Towards a science-based testing strategy to identify maternal thyroid hormone imbalance and neurodevelopmental effects in the progeny – Part IV: the ECETOC and CLE Proposal for a Thyroid Function-Related Neurodevelopmental Toxicity Testing and Assessment Scheme (Thyroid-NDT-TAS)

Abstract

Following the European Commission Endocrine Disruptor Criteria, substances shall be considered as having endocrine disrupting properties if they (a) elicit adverse effects, (b) have endocrine activity, and (c) the two are linked by an endocrine mode-of-action (MoA) unless the MoA is not relevant for humans. A comprehensive, structured approach to assess whether substances meet the Endocrine Disruptor Criteria for the thyroid modality (EDC-T) is currently unavailable. Here, the European Centre for Ecotoxicology and Toxicology of Chemicals Thyroxine Task Force and CropLife Europe propose a Thyroid Function-Related Neurodevelopmental Toxicity Testing and Assessment Scheme (Thyroid-NDT-TAS). In Tier 0, before entering the Thyroid-NDT-TAS, all available in vivo, in vitro and in silico data are submitted to weight-of-evidence (WoE) evaluations to determine whether the substance of interest poses a concern for thyroid disruption. If so, Tier 1 of the Thyroid-NDT-TAS includes an initial MoA and human relevance assessment (structured by the key events of possibly relevant adverse outcome pathways) and the generation of supportive in vitro/in silico data, if relevant. Only if Tier 1 is inconclusive, Tier 2 involves higher-tier testing to generate further thyroid- and/or neurodevelopment-related data. Tier 3 includes the final MoA and human relevance assessment and an overarching WoE evaluation to draw a conclusion on whether, or not, the substance meets the EDC-T. The Thyroid-NDT-TAS is based on the state-of-the-science, and it has been developed to minimise animal testing. To make human safety assessments more accurate, it is recommended to apply the Thyroid-NDT-TAS during future regulatory assessments.

Keywords:

• Endocrine disruptor
• thyroid modality
• adverse outcome pathway (AOP)
• mode-of-action (MoA)
• thyroid peroxidase (TPO)
• sodium-iodide symporter (NIS)
• deiodinase (DIO)
• uridine diphosphate glucuronyltransferase (UGT)
• liver enzyme induction
• weight-of-evidence (WoE)
• extended one-generation reproductive toxicity study (EOGRTS; OECD TG 443)
• developmental neurotoxicity study (DNT study; OECD TG 426)
• motor activity
• cognitive function (learning and memory)

1. Introduction

1.1. Background and scope

Thyroid perturbations during pregnancy and lactation can result in a variety of developmental alterations including neurodevelopmental impairment (Zoeller et al. 2007; Gilbert et al. 2012, 2020). Therefore, evaluations of thyroid function during offspring development can be relevant for toxicological assessments. This is also reflected in the European Food Safety Authority (EFSA) and European Chemicals Agency (ECHA) Guidance for the identification of endocrine disruptors in the context of Regulations (EU) No 528/2012 and (EC) No 1107/2009 (EFSA and ECHA 2018) that was developed for substances regulated under the Plant Protection Products and Biocidal Products Regulations (EP and Council 2009, 2012).

Appendix A of the EFSA and ECHA (2018) Endocrine Disruptor Guidance presents Additional considerations for how to assess the potential for thyroid disruption for human health. Specifically, Appendix A describes patterns of thyroid-related effects in experimental animals that are considered to represent a concern for thyroid perturbation-mediated neurodevelopmental impairment in humans so further investigations are required. A general testing scheme is then proposed that is applicable to liver enzyme inducers only, i.e. substances that mediate enhanced thyroid hormone clearance (Curran and DeGroot 1991). The general testing scheme presented in Appendix A includes serum thyroid hormone measurements in the most sensitive populations, comparative in vitro studies of liver enzyme activities induced by the test substance in animal and human test systems, and the exclusion of other possible thyroid-related modes-of-action (MoAs).

However, neither Appendix A nor the further clarifications that have since been published in the EFSA (2020) Technical report on the outcome of the pesticides peer review meeting on general recurring issues in mammalian toxicology provide broader guidance covering different thyroid-related MoAs or indicate how the data should be evaluated in a weight-of-evidence (WoE) approach to reach a conclusion on whether, or not, a substance meets the European Commission (2017, 2018) Endocrine Disruptor Criteria (EDC). Appendix A recognises that the identification of thyroid-related hazards is currently hampered by a lack of internationally validated test methods. Overall, it is currently unclear how specific thyroid-related MoAs should be identified and how the (non-)human relevance of thyroid effects and/or neurodevelopmental effects observed in rats should be established (see also reviews by Gilbert et al. 2012, 2020; Kortenkamp et al. 2020).

To address these uncertainties, the European Centre for Ecotoxicology and Toxicology of Chemicals (ECETOC) convened the Thyroxine (T4) Task Force (TF). It is the overarching goal of this TF to review the available evidence in order to contribute to the development of a science-based tiered testing strategy to identify (1) if a substance has the ability to elicit thyroid hormone imbalance and potentially also neurodevelopmental effects in the progeny; (2) if effects observed in rodents are not relevant for humans in line with the EDC (i.e. in accordance with the EDC, effects observed in rodents are by default considered relevant for humans, and it is the non-human relevance that would need to be established); and (3) if a threshold for thyroid hormone decrements can be identified below which neurodevelopmental effects are not to be expected. Building upon the evidence collated in the three previous reviews by the ECETOC T4 TF (Sauer et al. 2020; Marty et al. 2021, 2022), the present review fulfils the overarching goal of the TF and proposes a science-based tiered testing strategy. To enhance the range of underlying expertise, the ECETOC T4 TF has collaborated with CropLife Europe (CLE) in developing the ECETOC and CLE Proposal for a Thyroid Function-Related Neurodevelopmental Toxicity Testing and Assessment Scheme (Thyroid-NDT-TAS).

Note, throughout this article, the term neurodevelopmental toxicity (NDT) is used to describe adverse neurodevelopmental effects which are specifically related to thyroid hormone perturbations. This is in contrast to developmental neurotoxicity (DNT), which is used throughout as term to describe all types of substance-mediated neurodevelopmental adverse effects regardless of the underlying cause, i.e. also those that are unrelated to thyroid hormone perturbations.

The Thyroid-NDT-TAS, as it is presented here, only considers NDT as the adverse outcome of thyroid-related MoAs but not, e.g. thyroid carcinoma in rats. The human relevance of these tumours has generally been questioned (Bartsch et al. 2018; Foster et al. 2021). However, the Thyroid-NDT-TAS is not prescriptive. It presents a generic concept for the assessment of thyroid hormone imbalance and possibly resulting adverse effects.

The Thyroid-NDT-TAS is proposed to help overcome the impasse resulting from the fact that the current Appendix A of the EFSA and ECHA (2018) Endocrine Disruptor Guidance does not provide clear guidance on how to establish whether, or not, a substance meets the European Commission (2017, 2018) EDC for the thyroid modality (EDC-T; see Section 1.2 below for details on the EDC). The Thyroid-NDT-TAS is based on the state-of-the-science; it has been developed to make human safety assessments more accurate while at the same time minimising animal testing in line with the 3Rs principle to replace, reduce and refine animal testing (Russell and Burch 1959) that has been implemented in the Plant Protection Products and Biocidal Products Regulations (EP and Council 2009, 2012) as well as in Directive 63/2010/EU on the protection of animals used for scientific purposes (EP and Council 2010).

Further, the Thyroid-NDT-TAS is conceived to comply with current European Union (EU) legislation on the determination of endocrine disrupting properties. This legislation is outlined in Section 1.2 below as the starting point for the rationale underlying the Thyroid-NDT-TAS. Based thereupon, Section 2 describes the elements of the Thyroid-NDT-TAS, and Section 3 draws conclusions on the applicability of the Thyroid-NDT-TAS to determine whether, or not, a substance meets the European Commission (2017, 2018) EDC-T, including the determination whether effects observed in rat studies are not relevant for humans.

1.2. EU legislation of relevance for the Thyroid-NDT-TAS

The Commission Delegated Regulation (EU) 2017/2100 (European Commission 2017) and the Commission Regulation (EU) 2018/605 (European Commission 2018) provide the legal framework setting out scientific criteria for the determination of endocrine disrupting properties in the context of the Biocidal Products Regulation (EP and Council 2012) and the Plant Protection Products Regulation (EP and Council 2009), respectively.

A substance shall be considered as having endocrine disrupting properties if it (a) elicits adverse effects (unless it can be shown that these are not relevant to humans), (b) has an endocrine activity, and (c) the adverse effect is the consequence of the endocrine activity, i.e. the two are linked by an endocrine MoA (Box 1). Further, this link between the endocrine activity and the adverse effect should be biologically plausible (discussed in Section 2.1.1 and Section 2.3.1.2).

Point 1(a) of the EDC (Box 1) clearly states that the observed effects must be adverse, i.e. they must result in an impairment of functional capacity, of the capacity to compensate for additional stressors or of an increase in susceptibility to other influences. This definition for adversity follows the definition for adversity by the World Health Organisation/International Programme on Chemical Safety (WHO/IPCS 2009). By comparison, if a substance that has endocrine activity only causes adaptive, non-adverse effects, it does not fulfil the EDC and shall not be considered as having endocrine disrupting properties. Adaptive, non-adverse effects have been defined as biological effects that do not cause biochemical, behavioural, morphological or physiological changes that affect the general well-being, growth development or life span of an animal (Lewis et al. 2002).

Box 1. Endocrine disruptor criteria (European Commission 2017, 2018).

For active substances in biocidal products, Point 1 of Section A (Endocrine-disrupting properties with respect to humans) of the Annex to European Commission (2017) states:

1. A substance shall be considered as having endocrine-disrupting properties that may cause adverse effect in humans if, based on points (a) to (d) of point (2)^[#], it is a substance that meets all of the following criteria, unless there is evidence demonstrating that the adverse effects identified are not relevant to humans:

   1. it shows an adverse effect in an intact organism or its progeny, which is a change in the morphology, physiology, growth, development, reproduction or life span of an organism, system or (sub)population that results in an impairment of functional capacity, an impairment of the capacity to compensate for additional stress or an increase in susceptibility to other influences;

   2. it has an endocrine MoA, i.e. it alters the function(s) of the endocrine system;

   3. the adverse effect is a consequence of the endocrine MoA.

^[#] See Box 2 in Section 2.1.1 for details of Point 2(a-d).

A widely identical wording has been implemented for active substances in plant protection products in the Annex to European Commission (2018).

Point 1(a) of the EDC also clearly states that a substance shall not be considered as having endocrine disrupting properties if any adverse effects observed in laboratory animals are not relevant to humans. Hence, in the absence of evidence demonstrating irrelevance, effects on thyroid function and neurodevelopment observed in animal studies are considered relevant for humans.

Both Point 1(b) and 1(c) of the EDC refer to endocrine MoA (Box 1). Clarification regarding the use of this term is provided in the EFSA and ECHA (2018) Endocrine Disruptor Guidance (p. 7):

It should be highlighted that the “endocrine MoA” as stated in point (b) should be interpreted as “endocrine activity” while the term “endocrine MoA” in point (c) covers the link between the adverse effect and the endocrine activity identified in points a) and b), respectively.

In this context, endocrine activity is defined as follows (EFSA and ECHA 2018, p. 99):

Interaction with the endocrine system that can potentially result in a response of the endocrine system, target organs and tissues. A substance that has an endocrine activity it has the potential to alter the function(s) of the endocrine system.

The EDC implemented in European Commission (2017, 2018) for biocidal products and plant protection products are also cited in Regulation (EC) No. 1907/2006 concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH; EP and Council 2006). Specifically, in Section 2.3 (Other hazards) of Annex II (Requirements for the compilation of safety data sheets) of the REACH Regulation, it is stated: “Information shall be provided on…. whether the substance is a substance identified as having endocrine disrupting properties in accordance with the criteria set out in Commission Delegated Regulation (EU) 2017/2100 […] or Commission Regulation (EU) 2018/605 […].” Hence, the EDC are also generally applicable to substances under the REACH Regulation.

2. The ECETOC and CLE Proposal for a Thyroid Function-Related Neurodevelopmental Toxicity Testing and Assessment Scheme (Thyroid-NDT-TAS)

The ECETOC and CLE Proposal for a Thyroid-NDT-TAS provides a structural concept that consists of these elements (Figure 1):

Figure 1. Overview of the ECETOC and CLE Thyroid-NDT-TAS (see Figures 2–6 for details). BP: biocidal product; EDC-T: endocrine disruptor criteria for the thyroid modality; MoA: mode-of-action; PPP: plant protection product; REACH: Registration, Evaluation, Authorisation and Restriction of Chemicals; WoE: weight-of-evidence.

