCs would easily outstrip the ability of currently qualified contract testing laboratories to keep up with demand.

A more pragmatic (though more complex) approach would be to design some type of tiered testing strategy. This could take many forms, but the following draft scheme would be one way of tackling the problem (once the full suite of tests has been standardised):-

SID 5 (Rev. 3/06) of 30

Step 1Initially, it will be important to make full use of any information that already exists on the substance in question. This might include effects observed in field trials or during normal use (in the case of PPPs already on the market), predictions based on structure-activity relationships (e.g. a new conazole would be flagged as a likely aromatase inhibitor), data from various in vitro tests of endocrine activity, and ‘read-across’ from in vivo mammalian tests such as the Hershberger assay and the rodent 90-day and two-generation tests. Mammalian toxicity data are likely to be available from an early stage in the evaluation of a new PPP, and given the similarity of all vertebrate endocrine systems, such information can prove to be invaluable for predicting effects in other vertebrate classes including fish. If none of these pieces of information indicates likely endocrine activity at environmentally-relevant doses, it might be concluded that no further testing for this mode of action is required. However, if there are any remaining doubts, one would proceed to Step 2.Step 2If Step 1 suggests that a PPP might have oestrogenic, androgenic or aromatase-inhibiting activity, a next logical move would be to subject the substance to a 21 day fish screening assay to confirm or refute this activity in vivo. If anti-androgenic activity is suspected, the 21 day screen would not be appropriate, but a suitable alternative would be to run a variant of the screen using 3-spined stickleback Gasterosteus aculeatus to investigate spiggin levels in sexually-mature males or androgen-induced females (Katsiadaki et al., 2006; Allen et al., 2008). If anti-oestrogenic or thyroid activity is suspected, the value of the fish screening assay is as yet unknown, and there would currently be no alternative to proceeding immediately to Step 3. If a candidate PPP tests negative in the 21 day fish screen (or its stickleback variant) at environmentally-relevant concentrations, no further testing would be indicated. On the other hand, a positive result would raise the need for definitive in vivo data (Step 3 et seq.) that could be used in risk assessment.Step 3We do not yet know enough about the relationship between effect levels of EDCs in PLC fish tests and effect levels in FFLC tests, but information from an unpublished review (Peter Matthiessen, pers. comm. 2009) suggests that FSDTs may be of similar sensitivity to FFLC tests for many EDCs. The best approach in Step 3 may therefore be to run an FSDT. A positive result at environmentally-relevant concentrations might be considered sufficient information with which to proceed to risk assessment, although some regulators might then wish to obtain the added information available from an FFLC test (Step 5). On the other hand, a negative FSDT could not necessarily be interpreted as an ‘all clear’ (particularly if there has been a positive result in an in vivo screening test, or if there is a possibility that reproduction might be the most sensitive stage in the life cycle). In that event, it would probably be desirable to proceed to Step 4.Step 4This would involve running an FSTRT. If neither the FSTRT not the FSDT give positive responses at environmentally-relevant concentrations, then proceeding to risk assessment without further testing would probably be justifiable. If, on the other hand, the FSTRT is positive, most regulators would then wish to proceed to Step 5 in order to explore more fully the longer-term implications of interference with either sexual development or reproductive success (or both).Step 5This final step involves running a full FFLC test. It is conceivable that some regulators might require such a test as a direct consequence of a positive screening test, but this would probably be very wasteful of resources. If both PLC tests are negative, the current indications are that an FFLC test will probably only give a positive result for strongly bioaccumulative substances (Peter Matthiessen, pers. comm.., 2009), which are unlikely to receive approval as PPPs for that reason alone. Another factor to be considered is the complexity of FFLC tests which unfortunately sometimes requires the performance of a repeat. The need for repeats can be minimised by knowing the likely most sensitive endpoint in advance, information which is not available from a screening assay, but which can be obtained from the PLC tests.

As indicated above, Steps 1-5 are merely one potential way in which to proceed with the aquatic hazard assessment of suspected endocrine disrupting PPPs. Neither this proposed strategy, nor any other, can be considered infallible. This is especially the case because it seems probable that not all EDCs will provoke a response in the fish tests as currently envisaged. The magnitude of this risk of false negatives is unknown at present, but probably justifies some further research before a scheme of this type is implemented. It would be particularly useful to know, inter alia, whether the FSTRT and FFLC test can be modified to make them maximally sensitive to interference with the pheromonal priming of male maturation, a mode of action observed in salmonids which makes several PPPs extremely potent.

Further work is also needed on the best way to assess the environmental risks posed by endocrine-active PPPs. Although the fundamentals of risk assessment apply to EDCs just as much as to other chemicals, there are a number of special features which need to be taken into account. These have been discussed in the context of the European Union’s Registration, Evaluation and Authorisation of Chemicals regulations (REACH) (Matthiessen and Johnson, 2007). Here is not the place to go into this in detail, but relevant issues addressed by Matthiessen and Johnson include mixture effects, non-linear responses and low dose effects, delayed effects, and extrapolation between species. Several of these issues apply to all chemicals, not just EDCs, but the recognition of endocrine disruption as a significant MOA has brought them to the fore.

SID 5 (Rev. 3/06) of 30

Birds and mammals For birds and mammals the risk assessment for endocrine disruptors was reviewed in the EFSA 2008 Scientific Opinion of the Panel on Plant protection products and their Residues (PPR) on the Science behind the Guidance Document on Risk Assessment for birds and mammals. In summary the following steps were proposed:Step 1Study the information available from tests performed on other taxa (fish, amphibians, mammals and birds) for the substance under assessment. Information from structurally related substances may also be considered. If the data give rise to concerns of potential endocrine-mediated effects of the substance, then mammalian screening tests should be assessed to clarify the mechanism of action, and/or the potential of the test substance to cause endocrine-mediated effect in birds/mammals (in vivo). With regard to mammals, and in contrast to birds, a number of in vitro and in vivo screening tests for assessing endocrine-disrupting properties have become available in recent years and are in various stages of (pre-)validation Step 2Study the information available from mammalian screening studies to clarify any potential of the substance to influence known endocrine mechanisms. In case (in vitro) screening studies in mammals show that the substance has an effect on a known endocrine mechanism, further assessment is needed to allow for the evaluation or generation of data relevant to risk assessment. The mammalian multi-generation study, performed for pesticide risk assessment, covers the entire reproductive cycle and therefore is able to provide information on overallproductivity at the population level. In addition to mammalian screens, fish and amphibian screens exist that can address the question of the likelihood of a material to be an endocrine disruptor, as well as its probable mode of action. This information should also be taken into account for the assessment as further weight of evidence. In cases where screens are ‘positive’, or where no screens are available but concerns for potential endocrine-mediated effects remain, step 3 should be taken.Step 3Assess the standard (multi-generation) mammalian reproductive study or any available relevant mammalian in vivo study for potential endocrine-mediated effects on reproduction. Derive an endpoint value for these effects to be used in risk assessment for wild mammals. The absence of endocrine-mediated effects in mammalian in vivo studies is not sufficient to conclude a risk assessment on birds. It is not possible to use the endpoints from a mammalian risk assessment in an avian assessment. Such endpoints can only be used as a source of information.Step 4Assess all information available from the standard one-generation avian reproduction study or a specific modified one-generation study modified to include endocrine endpoints. The information provided may help in determining an appropriate strategy for further testing but will, in general, not provide conclusive information on endocrine-mediated effects. Currently, no internationally accepted testing methodology is available, that can be used to adequately assess the impact of endocrine meditated effects of a substance on the reproductive risk assessment in birds.Step 5Assess any specific two-generation or sensitive life stage study in birds for endocrine-mediated endpoints. When assessing/selecting the appropriate test design and the appropriate endpoints, it is essential to evaluate all the available information on avian and/or other species. If available information allows, the likely mode of action and the part of the avian life-cycle likely to be the most sensitive (with associated behaviours) should be identified. Subsequently, an appropriate test design should be selected. There is no single test design that should automatically be followed. In addition, only those techniques should be applied that have been developedsufficiently to assess the various endpoints.