• Tier 0: collection and WoE evaluation of all available data to determine whether the substance of interest poses a concern for endocrine disruption via the thyroid modality (T-modality) (see Section 2.1 for details)

• Tier 1–3: follow up on concerns identified in Tier 0 (see Section 2.2 for outline and Sections 2.3 and 2.4 for details)

  • Tier 1, Step 1: initial MoA and human relevance assessment of thyroid-related effects observed in animal studies

  • Tier 1, Step 2: generation and evaluation of supporting in vitro / in silico data, if relevant

  • Tier 2: identification of higher-tier testing needs to generate further thyroid- and/or neurodevelopment-related data, if relevant

  • Tier 3: final MoA and human relevance assessment and final WoE evaluations of all available data to draw a conclusion on whether, or not, the substance of interest meets the EDC-T

2.1. Tier 0: collection and WoE evaluation of all available data

2.1.1. Introduction to Tier 0 and general criteria for WoE evaluations

Figure 2 presents the Tier 0 evaluation scheme to collect all available thyroid-related data (and neurodevelopment-related data, if available) and to conduct a WoE evaluation of the thyroid-related data to determine whether the substance of interest poses a concern for endocrine disruption via the T-modality so that the Thyroid-NDT-TAS should be entered. Hence, Tier 0, as it is described throughout Section 2.1, focusses on the identification of thyroid-related effects. The reasons for this focus are (1) that thyroid-related data will generally be available for the Tier 0 evaluation (Section 2.1.2.1), and (2) that neurodevelopmental findings in the absence of thyroid-related effects (i.e. in the absence of endocrine activity) do not indicate a concern for endocrine disruption via the T-modality (but only DNT) in which case the Thyroid-NDT-TAS is not entered. By comparison, the scenario that both thyroid- and neurodevelopment-related effects are recorded in Tier 0 is described in Section 2.2.2 (i.e. the Thyroid-NDT-TAS is entered to determine whether the endocrine activity and the adverse effect are linked by a thyroid-related MoA and whether this MoA is (not) relevant to humans). See Table Appendix 1, which is included in this article after the bibliography, for possible scenarios for the Tier 0 in vivo database.

Figure 2. Tier 0: Evaluation of all available data to decide on the need to enter the ECETOC and CLE Thyroid-NDT-TAS.

EDC-T: endocrine disruptor criteria for thyroid modality; HP: histopathology; H.R.: human relevance; MoA: mode-of-action; MTD: maximum tolerated dose; T-modality: thyroid modality for endocrine disruption; TH: thyroid hormone; WoE: weight-of-evidence.

Colour legend: dark grey boxes: types of substances; light grey boxes, from right to left: production volumes (tonnage ranges) as per REACH Annexes VII-X, respectively; light blue boxes: elements of the assessment; blue arrows: continuation of evaluation; green arrows and text: negative findings; green circle: conclusion from Tier 0 evaluation that the EDC-T are not met. Yellow shape: conclusion from Tier 0 to enter the Thyroid-NDT-TAS.

[a] See Section 2.1.2.2 for elements to consider when applying expert judgement to determine whether the maximum tolerated dose was reached or exceeded.

[b] See Section 2.1.2.3 for elements to consider during the WoE evaluation of the in vivo thyroid-related findings.

[c] See Section 2.1.3 for elements to consider during the WoE evaluation of in vitro mechanistic data and to conclude that there is no evidence for in vitro activity.

[d] See Section 2.1.4 for elements to consider during the WoE evaluation of in silico data and to conclude that there is no evidence for in silico structural alerts.

[e] In vitro negative includes in vitro effects that only occurred at dose levels exceeding the in vivo top doses (as determined via in vitro-to-in vivo extrapolations).

[f] See Section 2.1.5 for aspects to consider in determining whether the in vivo database is sufficient. Inconsistent results in vivo vs. in vitro/in silico includes the scenarios “in vivo negative (in vivo database insufficient) combined with in vitro/in silico positive” and “in vivo positive combined with in vitro/in silico negative.”

[g] Respect information requirements for REACH substances depending on tonnage band and applicability of the European Commission (2017, 2018) Endocrine Disruptor Criteria and EFSA and ECHA (2018) Endocrine Disruptor Guidance.

Potentially relevant thyroid-related data that should be gathered for the Tier 0 evaluation include available data from in vivo studies (Section 2.1.2), and these are often complemented by available data from in vitro mechanistic assays (Section 2.1.3) and possibly also by non-testing information from in silico modelling (Section 2.1.4). Point 2(a) of the European Commission (2017, 2018) EDC also requests that the assessment is based on all available in vivo, in vitro and in silico data (Box 2). This may include information from read-across, i.e. “the use of relevant information from analogous substance(s) (the “source” information) to predict properties for the ‘target’ substance(s) under consideration” (ECHA 2017). However, read-across is not further discussed in this manuscript since it focusses on the collection, generation and evaluation of substance-specific data.

Box 2. Collection of all available data and WoE evaluation to identify whether a substance has endocrine disrupting properties (European Commission 2017, 2018).

For active substances in biocidal products, Point 2 of Section A (Endocrine-disrupting properties with respect to humans) of the Annex to European Commission (2017) states:

(2) The identification of a substance as having endocrine-disrupting properties that may cause adverse effect in humans in accordance with point (1) [see Box 1] shall be based on all of the following points:

1. all available relevant scientific data (in vivo studies or adequately validated alternative test systems predictive of adverse effects in humans or animals; as well as in vivo, in vitro, or, if applicable, in silico studies informing about endocrine MoAs):

   1. scientific data generated in accordance with internationally agreed study protocols, in particular those referred to in Annexes II and III of Regulation (EU) No 528/2012;

   2. other scientific data selected applying a systematic review methodology;

2. an assessment of the available relevant scientific data based on a WoE approach in order to establish whether the criteria set out in point (1) [see Box 1] are fulfilled; in applying the WoE determination, the assessment of the scientific evidence shall, in particular, consider all of the following factors:

   1. both positive and negative results;

   2. the relevance of the study designs for the assessment of adverse effects and of the endocrine MoA;

   3. the quality and consistency of the data, considering the pattern and coherence of the results within and between studies of a similar design and across different species;

   4. the route of exposure, toxicokinetic and metabolism studies;

   5. the concept of the limit dose, and international guidelines on maximum recommended doses and for assessing confounding effects of excessive toxicity;

3. using a WoE approach, the link between the adverse effect(s) and the endocrine MoA shall be established based on biological plausibility, which shall be determined in the light of current scientific knowledge and under consideration of internationally agreed guidelines;

4. adverse effects that are non-specific secondary consequences of other toxic effects shall not be considered for the identification of the substance as endocrine disruptor.

A widely identical wording has been implemented for active substances in plant protection products in the Annex to European Commission (2018).

Following Tier 0 of the Thyroid-NDT-TAS, these different types of data (i.e. data from in vivo studies, in vitro assays and/or in silico modelling) are first submitted to separate WoE evaluations and then to an overarching WoE evaluation. Based thereupon, it is determined whether the substance of interest poses a human health concern for endocrine disruption via the T-modality. All WoE evaluations included in Tier 0 (just as those which form part of Tier 1–3) should be conducted in line with Point 2(b–d) of the EDC (Box 2). Accordingly, the WoE evaluation should consider both positive and negative results, the relevance of study designs, the quality and consistency of the data, the route of exposure, toxicokinetic and metabolism studies as well as the concept of the limit dose and “international guidelines on maximum recommended doses and for assessing confounding effects of excessive toxicity” (Section 2.1.2.2). Based upon the WoE evaluation, “the link between the adverse effect(s) and the endocrine MoA shall be established based on biological plausibility” (Point 2(c) of the EDC), and “adverse effects that are non-specific secondary consequences of other toxic effects” (Point 2(d) of the EDC) shall be ruled out.

The ECETOC T4 TF and CLE recommend conducting all WoE evaluations following pre-defined approaches thereby enhancing their transparency, objectivity and consistency. Such pre-defined approaches may be based on the ECHA templates for WoE and uncertainty evaluation in risk assessment (https://www.echa.europa.eu/web/guest/support/guidance-on-reach-and-clp-implementation/formats [accessed 2023 May]) and/or the EFSA Scientific Committee Scientific opinion on the guidance on the use of the WoE approach in scientific assessments (EFSA SC 2017).

2.1.2. Tier 0: collection and WoE evaluation of all available in vivo data

2.1.2.1. In vivo database that is generally available for Tier 0 WoE evaluation

For all substances that may need to be assessed for endocrine disrupting properties (i.e. active substances in biocidal products and plant protection products as well as REACH substances), mandatory information requirements have been implemented in the EU (EP and Council 2006, 2012; European Commission 2013). In accordance with the applicable legislation (Table 1), data from repeated dose toxicity studies and at least screening-level developmental and reproductive toxicity studies in rats are generally available for the Tier 0 evaluation (i.e. except for the very low tonnage REACH substances, which are generally assumed to have very low exposure potential). Further, a multi-generation reproductive toxicity study is generally available for active substances in biocidal products and plant protection products. All corresponding Organisation for Economic Co-operation and Development (OECD) Test Guidelines (TGs) include mandatory assessments of the thyroid gland, i.e. gross inspection, measurements of absolute and relative organ weight and histopathological investigations (for most recent versions of all OECD TGs, see https://www.oecd.org/chemicalsafety/testing/oecdguidelinesforthetestingofchemicals.htm [accessed 2023 May]). Measurements of serum T4 are mandatory in the 90-day repeated dose toxicity study (OECD TG 408, version of 2018), in the pre-natal developmental toxicity study (OECD TG 414, version of 2018), in the reproduction and developmental toxicity screening tests (OECD TG 421/422, versions of 2016) and in the extended one-generation reproductive toxicity study (EOGRTS; OECD TG 443, version of 2018). Measurements of serum triiodothyronine (T3) are only mandatory in the 90-day repeated dose toxicity study and in the prenatal developmental toxicity study, and measurements of thyroid stimulating hormone (TSH) are mandatory in these two studies as well as in the EOGRTS (Table 2). Please see Beekhuijzen et al. (2019), Li AA et al. (2019) and Marty et al. (2021, 2022) for in-depth discussions of issues to be considered during thyroid hormone measurements.

Table 1. Overview of in vivo database that will generally be available for different types of substances during the Tier 0 evaluation.

Test method	Active substances in biocidal products [a]	Active substances in plant protection products [b]	Substances regulated under the REACH Regulation [c]
			Annex VII: ≥ 1 tonne	Annex VIII: ≥ 10 tonnes	Annex IX: ≥ 100 tonnes	Annex X: ≥ 1,000 tonnes
28-day short-term repeated dose toxicity study (OECD TG 407)	✓, preferred species rat	✓, where available	REACH Annex VII for substances manufactured or imported in quantities of 1 tonne or more does not include standard information requirements for repeated dose, developmental or reproductive toxicity, or toxicokinetics. Nonetheless, existing data concerning these endpoints have to be submitted.	✓, unless reliable 90-day study available	✓	✓
90-day subchronic repeated dose toxicity study (OECD TG 408)	✓, preferred species rat	✓, rodents (usually rats) and non-rodent species		✓, only if indicated by specific concerns	✓	✓
Long-term (≥ 12 months) repeated dose toxicity study (e.g. OECD TG 452)	✓, preferred species rat	✓, rat (and mouse, unless shown that not necessary)		Generally, not available	Generally, not available	✓, only under certain conditions
Reproduction/developmental screening study (OECD TG 421 or 422) [d]	May be available as range finding study for 90-day or developmental toxicity study, but not mandatory	Generally, not available		✓, preferred species rat	✓, preferred species rat	✓, preferred species rat
(Prenatal) developmental toxicity study (e.g. OECD TG 414) [d]	✓, initially one species, “preferred species rabbit,” OECD TG 414 rat study will generally be available	✓, rabbit and rat; OECD TG 414 rat study will generally be available		Generally, not available, unless “serious concerns about potential adverse effects on sexual function, fertility or development”	✓, one species, generally rat, second species as relevant	✓, second species if first species negative
Two-generation reproduction toxicity study (OECD TG 416) or EOGRTS (OECD TG 443)	✓, rat preferred species, EOGRTS (with Cohort 1 A & 1B including 2^nd generation) preferred test method for new studies	✓, rat (majority of compounds: OECD TG 416; EOGRTS only if conducted recently)		Generally, not available, unless “serious concerns about potential adverse effects on sexual function, fertility or development”	✓, EOGRTS (Cohorts 1 A and 1B without F2 generation, one species), if concerns [e]	✓, EOGRTS preferred test method
Developmental neurotoxicity study (OECD TG 426)	✓, or any relevant study (set) providing equivalent information, or Cohorts 2 A and 2B of an EOGRTS with additional investigation for cognitive functions	Generally, not available		Generally, not available	Generally, not available	Generally, not available
Toxicokinetics study (OECD TG 417)	“The toxicokinetics and metabolism studies should provide basic data about the rate and extent of absorption, the tissue distribution and the relevant metabolic pathway including the degree of metabolism, the routes and rate of excretion and the relevant metabolites. “	“Information on blood and tissues concentration of the active substance and relevant metabolites… shall be generated in short and long-term studies on relevant species to enhance the value of the toxicological data generated in terms of understanding the toxicity studies.”		“Assessment of the toxicokinetic behaviour of the substance to the extent that can be derived from the relevant available information”	No additional standard information requirements [f]	No additional standard information requirements [f]

EOGRTS: extended one-generation reproductive toxicity study; GD: gestational day; LD: lactational day; OECD: Organisation for Economic Co-operation and Development; REACH: Registration, Evaluation, Authorisation and Restriction of Chemicals; TG: Test Guideline.

[a] Database available for active substances under the Biocidal Products Regulation (EP and Council 2012, consolidated version of 15 April 2022), as per “Core Data Set” presented in its Annex II (Information requirements for active substances), Title 1, considering only information requirements related to Section 8.8 (toxicokinetics and metabolism data in mammals), Section 8.9 (repeated dose toxicity) and 8.10 (reproduction toxicity). Waiving of reproduction toxicity testing is possible as per Section 8.10 if “the substance is of low toxicological activity (no evidence of toxicity seen in any of the tests available provided that the dataset is sufficiently comprehensive and informative), it can be proven from toxicokinetic data that no systemic absorption occurs via relevant routes of exposure (e.g. plasma or blood concentrations below detection limit using a sensitive method and absence of the substance and of metabolites of the substance in urine, bile or exhaled air) and the pattern of use indicates that there is no or negligible human or animal exposure.”

[b] Database for active substances under the Plant Protection Products Regulation (EP and Council 2009) as per data requirements set out in the Annex to Commission Regulation (EU) No 283/2013 (European Commission 2013, consolidated version 17 November 2014), considering only data requirements related to Annex, Section 5.1 (studies on absorption, distribution, metabolism and excretion in mammals), Section 5.3 (short-term toxicity), Section 5.5 (long-term toxicity) and Section 5.6 (reproductive toxicity).

[c] Database for substances under the REACH Regulation (EP and Council 2006, consolidated version of 14 October 2022), as per standard information requirements set out in its Annexes VII, VIII, IX and X for substances manufactured or imported in quantities of ≥ 1 tonne, ≥ 10 tonnes, ≥ 100 tonnes and ≥ 1,000 tonnes, respectively, considering only information requirements related to the corresponding Sections 8.6 (repeated dose toxicity), 8.7 (reproduction toxicity) and 8.8 (toxicokinetics). For reproduction toxicity, these Annexes include the following rules for adaptation from the standard information requirements:  • Annex VIII, Section 8.7: The OECD TG 421/422 does not need to be conducted when there is no relevant human exposure.   • Annex IX and Annex X, Section 8.7: The reproductive toxicity studies do not need to be conducted if: “the substance is of low toxicological activity (no evidence of toxicity seen in any of the tests available), it can be proven from toxicokinetic data that no systemic absorption occurs via relevant routes of exposure (e.g. plasma/blood concentrations below detection limit using a sensitive method and absence of the substance and of metabolites of the substance in urine, bile or exhaled air) and there is no or no significant human exposure.” (See also Footnote [e] below for conditions under which the EOGRTS is required as per Annex IX.).

[d] The (combined repeated dose toxicity with the) reproduction/developmental toxicity screening test (OECD TG 421/422, version of 29 July 2016) includes exposure from at least 14 days pre-mating until LD 13 and mandatory measurements of thyroid weight, histopathology and serum thyroid hormones. The prenatal developmental toxicity study (OECD TG 414, version of 25 June 2018) includes substance exposure from GD 6 – GD 19 in the rat and mandatory measurements of thyroid weight and histopathology as well as serum thyroid hormone measurements.

[e] Conditions under which EOGRTS is required as per REACH Annex IX standard information requirements: “the available repeated dose toxicity studies indicate adverse effects on reproductive organs or tissues or reveal other concerns in relation with reproductive toxicity.” Further, as per rules for adaptation of the standard information requirements:  • the extension of Cohort 1B to include the F₂ generation may be necessary if “(a) the substance has uses leading to significant exposure of consumers or professionals…, and (b) any of the following conditions are met: the substance displays genotoxic effects in somatic cell mutagenicity tests in vivo…, or there are indications that the internal dose for the substance and/or any of its metabolites will reach a steady state in the test animals only after an extended exposure, or there are indications of one or more relevant MoAs related to endocrine disruption from available in vivo studies or non-animal approaches.”  • the inclusion of the DNT Cohort 2 A may be necessary if “existing information on the substance itself derived from relevant available in vivo or non-animal approaches (e.g. abnormalities of the CNS, evidence of adverse effects on the nervous or immune system in studies on adult animals or animals exposed prenatally), or specific mechanisms/MoAs of the substance with an association to (developmental) neurotoxicity and/or (developmental) immunotoxicity (e.g. cholinesterase inhibition or relevant changes in thyroidal hormone levels associated to adverse effects), or existing information on effects caused by substances structurally analogous to the substance being studied, suggesting such effects or mechanisms/MoAs.”

[f] As per REACH Annex I (General provisions for assessing substances and preparing chemical safety reports), Section 11 (Toxicological information), the safety data sheet shall include “where appropriate information on toxicokinetics, metabolism and distribution.” Also, as per specific rules for adaptation from the standard information requirements implemented in REACH Annexes VII-X, toxicokinetic data can be used to support, e.g. the waiving of repeated dose and/or reproduction toxicity studies, e.g. when “no systemic absorption occurs via relevant routes of exposure.”

Table 2. Measurements of serum T3, T4 and TSH in OECD test guidelines.

Test method	OECD TG, date version was adopted	T4	T3	TSH
28-day short-term repeated dose toxicity study	OECD TG 407 3 October 2008	“…it may be helpful to retain plasma or serum samples to measure T3, T4 and TSH (optional) if there is an indication for an effect on the pituitary-thyroid axis”
90-day subchronic repeated dose toxicity study	OECD TG 408 27 June 2018	Mandatory at study termination
(Prenatal) developmental toxicity study	OECD TG 414 25 June 2018	Dams: mandatory at study termination (GD 20/21)
Two-generation reproduction toxicity study	OECD TG 416 22 January 2001	No mandatory measurements of T3, T4 or TSH
(Combined repeated dose toxicity with the) reproduction/ developmental screening test	OECD TG 421/422 29 July 2016	Mandatory: PND 13 pups, adult males Optional: Dams, PND 4 pups	Optional	Optional
Developmental neurotoxicity (DNT) study	OECD TG 426 16 October 2007	No mandatory measurements of T3, T4 or TSH
Extended one-generation reproductive toxicity study (EOGRTS)	OECD TG 443 25 June 2018	Mandatory: PND 22 pups, P0/F1 adults at study termination Optional: PND 4 pups	Not mentioned	Mandatory: PND 22 pups, P0/F1 adults at study termination
Long-term (≥12 months) repeated dose toxicity study	OECD TG 451–453 27 June 2018	No mandatory measurements of T3, T4 or TSH

GD: gestational day; PND: postnatal day; T3: triiodothyronine; T4: thyroxine; TG: Test Guideline; TSH: thyroid stimulating hormone.

In accordance with the OECD (2012) Conceptual Framework for testing and assessment of endocrine disruptors and the OECD (2018) Guidance Document No. 150 on standardised test guidelines for evaluating chemicals for endocrine disruption, these OECD TGs are assigned to different levels that reflect the extent of information they can provide for the assessment of endocrine disruption:

Level 5 studies are defined as “in vivo assays providing more comprehensive data on adverse effects on endocrine relevant endpoints over more extensive parts of the life cycle of the organism” (OECD 2012). Two OECD TGs have been identified as Level 5 studies, the EOGRTS (OECD TG 443) and the two-generation reproductive toxicity study (OECD TG 416; version of 2001).

Level 4 studies are defined as “in vivo assays providing data on adverse effects on endocrine relevant endpoints” (OECD 2012). Level 4 studies include, amongst others, the 28- and 90-day repeated dose toxicity studies (OECD TG 407 and 408), the prenatal developmental toxicity study (OECD TG 414), the (combined repeated dose toxicity study with the) reproduction and developmental toxicity screening studies (OECD TG 421/422), if optional thyroid endpoints are included, the DNT study (OECD TG 426, version of 2007) as well as the pubertal development and thyroid function assays in peripubertal male and female rats (Environmental Protection Agency (EPA) Office of Chemical Safety and Pollution Prevention (OCSPP) Guidelines 890.1450 and 890.1500; https://www.epa.gov/sites/default/files/2019-10/documents/ocspp-testguidelines_masterlist-2019-09-24.pdf [accessed 2023 May]).

Level 3 studies include the uterotrophic assay and the Hershberger assay, which are generally not relevant for the T-modality.

Level 2 assays cover in vitro assays (Section 2.1.3).

Level 1 includes physico-chemical information, in silico modelling (Section 2.1.4) and other non-testing approaches.

The EFSA and ECHA (2018) Endocrine Disruptor Guidance has adopted the OECD (2012, 2018) scheme to assign different types of studies and information to these same levels.

2.1.2.2. Consideration of maximum tolerated dose

According to the Tier 0 scheme to decide on the need to enter the Thyroid-NDT-TAS, the WoE evaluation of the available in vivo data considers whether thyroid-related effects, if present, were only observed above the maximum tolerated dose. If this is the case, the effects are considered “adverse effects that are non-specific secondary consequences of other toxic effects” (Point 2(d) of the EDC) and thus do not necessarily indicate a concern for endocrine disruption via the T-modality (further discussed below). The EFSA and ECHA (2018) Endocrine Disruptor Guidance lists elements to consider in evaluating whether the maximum tolerated dose was reached in a study and emphasises that expert judgement is required for this evaluation (p. 21):

Elements to consider are alterations in physiological function, including: no more than 10% decrease in body weight gain relative to control, target organ toxicity and alterations in clinical pathological parameters. Although these parameters can only be considered indicative and expert judgement is necessary to define the maximum tolerated dose on a case-by-case basis. Elements which indicate that the maximum tolerated dose has been exceeded are reported in the OECD Guidance on the Recognition, Assessment and Use of Clinical Signs as Humane Endpoints for Experimental Animals Used in Safety Evaluation (OECD 2000).

Guidance related to the maximum tolerated dose has been published in several OECD documents. In the OECD (2002) Guidance notes for analysis and evaluation of chronic toxicity and carcinogenicity studies, the maximum tolerated dose is defined as the highest dose to produce toxic effects without causing death or significant morbidity and a no more than 10% decrease in body weight relative to controls. OECD TGs generally recommend using either a limit dose of 1,000 mg/kg body weight/day or the maximum tolerated dose as top dose. Further, the OECD (2018) Guidance Document No. 150 on standard test guidelines for evaluating chemicals for endocrine disruption highlights that endocrine effects observed in the presence of clear systemic toxicity are unlikely to be due to endocrine activity:

The top dose or concentration should be sufficiently high to give clear systemic (i.e. non endocrine-specific) toxicity in order to ensure that a wide range of exposures (high to low) is tested. However, endocrine effects observed solely in the presence of clear systemic toxicity should be interpreted with caution and may be disregarded when sufficiently justified to be caused by secondary effects which are unlikely to be due to endocrine activity.

More recently, Sewell et al. (2022), building on concepts developed in the ECETOC (2021) Technical Report No. 138 on Guidance on dose selection, recommended that the highest dose in repeated dose toxicity studies should cause “minimal but evident toxicity to the test animals without significantly compromising their well-being.” Sewell and co-authors highlighted the importance of understanding the human relevance of kinetics to inform study design:

For animal data, a translational understanding of human relevance of kinetics is important to inform study design in relation to several scopes, including hazard identification, risk assessment […] Typically, the risk of non-relevance may be higher at higher dose levels, where: A) high systemic exposures may disrupt physiological detoxification processes or other homeostatic processes leading to overt toxicity, potentially confounding appropriate evaluation of the toxicological results, and B) high systemic exposures may be quantitatively and qualitatively different from potential human systemic exposure.

The ECETOC (2021) Technical Report includes further details on “pragmatic approaches to selecting dose levels that allow accurate risk assessment and also enable hazard-based classification based on identification of relevant hazards.”

Beyer et al. (2011) provided recommendations from International Life Sciences Institute/Health and Environmental Sciences Institute (ILSI/HESI) workshops on how to establish maternal toxicity in developmental and reproductive toxicity studies. Regarding the maximum tolerated dose, Beyer and colleagues concluded that “a decrease in body weight gain of 20% [in pregnant animals] was considered excessive for most test articles/test materials.”

A European Society of Toxicologic Pathology expert workshop provided advice regarding the types of liver effects that indicate that the maximum tolerated dose was likely exceeded in studies investigating liver enzyme inducers (Hall et al. 2012):

A dose level of a xenobiotic that in short-term tests induced either structural or biochemical evidence of hepatocellular damage, or produced increases in liver weight of approx. ≥150% would be considered adverse in the context of dose setting and exceed the maximum tolerated dose.

Hall et al. (2012) also indicated which types of liver effects should be considered non-adverse adaptive reactions: “Hepatomegaly as a consequence of hepatocellular hypertrophy without histologic or clinical pathology alterations indicative of liver toxicity was considered an adaptive and a non-adverse reaction.”