Mammalian test methodsUnder 91/414EEC reproductive and developmental effects studies are required and specifically identify multi-generation and developmental studies as required for authorisation. It is likely therefore that effects on mammalian reproduction and growth will be identified during the authorisation process. If screening assays are required to address specific effects the proposed tiered testing strategy under development by the US EPA provides an appropriate approach (see EPA EDSTAC below)The non-acute mammalian toxicity test protocols that have been internationally standardised by the Organisation for Economic Cooperation and Development (OECD) are also available on the OECD website. These guidance documents and guidelines on mammalian toxicity testing are the basis of most regulatory work concerning PPPs and mammals, and they are listed below:Test No. 407: Repeated Dose 28-day Oral Toxicity Study in Rodents Test No. 408: Repeated Dose 90-Day Oral Toxicity Study in Rodents Test No. 409: Repeated Dose 90-Day Oral Toxicity Study in Non-Rodents Test No. 414: Prenatal Development Toxicity Study Test No. 415: One-Generation Reproduction Toxicity Study Test No. 416: Two-Generation Reproduction Toxicity Test No. 421: Reproduction/Developmental Toxicity Screening Test Test No. 422: Combined Repeated Dose Toxicity Study with the Reproduction/Developmental Toxicity Screening

SID 5 (Rev. 3/06) of 30

http://puck.sourceoecd.org/vl=4805188/cl=14/nw=1/rpsv/ij/oecdjournals/1607310x/v1n4/s23/p1








Test Test No. 440: Uterotrophic Bioassay in Rodents: A short-term screening test for oestrogenic properties

Within these tests obviously the 2-generation test includes the both sexual development and reproduction phases of the rat life cycle and therefore is the most relevant to the identification and evaluation of the impact of endocrine disrupters. Recently a call for data (deadline February 2009) for a review of the need for the 2-generation test has been published by the OECD (led by the US EPA) with a current proposal to use the extended one generation reproductive toxicity study, currently under discussion, and any other data available to identify cases where a 2-generation study should be performed. In addition the screening assays, such as the uterotrophic bioassay in rodents, can provide additional information during the screening process if there are concerns based on effects of chemicals with similar structure. A number of other test methods are currently under development and validation under the auspices of the OECD:The validation of the Test Guideline 407 Repeat Dose 28-day Oral Toxicity Study has been updated to possibly detect endocrine disruption effects. Validation studies of the updated Test Guideline have been performed under the OECD Test Guidelines Programme. The validation is now being reviewed by an international panel of reviewers. The Hershberger bioassay is intended to detect androgen agonists, antagonists and 5α-reductase inhibitors. Validation studies of this test have been performed under the OECD Test Guidelines Programme. The validation of the Hershberger assay is now being reviewed by an international panel of reviewers.The Stably Transfected Transcriptional Activation Assay is intended to detect estrogenic activity. Validation studies of this test have been performed by Japan. The validation is now reviewed by an international panel of reviewers. The OECD have also published a number of review documents on their website:No 97: Detailed Review Paper on the Use of Metabolising Systems for In Vitro Testing of Endocrine DisruptorsNo. 57: Detailed Review Paper on Thyroid Hormone Disruption AssaysNo. 21: Detailed Review Paper: Appraisal of Test Methods for Sex Hormone Disrupting Chemicals

Avian test methodsEFSA recognises the distinction between birds and mammals in the recent Bird and Mammal guidance document opinion “For mammals, screening and testing methodology is available, that allows for adequate assessment of endocrine mediated effects on reproduction. This is much less the case for birds. Due to differences in mechanisms for sex differentiation, absence of endocrine mediated effects in in vivo mammalian studies cannot be considered sufficient to negate concerns for potential endocrine effects in birds when (in vitro-) screening tests have demonstrated for a substance to have a potential to influence endocrine processes. Therefore, at present a fully conclusive assessment for birds is not feasible.” The proposed steps in the risk assessment are outlined above and this section discussion the available tests in more detail

The EPA EDSTAC (Endocrine Disruptor Screening and Testing Advisory Committee) proposed that a tier 1 screening battery of in vitro and in vivo assays aimed at detecting effects based on estrogen, androgen and thyroid hormones would be sufficiently comprehensive to detect any activity likely to also occur in birds. The list of tier 1 assays currently under pre-validation or validation are: Estrogen receptor binding or transcriptional activation Androgen receptor binding or transcriptional activation In vitro steroidogenesis assay Uterotrophic Hershberger Pubertal female with thyroid Frog metamorphosis assay for thyroid Fish screening assay There is currently no equivalent for birds of the uterotrophic and Hershberger screening bioassays in mammals. However, Millam et al (2002) have reported an avian bioassay for estrogens based on the growth response of the zebra finch chick oviduct based on increased oviduct weight and at the highest doses of estradiol benzoate changes in the differentiation of the oviduct. Such a screening bioassay may offer some advantages for birds over large scale in vivo studies.The non-acute avian toxicity test protocols that have been internationally standardised by the Organisation for Economic Cooperation and Development (OECD) are also available on the OECD website. These guidance documents and guidelines on avian toxicity testing are the basis of most regulatory work concerning PPPs and they are listed below:Test No. 205: Avian Dietary Toxicity Test Test No. 206: Avian Reproduction Test However, the endpoints currently included are unlikely to reliably identify endocrine related effects. Therefore a long-standing working group has been developing a 2-generation study in the Japanese quail with the aims of addressing endocrine disruptor effects in birds and part of this has included an OECD document:No. 74: Detailed Review Paper for Avian Two-generation Toxicity Testing

SID 5 (Rev. 3/06) of 30

http://appli1.oecd.org/olis/2007doc.nsf/linkto/env-jm-mono(2007)21



http://appli1.oecd.org/olis/2002doc.nsf/linkto/env-jm-mono(2002)8

http://www.oecd.org/dataoecd/49/35/37235405.pdf

http://www.olis.oecd.org/olis/2008doc.nsf/linkto/env-jm-mono(2008)24



At tier 2 the recommended assay for birds is the 2-generation test. This recommendation is based on the recognition of four life stages of birds where endocrine related processes are most vulnerable: in ovo, development of offspring, sexual maturation and reproduction. Thus the current one-generation reproduction test is not sufficient as it does not assess the ability of the F1 generation to reproduce. The proposed 2-generation study in quail is in the pre-validation phase (Touart 2004) and it is likely to be many years before it is routinely used. During the development and validation of the test there are a number of issues that have to be addressed, such as the modes of action for which the test design is able to detect effects and, related to this, the most suitable endpoints. Current outstanding modes of action include the effects of anti-estrogens, anti-androgens, thyroid hormone agonists and the links of effects to other endpoints such as hormone levels.

The clear omission in all approaches to date is the effects of endocrine disruptors on altricial species. All OECD test methods are based on precocial species (Japanese quail, northern bobwhite quail, mallard) and therefore studies do not take into account parental care, e.g. nest building and caring for young, which are important in a wide range of species. Thus the species tested is a particular issue where endocrine mediated effects occur in endpoints not measured in current precocial species (eg nest building); of course the current precocial species testing approach may also miss other non-endocrine pesticide effects. Hoogenstein et al (2005) investigated the effects of a known endocrine disruptor (Arochlor 1248) on the reproductive behaviour of the zebra finch (an altricial species). They demonstrated that exposure resulted in more clutches laid and nests constructed per pair and incubation time increased but parameters normally included in reproduction tests were not significantly affected. Their conclusions were that evaluation of parameters related to parental care is more important for endocrine disruption assessment in altricial species than simple reproductive success parameters evaluated in routine reproductive studies. Rochester et al (2008) used a similar zebra finch based assay to demonstrate masculinsed nest building behaviour in females such as changes in nest structure, nest building speed. Parental investment in raising young by barn swallows has been shown by Jenni-Eiermann et al (2008) to be related to plasma corticosterone levels in parents with a negative correlation between levels and body mass of nestlings. Schew et al (1996) also demonstrated significant differences between altricial and precocial species in the development of the thyroid with altricial species increasing plasma thyroid hormone levels post hatching whereas precocial species thyroid development occurs during the embryonic and peri-hatch periods. This suggests that precocial species are likely primarily affected by exposure in ovo whereas altricial species continue to be vulnerable post-hatching.