Taken together, the ECETOC T4 TF and CLE conclude that expert judgement is required to determine on a case-by-case basis whether the maximum tolerated dose was reached or exceeded in a study. Use of the kinetically derived maximum dose, i.e. the highest dose at kinetically linear doses or slightly above the point-of-departure from linear kinetics, has been suggested as an alternative to the maximum tolerated dose to address potential discrepancies between external and internal doses (Terry et al. 2016; Sewell et al. 2020; Felter et al. 2021). While a further discussion of the concepts of the maximum tolerated and kinetically derived maximum doses would exceed the scope of the present article, the ECETOC T4 TF and CLE recommend considering toxicokinetic data and/or physiologically based kinetic (PBK) modelling in establishing meaningful dose ranges for subchronic or longer-term studies.

2.1.2.3. Recommendations for the weighting of in vivo findings

During Tier 0 to decide on the need to enter the Thyroid-NDT-TAS, the available in vivo thyroid-related data are evaluated to determine whether there is a concern for endocrine disruption via the T-modality. Hence, the available in vivo data are evaluated to determine whether the substance of interest elicits effects on the thyroid gland and/or on serum thyroid hormone levels. Thereby, the in vivo data may provide information related to endocrine activity (Point 1(b) of the EDC) and a potential thyroid-related MoA (Point 1(c) of the EDC). If neurodevelopmental data from the EOGRTS (OECD TG 443) or DNT study (OECD TG 426) are available during Tier 0, they provide information on the substance’s potential to elicit an adverse effect (Point 1(a) of the EDC). In this case, the Thyroid-NDT-TAS is entered to determine whether the endocrine activity and the adverse effect are linked by a thyroid-related MoA (Point 1(c) of the EDC) and whether this MoA is not relevant to humans (Section 2.2.2).

If effects on the thyroid gland and/or altered serum T4 (TSH and T3) levels are recorded in Tier 0, they are jointly evaluated in a WoE approach following the general criteria for WoE evaluations outlined in Section 2.1.1. Further, the ECETOC T4 TF and CLE recommend considering the following elements in the WoE evaluation:

• Absence of effects on thyroid weight and histopathology observed in 28-day (or longer-term) repeated dose toxicity studies is generally sufficient to conclude that the EDC-T are not met even if serum thyroid hormone data are unavailable. Justification: Prolonged thyroid perturbations lead to compensatory reactions of the thyroid gland regardless of the substance’s thyroid-related MoA (further discussed in Section 2.3.2). In this regard, the EFSA (2020) Technical Report states (p. 7): “…the dataset for thyroid can be considered complete on a case-by-case basis, pending whether the duration and doses selection allow a proper assessment of the thyroid histopathology (thyroid histopathology is generally considered more sensitive and informative than thyroid weight).” This premise stands in line with findings from the EFSA (2019a) Scientific Report Establishment of cumulative assessment groups of pesticides for their effects on the thyroid. Of 127 active substances causing T4/T3 decrements and/or TSH increases, only two did not also alter relative thyroid weight or thyroid gland histopathology, and EFSA concluded that these two active substances were unlikely to cause hypothyroidism (EFSA 2019a).

• Thyroid-related effects are weighted lower if only one of several thyroid-related parameters is altered.

• Altered thyroid weight/histopathological findings are weighted lower if only observed in one (or few) out of several studies.

• Altered serum T4, T3 and/or TSH levels are weighted lower if observed in studies including small group sizes (e.g., n < 10). Justification: Due to the versatility of the hormone system, hormone data that are based on such small group sizes generally have weak statistical power. Assuming group sizes of 10 animals, T3/T4 decreases by approx. 25% and a TSH increase of approx. 40% as compared to the concurrent controls can be detected as statistically significant (for details, see Li AA et al. 2019; Marty et al. 2021). Appendix B of the EFSA and ECHA (2018) Endocrine Disruptor Guidance provides Recommendations for design, conduction and technical evaluation of hormonal studies.

• If thyroid effects are observed in more than one species (e.g. rats, rabbits and/or dogs) in studies with similar design, such concordances may add weight to the findings. However, if thyroid effects are only observed in one of several species, the aim should be to explain these using, e.g. information on absorption, distribution, metabolism, elimination (ADME) and/or data from in vitro mechanistic assays. Similarly, data from the amphibian metamorphosis assay (OECD TG 231, version of 2009) or the Xenopus eleutheroembryonic thyroid assay (OECD TG 248, version of 2019), as two ecotoxicological Level 3 assays (OECD 2012, 2019), may support WoE evaluations as the hypothalamic-pituitary-thyroid axis is highly conserved across vertebrates (see also EFSA and ECHA 2018, p. 9 and 102).

The WoE evaluation should also consider possible differences in sensitivity to substance-mediated thyroid perturbation (1) between adult males and/or non-pregnant/non-lactating females vs pregnant/lactating females; and (2) between pregnant/lactating females vs foetuses and/or new-born pups.

The ECETOC T4 TF and CLE also recommend considering the following elements to support the WoE evaluation of the available in vivo data. These elements reflect observational findings from Marty et al. (2022) who evaluated all thyroid- and brain-related data from 51 rat studies that included in utero/lactational exposure to substances causing thyroid hormone imbalance:

• Consider weighting offspring serum T4 decrements higher if they are ≥60% and ≥50% in the top- and lower-dose groups, respectively. Justification: Marty et al. (2022) found that offspring serum T4 decrements exceeding these thresholds indicated an increased likelihood for statistically significant adverse neurodevelopmental effects.

• Consider weighting offspring serum T3 decrements higher if they are ≥20% and statistically significant. Justification: Marty et al. (2022) found that offspring serum T3 decrements exceeding this threshold indicated an increased likelihood for statistically significant adverse neurodevelopmental effects.

• Consider weighting offspring serum TSH increases higher if they are ≥400%. Justification: Marty et al. (2022) found that thyroid peroxidase (TPO) and sodium – iodide symporter (NIS) inhibitors (i.e. substances with a direct thyroid-related MoA; Section 2.1.3) that elicited statistically significant adverse neurodevelopmental effects in the respective studies (that included in utero/lactational substance exposure) also elicited dose-dependent and statistically significant offspring serum TSH increases ≥ 400%.

• Consider weighting thyroid-related effects lower if offspring serum T4 decrements attenuate over the course of an in utero/lactational exposure period. Justification: Marty et al. (2022) found that attenuations of offspring serum T4 decrements over the course of in utero/lactational exposure periods were associated with absence of statistically significant adverse neurodevelopmental effects.

• Consider weighting thyroid-related effects lower if they are not accompanied by offspring body weight changes. Justification: Marty et al. (2022) found that absence of offspring body weight changes was associated with absence of statistically significant adverse neurodevelopmental effects.

The ECETOC T4 TF and CLE recommend making use of future assessments of developmental and reproductive toxicity addressing the T-modality to further investigate the relevance and reliability of these empirical parameters and thresholds observed by Marty et al. (2022) and hence their possible applicability and robustness for regulatory assessments.

The database considered by Marty et al. (2022) was insufficient to conclude on the suitability of maternal or offspring serum free T4 (fT4)/free T3 (fT3) data to predict the likelihood for statistically significant neurodevelopmental findings, let alone to suggest thresholds for fT4/fT3 alterations indicating adverse outcomes. Serum fT4 is the most frequent thyroid-related parameter measured in human studies addressing maternal thyroid function and child neurodevelopment (Sauer et al. 2020). The ECETOC T4 TF and CLE recommend considering the generation of offspring serum fT4/fT3 data in future rodent developmental and reproductive toxicity studies to support the further investigation of these parameters.

Finally, a comprehensive evaluation of a substance’s potential to elicit thyroid perturbations should also consider its toxicokinetic properties (Marty et al. 2022). The Tier 0 WoE evaluation considers all available toxicokinetics data to ensure that any observed effects reflect toxicological properties of the substance at doses that are relevant to the potentially exposed human population (Tan et al. 2021), with due consideration of the toxicokinetics-related part of the uncertainty factors that may be applied on a case-by-case basis. This stands in line with Point 2(b)(iv) of the EDC (Section 1.2).

2.1.3. Tier 0: collection and WoE evaluation of available in vitro data

During Tier 0 to decide on the need to enter the Thyroid-NDT-TAS, all available in vitro mechanistic data are collected and evaluated to record possible endocrine activity via the T-modality. A spectrum of in vitro assays has been developed to investigate whether substances may trigger the molecular initiating events (MIEs) or subsequent key events of the most important thyroid-related MoAs, including (1) inhibition of NIS, which mediates uptake of iodide into the thyroid gland as first step in thyroid hormone synthesis; (2) inhibition of TPO as enzyme mediating thyroid hormone synthesis; (3) induction of liver enzymes mediating thyroid hormone metabolism (predominantly: uridine diphosphate glucuronyltransferase (UGT)); and (4) displacement of thyroid hormones from serum binding proteins. The latter two key events may lead to enhanced thyroid hormone clearance and thus form part of indirect thyroid-related MoAs. By comparison, TPO inhibition and NIS inhibition, which both affect thyroid hormone synthesis, are MIEs for two direct thyroid-related MoAs. Less frequent MIEs of thyroid-related MoAs that can be investigated in vitro include the inhibition of different deiodinases (DIOs), as enzymes mediating thyroid hormone metabolism, and substance interaction with thyroid hormone receptors; see Table 3 for an overview of potentially relevant in vitro assays and Noyes et al. (2019) and Marty et al. (2021) for comprehensive reviews of the scientific evidence on the MIEs and key events of thyroid-related MoAs.

Table 3. Overview of in vitro assays, chemical methodologies and in silico models that allow investigating MIEs or key events of thyroid-related MoAs in mammals.

In vitro assay, chemical assay, in silico model, PBK model	MIE or key event addressed	Test system (from rat and human = facilitates H.R. assessment)	EU TSAR status [a]	EPA: HTS readiness [b]	References
Radioactive-iodide uptake assay using human hNIS-HEK293T-EPA cells	NIS inhibition (MIE of AOP 54, 134)	Rat and human cell lines	NA	Existing	Hallinger et al. (2017) Wang et al. (2018)
Radioactive-iodide uptake inhibition screening assay using the FRTL-5 cell line			NA	Existing	Buckalew et al. (2020)
Activation of NIS based on Sandell–Kolthoff reaction		Human liver microsomes as per Zuang et al. (2022)	ThyM 2d: Val. ongoing	NA	TSAR [a]; Zuang et al. (2022); JRC (2023)
Amplex UltraRed-TPO assay	TPO inhibition (MIE of AOP 42)	Paul Friedman et al.: rat microsomes ThyM 2a: FTC-238 cells transfected with human recombinant TPO	ThyM 2a: Val. finalised, evaluation ongoing [c]	Existing	Paul Friedman et al. (2016); TSAR [a]; JRC (2023)
Guaiacol oxidation assay		Paul et al.: rat and pig microsomes	NA	NA	Paul et al. (2013)
Characterisation of distinct multiple catalytic activities of TPO (LC-MS/MS)		Tater et al.: rat microsomes			Tater et al. (2021)
Tyrosine iodination using LC	TPO-mediated tyrosine iodination (MIE of AOP 42)	FTC-238 cells transfected with human recombinant TPO	ThyM 2c: Val. finalised, evaluation ongoing [c]	NA	TSAR [a]; JRC (2023)
QSAR	TPO inhibition (MIE of AOP 42)	Limited opportunity for human relevance assessment	NA	NA	Gadaleta et al. (2021)
QSAR	TPO inhibition (MIE of AOP 42) NIS inhibition (MIE of AOP 54, 134) TR inhibition; DIO1/2/3 inhibition (MIEs in AOP network [d])		NA	NA	Garcia de Lomana et al. (2021) Dracheva et al. (2022)
PBK modelling for TPO inhibition	Modelling of AOs in response to potencies of MIEs for AOP 42, 54	Modelling of effects in rats	NA	NA	Handa et al. (2021) Hassan et al. (2017)
In vitro assays to address a substance’s potential to displace TH from TTR, TBG (and/or albumin)	Interaction with serum binding proteins (MIE of AOP 152)	ThyM 3a: human TBG, TTR ThyM 3b: human TTR	ThyM 3a/3b: Val. finalised, evaluation ongoing [c]	Existing	Collet et al. (2020) Zuang et al. (2022) JRC (2023)
In vitro HTS assays to assess chemical binding and activation of specific nuclear receptors, which mediate liver enzyme induction	Liver nuclear receptor activation (MIE of AOP 8)	PXR (Rosenfeld et al.): liver samples from knock out mice; CAR (Maglich et al.): primary human hepatocytes; AhR (He et al.): plasmids transfected into mouse hepatoma Hepa1c1c7 cells	NA	Existing	Noyes et al. (2019) citing: He et al. (2011) Maglich et al. (2003) Romanov et al. (2008) Rosenfeld et al. (2003)
Comparative liver enzyme induction (Phase 2)	UGT induction (key event of AOP 8)	Primary rat and human hepatocytes	NA	Promising	Bomann et al. (2021) Parmentier et al. (2022) Tinwell and Bars (2022) Wiemann et al. (2023)
LC/MS assay addressing inhibition of TH glucuronidation	TH metabolism (relates to key events of AOP 8)	Human liver microsomes	ThyM 4b: Val. finalised, evaluation ongoing [c]	NA	Zuang et al. (2022) JRC (2023)
Thyroid organ-on-a-chip: spheroids of primary human (and/or rodent) thyroid microtissue	Functional assay covering TH synthesis (relates to key events of AOP 42, 54, 134)	Rat and human microtissues	NA	NA	Deisenroth et al. (2020) Moroni et al. (2020)
Liver-thyroid multi-organ-on-a-chip, rat and human	Functional assay covering TH synthesis and transport (relevant for all thyroid-related MoAs)		NA	NA	Karwelat et al. (2023) Kühnlenz et al. (2023)
PBK modelling to estimate serum/brain TH concentrations in rats vs humans	Modelling of functional assay for homeostasis of TH after exposure to TH disruptors (relevant for all thyroid-related MoAs)	Model developed to enable species differences assessments	NA	NA	Leonard et al. (2016)
In vitro inhibition of human DIO1	DIO inhibition (MIE in AOP network [d])	Rat or human liver microsomes	NA	Existing	Renko et al. (2015) Olker et al. (2019) Weber et al. (2022)
DIO1 activity based on Sandell–Kolthoff reaction	DIO activity (MIE in AOP network [d])	Human liver microsomes	ThyM 4a: Val. finalised, evaluation ongoing [c]	NA	TSAR [a]; JRC (2023)
In vitro inhibition of human DIO2 and DIO3	DIO inhibition (MIE in AOP network [d])	Lysate from recombinant cells	NA	Existing	Hornung et al. (2018) Olker et al. (2019)
Inhibition of MCT8 based on Sandell–Kolthoff reaction	TH transport through cell membranes (MIEs in AOP network [d])	MDCK cells stably transfected with human MCT8	ThyM 5a: Val. ongoing	Early	TSAR [a]; JRC (2023)
TR transactivation assays (ThyM 6a: human TRα and TRβ reporter gene transactivation – agonist activities ThyM 6b: CALUX human TRβ reporter gene transactivation – agonist and antagonist activities)	TR agonism/antagonism (MIE in AOP network [d])	ThyM 6a: engineered proprietary human HEK293 cells ThyM 6b: proprietary TRβ CALUX^® cells (origin: human osteoblastic osteosarcoma U2OS line)	ThyM 6a: Val. finalised, evaluation ongoing [c] ThyM 6b: Val. ongoing	Existing	Paul-Friedman et al. (2019) TSAR [a]; JRC (2023)
TSH receptor activation based on cAMP measurement	Integrated tissue responses to alterations in TH levels (MIE in AOP network [d])	CHO-K1 cells transfected with human TSH receptor	ThyM 1b: Val. ongoing	Promising	TSAR [a]; JRC (2023)
Measurement of proliferation of rat pituitary-derived cell line GH3	Interference of test item with TH on cellular level (pituitary weight change)	Rat pituitary-derived cell line GH3	ThyM 8a: Val. ongoing	NA	TSAR [a]; JRC (2023) Jomaa et al. (2013)

AhR: aryl hydrocarbon receptor; AO(P): adverse outcome (pathway); CAR: constitutive androstane receptor; CHO: Chinese hamster ovary (cells); EURL ECVAM: European Union Reference Laboratory for alternatives to animal testing; FRTL: Fischer rat thyroid follicular (cell line); DIO: deiodinase; HTS: high-throughput screening; LC: liquid chromatography; MCT: monocarboxylate transporter; MDCK: Madin Darby Canine Kidney; MIE: molecular initiating event; MoA: mode-of-action; MS: Mass spectrometry; NA: not addressed; NIS: sodium – iodide symporter; PBK: physiologically based kinetic; QSAR: quantitative structure activity relationship; TBG: thyroid binding globulin; TH: thyroid hormone(s); ThyM: thyroid method (key word in TSAR, followed by method number as assigned in TSAR [a]); TPO: thyroid peroxidase; TR: thyroid receptor; TRH: thyrotropin-releasing hormone; TTR: transthyretin; UGT: uridine diphosphate glucuronyltransferase; Val.: validation.

[a] EURL ECVAM TSAR – Tracking System for Alternative methods towards Regulatory acceptance; https://tsar.jrc.ec.europa.eu/ [accessed 24 May 2023].

[b] Adapted from Table 2 in Noyes et al. (2019): “In vitro HTS assay development denoted as “promising” indicates MIEs in the thyroid axis for which there is interest and/or activity in developing in vitro HTS approaches, typically with supportive in vivo and slow- or medium-throughput in vitro toxicity studies indicating chemical interactions. In vitro HTS readiness denoted as “early” indicates putative MIEs but with limited toxicity evidence and with little current activity to develop high-throughput alternatives.”

[c] Validation Study Part 1 (Assessment of method robustness, reliability and within laboratory reproducibility using few test items) and Part 2 (Assessment of the method’s mechanistic relevance using selected validation set chemicals covering the underlying mechanism of this method) finalised. Assessment of the data and study reports and decision if further validation work is needed in planning or ongoing; see JRC (2023) for details on the Thyroid Validation Study.

[d] See Figure 2 in Noyes et al. (2019): “AOP network for chemically induced thyroid activity showing the integration of multiple individual AOPs under development and proposed.”

The in vitro assays are Level 2 assays in accordance with the OECD (2012) Conceptual Framework, i.e. “in vitro assays providing data about selected endocrine mechanism(s)/pathways(s).” Hence, the in vitro data may provide information that is relevant to determine (1) whether the substance of interest has endocrine activity (Point 1(b) of the EDC) and (2) to support the determination of its likely MoA (related to Point 1(c) of the EDC; Box 1).

Regarding regulatory applicability, the validation of in vitro assays to investigate thyroid activity is a matter of extensive research work worldwide. The EU Reference Laboratory for alternatives to animal testing (EURL ECVAM) and the EU Network of Laboratories for the Validation of Alternative Methods (EU-NETVAL) are engaged in the validation of in vitro assays targeting the MIEs and/or key events of all major thyroid-related MoAs (Zuang et al. 2022; JRC 2023). Details on the validation status of the in vitro thyroid methods are provided in the EURL ECVAM Tracking System for Alternative Methods towards Regulatory Acceptance (TSAR; https://tsar.jrc.ec.europa.eu/). As per 24 May 2023, the experimental parts of the validation have been finalised for in vitro assays addressing TPO inhibition, thyroid hormone glucuronidation, substance binding to transthyretin or thyroxin binding globulin, DIO1 activity and thyroid receptor transactivation (see Column “EU TSAR status” in Table 3). All finalised validation study reports and associated data are provided to an ad hoc OECD Expert Group for Thyroid Disruption Methods for data analysis and the decision if further validation work is needed (JRC 2023).

In the USA, the EPA has been using a broad spectrum of in vitro assays that address MIEs or key events for all major thyroid-related MoAs for high-throughput screening (HTS). Noyes and colleagues from the EPA have marked the “HTS readiness” as “existing” for in vitro assays addressing NIS, TPO and DIO inhibition, interaction with serum binding proteins, interaction with hepatic nuclear receptors and transactivation of nuclear thyroid hormone receptors (Noyes et al. 2019; Table 3; Column “EPA: HTS readiness”). These assays have been used within the EPA Endocrine Disruptor Screening Program (https://www.epa.gov/endocrine-disruption) and the EPA research effort ToxCast (https://www.epa.gov/chemical-research/toxicity-forecasting) [both websites accessed 2023 May].

The ECETOC T4 TF and CLE recommend considering all available relevant in vitro data for the Tier 0 evaluation, giving due consideration to the scientific validation status and applicability domains of the underlying methods. The evaluation of in vitro mechanistic data should consider the physico-chemical properties and analytical chemistry data of the test substance to ensure that it was amenable to screening and likely to be present in the test system. Importantly, in vitro effects should only be considered relevant if observed at non-cytotoxic doses. Further, if permitted by the available in vitro and in vivo database, in vitro-to-in vivo extrapolations and/or reverse dosimetry based on PBK modelling should be conducted to determine whether in vitro effects were observed at dose levels reflecting relevant in vivo plasma concentrations (see, e.g. Li H et al. 2017; Louisse et al. 2017). If in vitro effects only occurred at dose levels exceeding the internal/plasma concentrations recorded in the top dose groups from animal studies, the outcome is assessed as negative. If PBK modelling and/or in vivo toxicokinetics data are not available for the substance of interest, the ECETOC T4 TF and CLE recommend only using in vitro effects for the assessments if they were observed at or below the respective assay-specific top concentrations (e.g. ≤ 100 µM as used for ToxCast assays; Judson et al. 2016; Filer et al. 2017; Whalley et al. 2017; Franzosa et al. 2021) as threshold values to optimise the relevance of in vitro findings for human safety assessments.

Finally, when jointly evaluating in vitro data from different assays, it is important to note that these assays generally have high false positive rates (see, e.g. Paul Friedman et al. 2016; Noyes et al. 2019). This suggests high biological relevance of an overall negative outcome of a battery of in vitro assays.

Taken together, the ECETOC T4 TF and CLE recommend that all available and relevant in vitro data on thyroid activity should be considered within the Tier 0 WoE evaluations. Further, it is recommended that “sufficient in vitro database and no evidence for in vitro effects” may be concluded if negative in vitro data are available for MIEs or key events covering all major thyroid-related MoAs that are potentially relevant for the substance of interest. These may be TPO inhibition, NIS inhibition, UGT induction and interference with serum binding proteins. In addition, in vitro effects on DIO interaction and interference with thyroid receptors may need to be considered if deemed relevant based on in silico structural alerts (Section 2.1.4). As described above, HTS assays are available for all these events (Noyes et al. 2019). Even though such HTS assays have generally not yet been formally adopted (e.g. as OECD TG), the ECETOC T4 TF and CLE recommend seeking opportunities to facilitate their applicability within the Thyroid-NDT-TAS. For example, Ball et al. (2022) presented a framework for chemical safety assessment as pragmatic approach to integrate different types of data from TG-conforming and non-TG-conforming studies. This framework allows new in silico models and in vitro assays to be incorporated in the assessment scheme “as they develop through continuous selective evolution rather than periodic revolution.” Within this framework, a simple categorisation scheme for exposure estimation is suggested, which considers both the level and duration of exposure, and a tiered approach for hazard assessment. Importantly, the framework by Ball et al. includes derivation of regulatory decision-making outputs (i.e. classification/categorisation, limit values and exposure estimates) after each tier. Thereby, the level of precision required to make an appropriate safety assessment can be adapted to the given information needs (Ball et al. 2022). The ECETOC T4 TF and CLE suggest following the tiered chemical safety assessment framework described by Ball et al. (2022) when considering whether data from non-TG-conforming studies or assays may be useful for the WoE evaluations embedded in the Thyroid-NDT-TAS.

2.1.4. Tier 0: collection and WoE evaluation of available in silico data

The Tier 0 evaluation of the available data to decide on the need to enter the Thyroid-NDT-TAS should preferably also consider non-testing information from in silico modelling (see, e.g. Garcia de Lomana et al. 2021; Dracheva et al. 2022). Generally, in silico models are Level 1 assays in accordance with the OECD (2012) Conceptual Framework, i.e. “existing data and non-test information.”

In silico models to predict a substance’s potential to trigger the MIEs of thyroid-related MoAs have not yet been formally accepted for regulatory purposes. Nonetheless, it is recommended to seek opportunities to integrate such approaches into regulatory assessments as this may serve the goal to replace, reduce and refine animal testing (Russell and Burch 1959; EP and Council 2010), for example by applying the Ball et al. (2022) framework to incorporate new in silico models and in vitro assays into assessment schemes (Section 2.1.3).

2.1.5. Tier 0: overarching WoE evaluation of all available data and conclusion

After separate WoE evaluations of the available in vivo, in vitro and in silico data, the final step of Tier 0 includes (1) an evaluation of the sufficiency of the in vivo thyroid-related database, and (2) an overarching WoE evaluation of all available in vivo, in vitro and in silico data. Regarding the sufficiency of the in vivo database, the EFSA and ECHA (2018) Endocrine Disruptor Guidance states (p. 31):

To have the T-mediated adversity with regard to humans and mammals (as non-target organisms) sufficiently investigated, the thyroid parameters foreseen to be investigated in the following studies OECD test guidelines 407, 408, 409 (and/or the one-year dog study, if available), 416 (or 443 if available) and 451-3 should have been measured and the results included in the dossier. If there is no indication of effects in these studies, the T modality is considered to be sufficiently covered.

The further clarifications that have since been published in the EFSA (2020) Technical report on the outcome of the pesticides peer review meeting on general recurring issues in mammalian toxicology have detailed that if no effects are observed in comprehensive histopathological assessments of the thyroid gland, this is generally sufficient to conclude that the EDC-T are not met even if serum thyroid hormone data are unavailable. As described in Section 2.1.2.3, this scenario may apply to substances whose database includes older OECD TG versions that did not yet include serum hormone measurements. In addition, the ECETOC T4 TF and CLE suggest that absence of thyroid-related findings in an enhanced OECD TG 421/422 reproduction and developmental toxicity screening test, together with absence of thyroid-related findings upon subchronic or longer-term substance exposure, may be sufficient to conclude that there is no indication for thyroid-related effects (even if in vitro assays were positive). In this context, enhanced implies that the reproduction and developmental toxicity study includes both serum thyroid hormones and thyroid histopathology.