The absence of assays to detect effects on the complete life cycle of both precocial and altrical species is a key issue for the detection of endocrine effects. The major differences in the care and development of young requires that effects on both reproductive success and parenting are required in the F0 and F1 generations. Only reproductive success is currently addressed in the proposed 2-generation OECD test design. It is therefore recommended that a study design capable of identifying endocrine-related effects in altricial species is developed so as to address this major gap in the ability to detect effects of pesticides.. 7.0 Conclusions

1. The published data on the endocrine disrupting effects of PPPs in fish- and avian-based test systems are relatively sparse. There are more data for mammals but still these do not cover all classes of pesticides. There are few reports demonstrating the absence of effects of pesticides since these data are less likely to be published. These limitations should be borne in mind when considering any response to this report.

2. Potential or actual endocrine disrupting properties appear to be fairly widely distributed across different types of PPP for fish, birds and mammals. Endocrine activity is found in some insecticides, fungicides and herbicides, and although it can be clearly linked in some cases to structure (e.g. aromatase inhibition in conazole fungicides), in other cases the structural correlates of endocrine activity (e.g. interference with the cortisol response to stress; damage to the olfactory response to reproductive prostaglandins) are hard to discern at present.

3. Mechanisms of endocrine disruption by PPPs are as varied as in the wider universe of chemicals. Cases of in vitro or in vivo effects described in this report include altered steroidogenesis (e.g. via aromatase inhibition or induction), estrogen agonism or antagonism, androgen antagonism, progestin antagonism, damaged cortisol response to stress, damaged olfactory and gonadal response to reproductive pheromones, displacement of hormones from sex steroid binding protein, disruption of sensitivity to peptide hormone-like growth factors, and interference with several thyroid mechanisms.

4. In many cases, the causes of endocrine disruption in vivo are not completely understood. This is partly because several endocrine-active PPPs appear to have multiple MOAs, thus confounding attempts at interpretation. Furthermore, some endocrine mechanisms have not yet been replicated in vitro, making diagnoses of effects in vivo difficult.

5. Some effects of PPPs on endocrine systems in vivo occur at environmentally-relevant concentrations, and are clearly adverse in their own right (i.e. they have a direct bearing on apical processes such as growth and reproduction). Many others either occur at unrealistically high concentrations, or are more accurately described as biomarkers which may or may not have implications for higher order apical damage. Some biomarker responses have very important consequences for body function, but correlations with higher order

SID 5 (Rev. 3/06) of 30

damage have only been made in a few cases (e.g. in fish vitellogenin induction resulting from exposure to oestrogen mimics is closely correlated with adverse effects on reproduction).

Due to the sparse nature of the available data, it is difficult to make comparisons between endocrine effect levels and concentrations causing apical changes. However for fish, out of 29 PPPs for which comparisons have been made in this report, the most sensitive endocrine-related endpoints were more sensitive than the most sensitive apical ones in 35% of cases. This is probably due in part to the fact that almost all apical fish toxicity tests in common use are insensitive to endocrine MOAs. In mammals 10% of the PPPs studied, the most sensitive endocrine-related effects occurred at lower concentrations than the most sensitive apical effects. This difference from fish is probably because the 2-generation study is in widespread use in mammalian toxicity testing. In birds 40% of the PPPs studied, the most sensitive endocrine-related effects occurred at lower concentrations than the most sensitive apical effects. This is probably because the 1-generation study in widespread use in avian toxicity testing is insensitive to many endocrine effects.

1. This has obvious implications for the ways in which PPPs (and other chemicals) are tested in the future: 2. A suite of apical fish toxicity tests with proven or expected sensitivity to EDCs is being developed by

OECD, but is not expected to be fully validated for another 10 years. This includes a short-term screening test for endocrine activity (which should be finalised in 2009/10), two partial life cycle tests currently in validation, and a full life cycle test which has not yet begun validation. As it stands, this suite is probably sensitive to most, but not all, modes of endocrine action, and it will almost certainly require some further development.

3. One possible strategy for deploying this suite of improved fish toxicity tests in a hazard and risk assessment programme for PPPs involves a tiered approach that attempts to minimise the resources required to make a robust risk assessment. However, it has to be recognised that in many cases of PPPs with suspected endocrine activity, the only reasonably definitive method to define their apical effects may be the fish full life cycle (FFLC) test. This has both financial and ethical implications due to the complexity of the test and the numbers of animals which it requires.

4. For birds and mammals the proposed EFSA tiered risk assessment approach for endocrine effects has been outlined in the recently published opinion on the science behind the bird and mammal guidance document (EFSA 2008)

For mammals the current 2-generation study appears to present a robust method of detecting endocrine related effects but a move to reduce the number of such studies undertaken by relying on additional endpoints in a one generation study is underway and will require the extensive validation proposed.

For birds the absence of assays to detect effects on the complete life cycle of both precocial and altrical species is a key issue for the detection of endocrine effects. The major differences in the care and development of young requires that effects on both reproductive success and parenting are required in the F0 and F1 generations. Only reproductive success is currently addressed in the proposed 2-generation OECD test design. It is therefore recommended that a study design capable of identifying endocrine-related effects in altricial species is developed so as to address this major gap in the ability to detect effects of pesticides.

It is probably true that most of the PPPs with endocrine activity in fish and birds are also active in mammals. However, the absence of an effect in mammals and fish cannot be used as evidence of lack of an effect in birds. Effects may be observed in birds but not in mammals due to the differences in default sex during development. Even where effects occur across phyla the important point is that the ways in which this activity are expressed, and the concentrations at which it occurs, are often very different. For that reason, we do need to use fish and birds for this type of testing.

Finally, the way in which risk assessments should be carried out for PPPs which disrupt the endocrine systems has not yet been fully considered. Some modifications to standard risk assessment procedures will probably be needed to recognise the unique properties of EDCs.

References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.