If the in vivo thyroid-related database is sufficient and consistently indicates no substance-mediated effects on thyroid function, there is no evidence for endocrine activity via the T-modality (Point 1(b) of the EDC), and it is concluded that the EDC-T are not met. Accordingly, there is no need to enter the Thyroid-NDT-TAS.

If the in vivo thyroid-related database is not sufficient, or if in vivo thyroid-related effects were observed, all available in vivo, in vitro and in silico data are submitted to an overarching WoE evaluation, which should again follow the general principles for WoE evaluations (Section 2.1.1). This overarching WoE evaluation can yield four different outcomes:

• In vivo, in vitro and in silico thyroid-related data are consistently negative, but insufficient in vivo thyroid-related database: If this scenario applies and the in vitro database is complete and consistently provides negative outcomes for all potentially relevant thyroid-related MoAs (Section 2.1.3), there is no evidence for endocrine activity via the T-modality (Point 1(b) of the EDC), and it is concluded that the EDC-T are not met. This conclusion considers the high false positive rate of the in vitro assays, which enhances the biological relevance of an overall negative in vitro outcome.

• In vivo thyroid-related data are negative (insufficient in vivo thyroid-related database), in vitro and/or in silico data are positive: If this scenario applies, the further evaluation should consider which in vitro assays were positive and whether the discrepancy is explainable (e.g. by in vitro-to-in vivo extrapolations). If the in vitro response was observed at relevant concentrations, the WoE evaluation should aim to establish the level of confidence in the negative in vivo results. It may be considered, for example, whether serum thyroid hormone data are already available, and, if so, whether they were recorded at sensitive stages (e.g. pregnancy) and/or in the offspring. Further, the WoE evaluation may consider whether the compound is rapidly metabolised in vivo.

  • If the WoE evaluation conclusively shows that the in vivo negative findings are biologically more relevant than the in vitro positive findings (e.g. because discrepancies in the in vitro-to-in-vivo extrapolations can be shown), there is no evidence for endocrine activity via the T-modality (Point 1(b) of the EDC), and it is concluded that the EDC-T are not met.

  • If the WoE evaluation indicates that the in vitro positive findings may be biologically more relevant than the in vivo negative findings (e.g. because the in vitro assays were conducted using human test systems and species differences in the metabolism of the substance are suspected), the Thyroid-NDT-TAS should be entered, and the generation of additional data should preferably first focus on the conduct of comparative metabolism assays to further investigate such potential species differences.

  • If the WoE evaluation is inconclusive, the Thyroid-NDT-TAS should be entered. As relevant, the generation of additional data may first focus on the generation of ADME data to help explain the discrepancy between the negative in vivo studies and positive in vitro assays (e.g. when the compound is rapidly metabolised in vivo) and/or on further investigating the MoA that likely underlies the positive in vitro response(s). Preferably, orthogonal assays, i.e. assays including different endpoint detection methods, should be applied for this purpose (Section 2.3.4).

• In vivo thyroid-related data are positive, but in vitro/in silico data are negative: If this scenario applies, the Thyroid-NDT-TAS should be entered to evaluate the sufficiency of the in vitro database to assess all potentially relevant thyroid-related MoAs. Further, the WoE evaluation may consider whether in vivo findings were likely due to an active metabolite that was not formed in vitro. Accordingly, the added value of comparative in vitro metabolism assays, in silico modelling and/or read-across should be considered to investigate whether active metabolites recorded in animal studies may also be formed in humans.

• In vivo, in vitro and in silico thyroid-related data are consistently positive, i.e. they indicate a concern for endocrine disruption via the T-modality: If this scenario applies, the Thyroid-NDT-TAS should be entered to conduct a MoA and human relevance assessment and, possibly, to determine higher-tier testing needs.

Taken together, if the available thyroid-related data consistently indicate no concerns for endocrine activity via the T-modality, the EDC-T are not met, and the Thyroid-NDT-TAS is not entered. Vice versa, if the available thyroid-related data consistently indicate concerns for endocrine activity via the T-modality, the Thyroid-NDT-TAS is entered. In case of inconsistent findings, it should be the goal to explain these inconsistencies and to identify the need to generate further data when passing through the Thyroid-NDT-TAS. If the database is already complete (in accordance with the state-of-the-art) and the overall WoE evaluation is inconclusive (e.g. because findings are equivocal), an overall consideration of the entire hazard profile of the compound is recommended on a case-by-case basis before risk management is conducted.

2.2. Tier 1–3 of the Thyroid-NDT-TAS

Section 2.2.1 presents the scenario that the in vivo thyroid- and/or neurodevelopment-related database is insufficient at the onset of Tier 1 of the Thyroid-NDT-TAS and outlines the principal elements of Tier 1–3 for this scenario. By comparison, Section 2.2.2 presents the scenario that the in vivo thyroid- and neurodevelopment-related database is sufficient at the onset of Tier 1, i.e. that data from all potentially relevant OECD TGs are already available. These two scenarios were selected as illustrative examples to demonstrate how the Thyroid-NDT-TAS, that is not prescriptive, can be adapted to different information needs. In practice, the available database, just as the spectrum of thyroid- and neurodevelopment-related findings recorded at the onset of Tier 1 will vary on a case-by-case basis (Table Appendix 1). Expert judgement is required to determine how the Thyroid-NDT-TAS should be passed through, i.e. how it needs to be adapted to the given information needs.

2.2.1. Tier 1–3 of the Thyroid-NDT-TAS; example: in vivo thyroid- and/or neurodevelopment-related database insufficient at onset of Tier 1

Figure 3 presents the Thyroid-NDT-TAS illustrating the scenario that the in vivo thyroid- and/or neurodevelopment-related database is insufficient when entering Tier 1 of the Thyroid-NDT-TAS. To avoid, as far as possible, new animal testing, the Thyroid-NDT-TAS begins by conducting, in Tier 1, Step 1, a MoA and human relevance assessment and by completing in Tier 1, Step 2, the in vitro/in silico database, as relevant (Section 2.3). If the outcome of Tier 1 permits the conclusion that the EDC-T are either met or not met, the Thyroid-NDT-TAS is terminated. Only if the MoA is unclear and/or the human relevance of thyroid effects observed in the rat studies (and/or in studies using further species) cannot be excluded, the Thyroid-NDT-TAS is continued to Tier 2 to determine higher-tier testing needs to complete the in vivo thyroid- and/or neurodevelopment-related database. Further, the Tier 2 evaluations should consider, on a case-by-case basis, whether toxicokinetic data are relevant to establish, e.g. test substance concentrations and half-lives in the maternal, foetal and/or pup blood, placental/lactational substance transfer and/or the substance’s potential to pass the blood-brain barrier.

Figure 3. The ECETOC and CLE Thyroid-NDT-TAS. Scenario I: insufficient in vivo thyroid- and/or neurodevelopment-related data at onset of Tier 1.

ADME: absorption, distribution, metabolism, elimination; CTA: comparative thyroid assay; DNT: developmental neurotoxicity; EDC-T: endocrine disruptor criteria for thyroid modality; EOGRTS: extended one-generation reproductive toxicity study; H.R.: human relevance; MoA: mode-of-action; NDT: neurodevelopmental toxicity; OECD: Organisation for Economic Co-operation and Development; T3: triiodothyronine; T4: thyroxine; TG: test guideline; TK: toxicokinetics; TSH: thyroid stimulating hormone.

Colour legend: yellow shape: conclusion from Tier 0 evaluation to enter the Thyroid-NDT-TAS; dark blue boxes: MoA and human relevance assessment (Figure 5); light blue boxes: elements of the assessment; blue arrows: continuation of evaluation; ochreous box and arrows: optional elements of the assessment (as the respective parameters have not yet been formally validated or adopted for regulatory use); grey shading: elements of the higher-tier testing; red-brown vs green arrows and text: findings leading to conclusion that the EDC-T are met (red-brown circle)/are not met (green circle).

[a] Consider offspring serum T4, T3, and TSH thresholds observed by Marty et al. (2022) to support the determination of the biological relevance of findings (Section 2.1.2.3).

[b] Following expert judgement, further serum thyroid hormone data may not be necessary. Measurements of maternal and offspring plasma concentrations of the test material may be used to calculate placental transfer.

[c] See Section 2.2.3 for details on neurodevelopmental assessments.

[d] Following expert judgement, consider additional investigations using culled pups from the EOGRTS (or DNT study) and/or the performance of in vitro mechanistic assays and/or (not TG-conforming) perinatal studies to measure, e.g. brain thyroid hormones and/or receptor occupancies using immunohistochemistry. As relevant, consider measuring brain thyroid hormone already during the performance of the EOGRTS (or DNT study). Measurements of maternal and offspring plasma concentrations of the test material may be used to calculate placental transfer.

As the first part of Tier 2, the in vivo thyroid-related database is completed. The measurement of serum levels of T4 in the offspring is considered pivotal for this purpose, possibly supplemented by offspring serum TSH and T3 data, as offspring serum hormone levels, at the end of gestation or during lactation, are more closely associated with neurodevelopmental outcomes than the corresponding maternal serum levels (Marty et al. 2022). In vivo studies to measure offspring serum thyroid hormone levels include the comparative thyroid assay (US EPA 2005) and the OECD TG 421/422 developmental toxicity tests if enhanced to include serum thyroid hormone measurements at suitable timepoints. The ECETOC T4 TF and CLE recommend considering such shorter-term studies to enhance the understanding of thyroid function before considering the EOGRTS (OECD TG 443) or the enhanced two-generation reproductive toxicity study (OECD TG 416) supplemented with thyroid and offspring brain parameters, as Level 5 assays (Section 2.1.2.1). Alternatively, a DNT study (OECD TG 426), enhanced with measurements of thyroid-related parameters, or a male/female pubertal assay, as Level 4 assays, may be considered useful. Generally, expert judgement is required to determine which test methods are best suited for the substance of interest while adhering to the 3Rs principle (EP and Council 2010).

If the selected studies do not show any effects on offspring serum thyroid hormone levels, endocrine activity for the T-modality (Point 1(b) of the EDC; Box 1) is not present in the offspring, and it is concluded that the EDC-T are not met. If, however, offspring serum thyroid hormones are statistically significantly altered, endocrine activity for the T-modality is likely present, and the next part of Tier 2 aims to complete the neurodevelopment-related database to determine whether the substance also causes an adverse effect (Point 1(a) of the EDC). Formally adopted TGs to assess neurodevelopmental effects include the Level 4 DNT study (OECD TG 426), and the Level 5 EOGRTS (OECD TG 443), which is the only OECD TG that mandates measurements of offspring serum thyroid hormone, thyroid histopathology and neurodevelopmental effects in one study. Following the parameters addressed in OECD TG 426 or 443, adverse neurodevelopmental effects that can possibly be observed in rats include, e.g. alterations in motor activity, acoustic startle response, learning and memory and/or brain morphology/histopathology (Section 2.2.3).

If the selected assessments provide no evidence for adverse neurodevelopmental effects, it is concluded that the EDC-T are not met since there is no adverse outcome. Of note, the Thyroid-NDT-TAS, as it is described in this article, only considers NDT as the adverse outcome of thyroid-related MoAs (Section 1.1). Further developmental manifestations of substance-mediated thyroid hormone imbalance (e.g. altered offspring growth, developmental delays, disrupted thermoregulation, increased mortality and early eye opening) are not further considered here but were discussed in Section 4.5 of Marty et al. (2022).

If, however, there is evidence for adverse neurodevelopmental effects, Tier 3 of the Thyroid-NDT-TAS addresses the final MoA and human relevance assessment (Section 2.4) to establish whether the adverse effect (Point 1(a) of the EDC) and the endocrine activity (Point 1(b) of the EDC) are linked by an endocrine MoA (Point 1(c) of the EDC).

2.2.2. Tier 1–3 of the Thyroid-NDT-TAS; example: in vivo thyroid- and neurodevelopment-related database sufficient at onset of Tier 1

Figure 4 presents the Thyroid-NDT-TAS illustrating the scenario that the in vivo thyroid- and neurodevelopment-related databases are sufficient at the onset of the Thyroid-NDT-TAS (and that in vivo thyroid effects were recorded in Tier 0). If this scenario applies, the Thyroid-NDT-TAS is generally passed through in the same manner as described in Section 2.2.1 for the scenario “in vivo database insufficient.” However, since neurodevelopmental data are already available, these are assessed first (Section 2.2.3), and Tier 1 is only entered if adverse neurodevelopmental effects were observed or could not be excluded. If the Tier 1 MoA and human relevance assessment is inconclusive, the supportive higher-tier testing will generally be restricted to the measurement of non-guideline parameters since all potentially relevant TG-conforming data are already available.