SID 5 (Rev. 3/06) of 30

Allen, Y.T., Katsiadaki, I., Pottinger, T.G., Jolly, C., Matthiessen, P., Mayer, I., Smith, A., Scott, A.P., Eccles, P., Sanders, M.B., Pulman, K.G.T. and Feist, S. (2008) Intercalibration exercise using a stickleback endocrine disrupter screening assay. Environ. Toxicol. Chem. 27, 404-412.Ankley, G.T., Jensen, K.M., Kahl, M.D., Korte, J.J. and Makynen, E.A. (2001). Description and evaluation of a short-term reproduction test with the fathead minnow (Pimephales promelas). Environ. Toxicol. Chem. 20, 1276-1290.Anon. (2002). The medaka (Oryzias latipes) full life-cycle test guideline. Ministry of the Environment, Japan. 14 pp.Balthazart J (1990) Comparative Physiology: Hormones, Brain and Behaviour in Vertebrates, Pt.2 - Behavioural Activation in Males and Females, Social Interactions and Reproductive Endocrinology v. 9 S Karger AGCrane, M. and Matthiessen, P. (2007). An evaluation of the reproducibility and power of various endpoints in partial and full fish life cycle tests. Unpublished paper submitted to the 6th meeting of the OECD Validation Management Group for Ecotoxicity Tests, Paris, 16-17 Jan. 2008. 90 pp.Benoit, D.A. (1981). User’s Guide for Conducting Life-Cycle Chronic Toxicity Testing With Fathead Minnows (Pimephales promelas). United States Environmental Protection Agency, USEPA 600/8-81-011, NTIS / PB82-238395.Dawson A (2000) Mechanisms of endocrine disruption with particular reference to occurrence in avian wildlife: A review. Ecotoxicology 9 59-69EFSA (2008) Scientific Opinion of the Panel on Plant protection products and their Residues (PPR) on the Science behind the Guidance Document on Risk Assessment for birds and mammals The EFSA Journal (2008) 734: 1-181Folmar, L.C., Gardner, G.R., Schreibman, M.P., Magliulo-Cepriano, L., Mills, L.J., Zaroogian, G., Gutjahr-Gobell, R., Haebler, R., Horowitz, D.B., and Denslow, N.D. (2001). Vitellogenin-induced pathology in male summer flounder (Paralichthys dentatus). Aquat Toxicol. 51, 431–441.Groshart, C. and Okkerman, P.C. (2000). Towards the establishment of a priority list of substances for further evaluation of their role in endocrine disruption. Report by BKH Consulting Engineers to the European Commission, Brussels, M 0355008/1786Q/10/11/00, 29 pp. + annexes.Hansen, D.J., Parrish, P.R., Schimmel, S.C. and Goodman, L.R. (1978). Sheepshead Minnows (Cyprinodon variegatus). Bioassay Procedures for the Ocean Disposal Permit Program. United States Environmental Protection Agency, Washington D.C., EPA-600/9:78-010.Hiller-Sturmhöfel, S., and Bartke, A. (1998) The Endocrine system: An overview Alcohol Health & Research World. 22, No. 3 153-164Holbech, H., Kinnberg, K., Petersen, G.I., Jackson, P., Hylland, K., Norrgren, L. and Bjerregaard, P. (2006). Detection of endocrine disrupters: evaluation of a fish sexual development test (FSDT). Comparative Biochemistry and Physiology 144C, 57-66.Hoogestein, A. L, deVoogd, T. J., Quinby, F. W, De Caprio, T., and Kollias, G. V.(2005) Reproductive impairment in zebra finches (Taeniopygia guttata). Environmental Toxicology and Chemistry 24, 219-223Jenni-Eiermann, S., Glaus, E. Gruebler M., Schwable, H., and Jenni, L. (2009) Glucocorticoid response to food availability in breeding barn swallows (Hirundo rustica). General and Comparative Endocrinology 155(3), 558-565. Jobling, S., Coey, S., Whitmore, J. Kime, D.E., van Look, KL, McAllister BG, Beresford, N., Henshaw, AC, Brighty, G., Tyler, C.R. and Sumpter, J.P. (2002). Wild roach (Rutilus rutilus: Cyprinidae) living in effluent contaminated rivers have reduced fertility. Biol. Reprod. 67, 515-524Katsiadaki, I., Morris, SW., Squires, C., Hurst, M.R., James, J.D. and Scott, A.P. (2006). Use of the three-spined stickleback (Gasterosteus aculeatus) as a sensitive in vivo test for detection of environmental antiandrogens. Environ. Health Perspect. 114 (suppl 1), 115-121.Kidd, K.A., Blanchfield, P.J., Mills, K.H., Palace, V.P., Evans, R.E., Lazorchak, J.M. and Flick, R.W. (2007). Collapse of a fish population after exposure to a synthetic estrogen. Proc. Nat. Acad. Sci. 104, 8897-8901.Matthiessen, P. (2003). An historical perspective on endocrine disruption in wildlife. Pure and Applied Chemistry 75, 2197-2206.Matthiessen, P. (2006a). Estrogenic contamination of surface waters and its effects on fish in the United Kingdom. In: Estrogens and xenoestrogens in the aquatic environment: an integrated approach for field monitoring and effect assessment (eds. Vethaak, D., Schrap, M. and de Voogt, P.), Society of Environmental Toxicology and Chemistry, Pensacola, pp. 329-357. ISBN 1-880611-85-6. Matthiessen, P. and Johnson, I. (2007). Implications of research on endocrine disruption for the environmental risk assessment, regulation and monitoring of chemicals in the European Union. Environmental Pollution 146, 9-18.Miller, D.H., Jensen, K.M., Villeneuve, D.L., Kahl, M.D., Makynen, E.A., Durhan, E.J. et al. (2007). Linkage of biochemical responses to population-level effects: a case study with vitellogenin in the fathead minnow (Pimephales promelas). Environ Toxicol Chem 26, 521-527OECD (2008a). Detailed Review Paper on Fish Life-Cycle Tests, OECD Series on Testing and Assessment No. 97, Organisation for Economic Cooperation and Development, Paris, 146 pp.OECD (2008b). Draft report of Phase 1 of the validation of the fish sexual development test. Unpublished report prepared for the 6th meeting of the OECD Validation Management Group for Ecotoxicity Testing, Organisation for Economic Cooperation and Development, Paris, 62 pp.