Figure 4. The ECETOC and CLE Thyroid-NDT-TAS: Scenario II: sufficient in vivo thyroid- and neurodevelopment-related data at onset of Tier 1.

DNT: developmental neurotoxicity; EDC-T: endocrine disruptor criteria for thyroid modality; EOGRTS: extended one-generation reproductive toxicity study; H.R.: human relevance; MoA: mode-of-action; NDT: neurodevelopmental toxicity; T3: triiodothyronine; T4: thyroxine; TG: test guideline.

Colour legend: yellow shape: conclusion from Tier 0 evaluation to enter the Thyroid-NDT-TAS; light blue boxes: elements of the assessment; blue arrows: continuation of evaluation; ochreous box and arrows: optional elements of the assessment (as the respective parameters have not yet been formally validated or adopted for regulatory use); dotted ochreous arrow: expert judgement that offspring brain thyroid hormones and/or further neurodevelopmental parameters are relevant to substantiate or rule out NDT; dark blue box: MoA and human relevance assessment (Figure 5); red-brown vs green arrows and text: findings leading to conclusion that the EDC-T are met (red-brown circle)/are not met (green circle).

[a] See Section 2.2.3 for details on neurodevelopmental assessments.

[b] Following expert judgement, consider additional investigations using culled pups from the EOGRTS and/or the performance of in vitro mechanistic assays and/or (not TG-conforming) perinatal studies to measure e.g. brain thyroid hormones and/or receptor occupancies using immunohistochemistry. Measurements of maternal and offspring plasma concentrations of the test material may be used to calculate placental transfer.

2.2.3. Neurodevelopmental assessments within the thyroid-NDT-TAS

The ECETOC T4 TF and CLE recommend performing all neurodevelopmental assessments within the Thyroid-NDT-TAS in line with the provisions from the OECD TG 426 DNT study (the OECD TG 443 EOGRTS refers to the OECD TG 426 in this regard). The OECD TG 426 (version of 2007) includes substance administration during gestation and lactation and observations of the offspring up until adulthood, and it requests addressing these landmarks and endpoints (see Table 1 in OECD TG 426):

• Physical and developmental landmarks:

  • Body weight and clinical observations: weekly pre-weaning; thereafter, at least every two weeks

  • Brain weight and neuropathology: on postnatal day (PND) 22 and at study termination

  • Sexual maturation and other developmental landmarks, such as eye opening: as appropriate

• Functional/behavioural endpoints:

  • Behavioural ontogeny: at least two measures pre-weaning

  • Motor activity including habituation: 1–3 times pre-weaning, once in the young adults (PND 60–70)

  • Motor and sensory function as well as learning and memory: once each in the adolescent offspring (PND 25 ± 2) and young adults (PND 60–70)

The ECETOC T4 TF and CLE consider all these landmarks and endpoints generally suitable to assess NDT. However, effects observed in the offspring should be put into context with, e.g. available data for adult animals across studies to determine whether such effects are due to systemic toxicity (including acute neurotoxicity) and not signs of developmental toxicity. Further, for substances with a known direct neurotoxic MoA, the Thyroid-NDT-TAS should be only entered if there is evidence that maternal hypothyroxinaemia has contributed to the neurotoxic effects in the offspring. For example, if thyroid function is only affected at higher dose levels than offspring neurotoxicity, the neurodevelopmental effect is unlikely to be caused by the thyroid perturbation (see Section 2.3.1.2 for evolved Bradford Hill consideration dose-response concordance).

Opportunities to test learning and memory (i.e. cognitive function) in rodents are limited. In the comprehensive review of 51 rat studies that included in utero/lactational exposure to substances causing thyroid hormone imbalance, Marty et al. (2022) found that:

…motor activity, acoustic startle response and periventricular heterotopia were sensitive parameters even though the nature of the finding was not necessarily consistent across the substances (e.g. decreased motor activity vs. increased motor activity vs altered habituation). Parameters related to cognitive function were never the most sensitive parameters among the neurodevelopmental parameters examined in the respective studies.

This outcome is consistent with an analysis of the results from OECD TG 426 DNT studies investigating a total of 69 pesticides, which showed that the standard learning and memory tests were less sensitive than measures of motor activity and acoustic startle habituation (Raffaele et al. 2010). If testing for learning and memory is to be performed, expert judgement is required to select the most appropriate test that the given laboratory is equipped to run.

Heterotopias (mentioned in the above quote) are clusters of ectopic neurons in the brain indicative of altered neuronal migration (Goodman and Gilbert 2007; Gilbert et al. 2014; O’Shaughnessy et al. 2018, 2019). The potential of thyroid-active substances to elicit periventricular heterotopia in rats has been a matter of extensive research work that is being led and coordinated by the EPA (see, e.g. Goodman and Gilbert 2007; Gilbert et al. 2014; O’Shaughnessy et al. 2018, 2019). While the assessment of periventricular heterotopia has not yet been standardised for regulatory use, brain histopathology performed around PND 16, or later up until adulthood in accordance with the respective TG, will inform on the presence of periventricular heterotopia provided that the histopathological sections include the respective brain location (Marty et al. 2022). To the best of the knowledge of the ECETOC T4 TF and CLE, the development of hypothyroid-mediated heterotopia has not yet been described in humans. Whereas cortical heterotopia has been associated with neuronal disorders and certain genetic mutations in afflicted patients, it is “unknown what, if any, environmental influences are also linked to this malformation” (O’Shaughnessy et al. 2018). Further, a causal relationship between heterotopias and serum thyroid hormone decrements has not been established, and control rat brains may also have heterotopias. It is also not clear whether neurodevelopmental disorders can be present without brain heterotopia. For all these reasons, the ECETOC T4 TF and CLE suggest caution in the interpretation of hypothyroid-mediated heterotopia in rats.

Generally, neuropathological evaluations (gross inspections, morphometry and histopathology of the brain) are important elements of neurodevelopmental assessments to identify substance-mediated structural abnormalities; however, the evaluation of neuropathological findings requires expert judgement (further discussed in Section 4.3.4 of Marty et al. 2022). Details on the interpretation of neuropathological evaluations are provided in Section 44 of the OECD TG 426 and references therein. The ECETOC T4 TF and CLE are aware of evidence relating the following neurodevelopmental insults (listed in Section 44 of the OECD TG 426) to maternal or offspring thyroid perturbation in rats. Hence, these histopathological findings appear relevant for the evaluation of NDT:

• Alterations in the relative size of various brain regions (e.g. external germinal layer of cerebellum, corpus callosum): rat studies by Li J et al. (1986) and Shibutani et al. (2009), see also reviews by Chen and Hetzel (2010) and Valdés Hernández et al. (2013); note that Shibutani et al. (2009) further recorded reduced numbers of oligodendroglial cells (reflecting impaired oligodendroglial development) in the cerebral deep cortex of rats exposed to propyl thiouracil or methimazole.

• Alterations in neuronal migration and differentiation: rat studies by, e.g. Goodman and Gilbert (2007), Gilbert et al. (2014) and O’Shaughnessy et al. (2018, 2019).

• Alterations in patterns of myelination: rat studies by Salas-Lucia et al. (2020).

If assessments addressing the standard neurodevelopmental landmarks and endpoints that are included in the OECD TG 426 (or OECD TG 443) do not provide conclusive evidence on the presence or absence of adverse neurodevelopmental effects, the ECETOC T4 TF and CLE recommend considering additional neurodevelopmental assessments. For example, it might be considered to measure brain T4 and T3 in the culled pups from the DNT study or EOGRTS. It is important to note, however, that brain T4/T3 measurements have not yet been formally standardised and are currently not included in any formal TG. In the review by Marty et al. (2022), extensive data gaps compromised the establishment of associations between offspring brain T4/T3 decrements and the occurrence or absence of specific neurodevelopmental effects. Marty et al. (2022) concluded:

Brain T4 and T3 levels may well be the most relevant thyroid-related parameters to predict whether neurodevelopmental impairment will occur upon in utero/lactational exposure to thyroid-active substances. It is recommended to include brain T4/T3 measurements in rat developmental toxicity studies [i.e. studies that include a developmental component] evaluating thyroid-active substances [especially in prenatal/preweaning offspring]. Thereby, opportunities to standardise brain T4/T3 measurements can be identified, and the understanding of how altered brain T4/T3 levels are linked to neurodevelopmental impairment will be enhanced.

The measurement of offspring brain T4/T3 levels is a matter of active research (O’Shaughnessy and Gilbert 2020; Ford et al. 2023; see also research project EMSG59 funded by the European Chemical Industry Council Long-Range Research Initiative (Cefic LRI) Developing a quantitative AOP for liver-mediated thyroid modulation after prenatal exposure to a xenobiotic compound in the rat; http://cefic-lri.org/projects/emsg59-developing-a-quantitative-aop-for-liver-mediated-thyroid-modulation-after-prenatal-exposure-to-a-xenobiotic-compound-in-the-rat/ [accessed 2023 May]). As measurements of brain T4/T3 levels become more established (and a historical control database is developed), this may enhance the understanding of how these parameters may contribute to regulatory assessments of NDT (see Table Appendix 2, which is included in this article after the bibliography, for relevant research needs). Further potentially relevant new approaches, that may complement serum T4 measures when evaluating NDT (but that have also not yet been standardised and validated for formal use), include immunohistochemical and/or gene expression assays to identify biomarkers for adverse brain-related effects (O’Shaughnessy and Gilbert 2020).

Taken together, expert judgement is required (1) to select the spectrum of standard neurodevelopmental parameters that appear relevant for the substance of interest; (2) to determine the feasibility and usefulness of non-standardised neurodevelopmental measurements and investigations; and (3) to conduct a WoE evaluation of all relevant data to determine whether, or not, NDT is present. Generally, the ECETOC T4 TF and CLE suggest that there is “sufficient” evidence to conclude on the absence of NDT if the spectrum of neurodevelopmental landmarks and endpoints listed in the OECD TG 426 DNT study or OECD TG 443 EOGRTS have been covered and the WoE evaluation provides no indication for statistically significant and biologically relevant adverse neurodevelopmental effects. Further, the ECETOC T4 TF and CLE recommend generally assessing NDT observed in rats as relevant for humans, unless it can be shown, e.g. that the underlying MoA is not relevant for humans (Section 2.3). Human observational studies addressing neurodevelopmental impairment secondary to maternal thyroid hormone imbalance include a broad spectrum of parameters including psychomotor and mental development, cognitive function (intelligence quotient), expressive vocabulary or educational attainment, and, in single studies, clinical diagnoses of autism or attention deficit hyperactivity disorder or brain morphology assessed by magnetic resonance imaging (see comprehensive review by Sauer et al. 2020). Of note, the outcomes of some of the human observational studies considered by Sauer and colleagues may have been influenced by non-consideration of confounding lifestyle and/or medical parameters when selecting mothers for the respective cohorts. While the neurodevelopmental assessments that can be included in rat studies are per se relatively insensitive (by addressing rather crude endpoints, including limited animal numbers per group and yielding high data variability), they are invariably conducted at high dose levels relative to possible human exposure.

2.3. Tier 1: initial MoA and human relevance assessment and generation and evaluation of supporting in vitro/in silico evidence

Figure 5 presents the decision-tree for the initial and final MoA and human relevance assessments that are embedded in Tier 1 and Tier 3 of the Thyroid-NDT-TAS, respectively. Tier 1 consists of two steps:

Figure 5. Decision-tree for MoA and human relevance assessment embedded in Tier 1 and Tier 3 of the ECETOC and CLE Thyroid-NDT-TAS.

DIO: deiodinase; EDC-T: endocrine disruptor criteria for thyroid modality; H.R.: human relevance; LEI: liver enzyme induction; MoA: mode-of-action; NIS: sodium – iodide symporter; SBP: serum binding protein; T3: triiodothyronine; T4: thyroxine; TH: thyroid hormone; TPO: thyroid peroxidase; TR: thyroid receptor (nuclear); TSH: thyroid stimulating hormone; UGT: uridine diphosphate glucuronyltransferase.

Colour legend: dark blue boxes: Step 1 and Step 2 of the MoA and human relevance assessment; light blue boxes: elements of the assessment; blue arrows: continuation of evaluation; dotted blue arrow: expert judgement that Step 2 of Tier 1 should be skipped to directly continue to Tier 2 to generate higher-tier data; light green boxes: optional elements of the assessment as the respective parameters have not yet been formally adopted for regulatory use; ochreous box: optional assessment as the corresponding MIEs seem to be less frequent; red-brown vs green arrows and text: findings leading to conclusion that the EDC-T are met (red-brown circle)/are not met (green circle). Yellow shape: continuation of Thyroid-NDT-TAS.

[a] TSH likely not increased, and no thyroid organ changes, if only in utero/developmental exposure to substances enhancing thyroid hormone clearance (Marty et al. 2022).

[b] See Section 2.3.1 for elements to consider during the WoE evaluation.

[c] Apply expert judgement to determine which type of supporting in vitro and/or in silico evidence may be relevant for the substance of interest.

[d] Primarily TPO and NIS inhibition need to be excluded, as thyroid-related parameters are similarly affected by substances acting via a direct thyroid-related MoA. Also, depending on the thyroid hormone effect pattern (e.g. increased serum T4), substance interaction with DIOs needs to be excluded.