SID 5 (Rev. 3/06) of 30

http://www.amazon.co.uk/Comparative-Physiology-Interactions-Reproductive-Endocrinology/dp/3805551851/ref=sr_1_40?ie=UTF8&s=books&qid=1238489392&sr=8-40

http://www.amazon.co.uk/Comparative-Physiology-Interactions-Reproductive-Endocrinology/dp/3805551851/ref=sr_1_40?ie=UTF8&s=books&qid=1238489392&sr=8-40

Ottinger M.A. ,Wu, J.M., Hazelton, J.L., Abdelnabi, M.A., Thompson, N., Quinn, M.L., donoghue, D., Schenk, F., Ruscio, M., Beavers, J., Jaber, M. (2005) Assessing the consequences of the pesticide methoxychlor: neuroendocrine and behavioural measures as indicators of biological impact of an estrogenic environmental chemical Brain Research Bulletin, 65, 199-209Ottinger, M.A., Quinn, M.J., Lavoie, E., Abdelnabi, M.A., Thompson, N., Hazelton, J.L., Wu, J.M., Beavers, J., Jaber, M. (2005) Consequences of endocrine disrupting chemicals on reproductive endocrine function in birds: Establishing reliable endpoints of exposure Domestic Animal Endocrinology 29, 411-419Petersen, G., Rasmussen, D. and Gustavson, K. (2007). Study on enhancing the endocrine disruptor priority list with a focus on low production volume chemicals. Report by DHI Water & EnvironmentRochester, J. R., Heiblum, R., Rozenbloim, I., and Millam, J. R. (2008) Post-hatch oral estrogen exposure reduces oviduct and egg mass and alters nest building behaviour in adult zebra finches (Taeniopygia guttata). Physiology and Behaviour 95(3), 370-380. 2Schew, W. A., Mcnabb, F. M. A., and Scanes, C. G. (1996) Comparison of the ontogenesis of thyroid hormones, growth hormone and insulin like growth factor-I in ad libitum and food restricted (altriacial) European starlings and (precocial) Japanese quail. General and Comparative Endocrinology 101(3), 304-316. Thomas, K.V., Hurst, M., Matthiessen, P., Sheahan, D. and Williams, R.J. (2001). Toxicity characterisation of organic contaminants in stormwaters from an agricultural headwater stream in south east England. Water Research 35, 2411-2416.Thorpe, K.L., Benstead, R., Hutchinson, T.H. and Tyler, C.R. (2007). Association between altered vitellogenin concentrations and adverse health effects in fathead minnows (Pimephales promelas). Aquat. Toxicol. 85, 176-183.Touart, L. W. (2004) Factors considered in using birds for evaluating endocrine-disrupting chemicals. ILAR Journal 45(4), 462-468. USEPA (2007). Validation of the fish short-term reproduction assay: integrated summary report. Unpublished report to OECD, United States Environmental Protection Agency, Washington DC, 102 pp + appendices. to the European Commission DG Environment, Brussels, ENV.D.4/ETU/2005/0028r, 249 pp.

SID 5 (Rev. 3/06) of 30

General enquiries on this form should be made to: -...

Documents

Transcript of General enquiries on this form should be made to: -...