[e] The final MoA and human relevance assessment shall serve to answer the questions: Is the adverse effect not a consequence of thyroid MoA? If the substance has a thyroid MoA, is it not relevant for humans?

• Step 1: The initial MoA and human relevance assessment (Section 2.3.2). At this step, focus is on the identification of indirect thyroid-related MoA(s). Thereby, best possible use is made of the available in vivo database, which will generally include liver-related data. The outcome of Step 1 of Tier 1 may indicate:

  • That the EDC-T are met since a human-relevant thyroid-related MoA is likely, or that the EDC-T are not met since there is conclusive evidence that the substance has an indirect thyroid-related MoA that is not relevant for humans (Section 2.3.1.3 and Section 2.3.3). In these cases, the Thyroid-NDT-TAS is terminated after Tier 1, Step 1.

  • That the available evidence is inconclusive. In this case, expert judgement is required to determine, which specific additional data are needed to determine the substance’s MoA and to establish whether the MoA in rats is (not) relevant for humans.

    • If the expert judgement indicates that the need for in vitro and/or in silico data is predominant, Step 2 of Tier 1 is entered.

    • If the expert judgement indicates that the need for higher-tier data is predominant, Step 2 of Tier 1 is skipped and Tier 2 is entered.

• Step 2: The generation and evaluation of supportive in vitro and/or in silico evidence, as relevant (Section 2.3.4). Step 2 aims to complete the in vitro/in silico database by generating additional information that appears relevant to identify both indirect and direct thyroid-related MoAs. Just as the outcome of Step 1, the outcome of Step 2 may indicate:

  • That the EDC-T are either met or not met. In these cases, the Thyroid-NDT-TAS is terminated, and higher-tier testing is not necessary.

  • That the available evidence is still inconclusive. In this case, expert judgement is required to determine which higher-tier data may be relevant (Section 2.3.5).

If Tier 2 (higher-tier testing) is entered as an outcome of either Step 1 or Step 2 of Tier 1, the relevant in vivo data are identified, generated and evaluated to then proceed to Tier 3, which includes the final MoA and human relevance assessment and the final WoE evaluation (Section 2.4).

As a starting point for the detailed description of the MoA and human relevance assessments, important underlying elements are presented below, i.e. (Section 2.3.1.1) the adverse outcome pathway (AOP) concept; (Section 2.3.1.2) the evolved Bradford Hill considerations; and (Section 2.3.1.3) scientific evidence on the non-human relevance of indirect thyroid-related MoAs observed in rats.

2.3.1. Background information to MoA and human relevance assessment

2.3.1.1. Adverse outcome pathway (AOP) concept

The ECETOC T4 TF and CLE recommend applying information on AOPs to organise the MoA and human relevance assessments as the AOP concept has proven useful to address the biological plausibility of substance-specific MoAs and to enhance an understanding of toxicological effects (Marty et al. 2021). AOPs are linear sequences of events beginning with an MIE that may lead to early cellular events, followed by events in organs and organ systems and, ultimately, an observable adverse outcome in the organism, or population in case of ecologically relevant AOPs (Ankley et al. 2010; OECD 2017; Vinken et al. 2017). As compared to MoAs, AOPs are not substance-specific and thus do not include exposure or metabolism considerations (Marty et al. 2021). AOPs provide a structured approach to investigate the sequence of MIEs, key events and adverse outcomes that a substance may trigger. In biological reality, however, AOPs are hardly ever truly linear but embedded in e.g. converging events and feedback loops that aim at restoring balance (Knapen et al. 2018; Villeneuve et al. 2018).

In the central AOP repository “OECD AOP Wiki” (https://aopwiki.org [accessed 2023 May]), six AOPs relate to thyroid hormone imbalance and adverse neurodevelopmental outcomes in mammals:

• AOP 8: Upregulation of thyroid hormone catabolism via activation of hepatic nuclear receptors (leading to UGT induction), and subsequent adverse neurodevelopmental outcomes in mammals

• AOP 42: Inhibition of TPO and subsequent adverse neurodevelopmental outcomes in mammals (Crofton et al. 2019)

• AOP 54: Inhibition of NIS leading to learning and memory impairment (Rolaki et al. 2019)

• AOP 134: NIS inhibition and subsequent adverse neurodevelopmental outcomes in mammals

• AOP 152: Interference with thyroid serum binding protein transthyretin and subsequent adverse human neurodevelopmental toxicity

• AOP 300: Thyroid receptor antagonism and subsequent neurodevelopmental adverse outcomes in mammals

Hence, AOP 8 and 152 are relevant for the evaluation of indirect thyroid-related MoAs (liver enzyme induction and interaction with serum binding proteins); AOPs 42, 54 and 134 are relevant for the further evaluation of direct thyroid-related MoAs (TPO and NIS inhibition); and AOP 300 reflects events in the brain cells. Table Appendix 3, which is included in this article after the bibliography, presents the MIEs, key events and adverse outcomes of these AOPs (adapted from Table 1 in Marty et al. 2021). Table Appendix 3 also illustrates the (estimated) strength of evidence for key event relationships and their quantitative understanding. While AOP 42 and AOP 54 have been endorsed (Crofton et al. 2019; Rolaki et al. 2019), the other AOPs are marked in the OECD AOP Wiki as being “under development.” Often, the evidence for key event relationships and their quantitative understanding has been tentatively indicated as “moderate” or “weak,” especially for the later key events that lie closer to the adverse outcome (Table Appendix 3).

Further, Noyes et al. (2019) presented and discussed an “AOP network for chemically induced thyroid activity showing the integration of multiple individual AOPs under development and proposed” (Figure 2 in Noyes et al. 2019). This AOP network includes additional MIEs, such as:

• Activation of the hepatic nuclear receptors constitutive androstane receptor (CAR), aryl hydrocarbon receptor (AhR) and peroxisome proliferator-activated receptor (PPAR) – in addition to pregnane X receptor (PXR) activation, the MIE for AOP 8

• Substance interaction with the serum binding proteins albumin and thyroid binding globulin – in addition to interaction with transthyretin, the MIE for AOP 152

• DIO1, DIO2, DIO3 inhibition

Despite its recognised scientific limitations, the AOP concept provides a useful, structured framework for MoA and human relevance assessments. Therefore, the ECETOC T4 TF and CLE recommend that all MoA and human relevance assessments within the Thyroid-NDT-TAS are structured following the sequence of MIE(s), key events and adverse outcome(s) of those AOP(s) that reflect the hypothesised MoA(s) (Section 2.3.3). Such an approach also stands in line with the EFSA (2019b) Administrative guidance on submission of dossiers and assessment reports for the peer-review of pesticide active substances and the EFSA (2020) Technical report on the outcome of the pesticides peer review meeting on general recurring issues in mammalian toxicology, which request that the relevant data should be transparently tabulated to evaluate the dose- and temporal-response.

2.3.1.2 Evolved Bradford Hill considerations

In addition to consideration of the AOP concept, the ECETOC T4 TF and CLE recommend conducting all MoA and human relevance assessments following pre-defined considerations. For example, the WHO/IPCS MoA and Species Concordance Analysis Framework includes five “evolved Bradford Hill considerations” (Meek, Boobis, et al. 2014; Meek, Palermo, et al. 2014; see also Table 2 in Meek, Palermo, et al. 2014):

• Consistency: “Is the pattern of effects across species/strains/organs/test systems what would be expected?”

• Essentiality of key events: “Is the sequence of events reversible if dosing is stopped or a key event prevented?”

• Temporal concordance: “Are the key events observed in hypothesised order?”

• Dose-response concordance: “Are the key events observed at doses below or similar to those associated with the end (adverse) effect?”

• Biological concordance: “Does the hypothesised MoA conflict with broader biological knowledge? How well established is the MoA in the wider biological database?”

The elements for WoE analysis presented in the ECHA templates for WoE and uncertainty evaluation in risk assessment (Section 2.1.1) are also based on these evolved Bradford Hill considerations. Similarly, the EFSA (2019b) Administrative guidance on submission of dossiers and assessment reports for the peer-review of pesticide active substances includes an Appendix I Template for presentation of the assessment of endocrine disrupting properties, which lists predefined criteria that should be considered during the MoA analysis. These predefined criteria are biological plausibility and empirical support for each key event relationship (MIE to key event 1, key event 1 to key event 2, etc. up until final key event to adverse outcome), essentiality of each key event as well as consistency, analogy and specificity of the MoA.

2.3.1.3 Indirect thyroid-related MoA, a rodent-specific mechanism

Evidence in the scientific literature indicates that thyroid perturbations caused by substance-mediated induction of Phase II liver enzymes (mainly: UGT) and/or interaction with serum binding proteins are rodent-specific MoAs that are generally not relevant for humans (see, e.g. Papineni et al. 2015; Plummer et al. 2021; Strupp et al. 2020; Bomann et al. 2021; Parmentier et al. 2022; Tinwell and Bars 2022). In a comprehensive review of the underlying mechanistic evidence, Marty et al. (2021) showed that rats and humans differ considerably in the major thyroid hormone metabolism pathways. Generally, UGT-mediated glucuronidation is the major metabolic pathway in rats but only a minor pathway in humans whose most prominent route for thyroid hormone metabolism is by deiodination (Cavalieri and Pitt-Rivers 1981) and to a lesser extent by sulphotransferases (Richardson et al. 2014). See Table 3 for in vitro assays to evaluate species differences in the activity of these enzymes. Further research may elucidate other pathways that impact thyroid hormone metabolism.

Rats and humans also differ considerably in thyroid hormone distribution across the three major serum binding proteins (thyroid binding globulin, transthyretin and albumin) and in the binding affinities of these proteins (Marty et al. 2021). Also, none of the publications reviewed by Sauer et al. (2020) provided relevant information to establish a link between substance-mediated induction of those liver enzymes that are relevant for thyroid hormone metabolism in humans and thus increased serum thyroid hormone clearance, let alone to establish a further link between maternal hypothyroxinaemia and child neurodevelopmental impairment.

The EFSA (2020) Technical Report on the outcome of the pesticides peer review meeting on recurring issues on mammalian toxicology also notes that liver enzyme induction-mediated thyroid effects may not be relevant for humans (EFSA 2020, p. 8):

…endocrine mediated adverse effects that are secondary to other toxicities (including liver toxicity) should not be considered for concluding that EDC are met. In this case, it is necessary to demonstrate by means of comparative MoA analysis that thyroid toxicity is secondary to e.g. liver toxicity. In the comparative MoA analysis, a MoA for thyroid toxicity and one for liver toxicity should be postulated in a comparative manner. The applicant should transparently tabulate the data in order to evaluate the dose- and temporal-response…

The assessment of human relevance is mainly applicable to those cases where the T-mediated effect is through a liver-mediated mechanism i.e. liver enzyme induction resulting in an increase of THs [thyroid hormones] clearance. In this case, three pieces of information should be provided to evaluate whether the thyroid findings are likely or not to be human relevant: 1) analysis of T3, T4 and TSH in the repeated dose studies; 2) in vitro comparative studies to evaluate liver enzyme induction in the tested species (i.e. rat, mouse and dog) and humans; 3) evaluation of other potential in vitro mechanisms involved in the thyroid disruption. Finally, all the available evidences should be weighed, including interspecies differences and lack of any concomitant molecular initiating event.

This EFSA Technical Report also states that “…EFSA confirmed that a CAR/PXR-mediated MoA that can also be expected to be functional in humans, leading to an increased clearance of THs would be considered relevant” (EFSA 2020, p. 8). This statement from EFSA (2020) underlines that the non-human relevance of MoAs needs to be shown on a case-by-case basis.

In line with the available scientific evidence, the human relevance assessments that are embedded in the Thyroid-NDT-TAS focus on indirect thyroid-related MoAs that are triggered by UGT induction and/or substance interaction with serum binding proteins.

2.3.2. Tier 1, Step 1: relevant data to establish indirect thyroid-related MoA

To make best possible use of the database that is available at the onset of the Thyroid-NDT-TAS, the Tier 1, Step 1, initial MoA and human relevance assessment (Figure 5) includes the evaluation of:

1. All available in vivo data on liver effects, including in vivo UGT induction data, if available

2. All available in vitro and/or in silico data on UGT induction and/or on substance interaction with serum binding proteins

3. The observed pattern of in vivo thyroid effects

Accordingly, Step 1 addresses the question whether the available evidence indicates that the substance of interest enhances thyroid hormone clearance via liver enzyme (UGT) induction and/or interaction with serum binding proteins and thus is likely to have an indirect thyroid-related MoA. As per Marty et al. (2022), some compounds trigger both MoAs, i.e. they induce liver enzymes and interact with serum binding proteins.

The available in vivo data on liver effects are jointly evaluated to determine whether adverse or adaptive liver effects are present, thereby indicating a concern for liver enzyme induction even if data on UGT induction are unavailable. Useful parameters for this evaluation include absolute and relative liver weight, gross necropsy and histological assessments of the liver as well as clinical pathology addressing the activities of, e.g. alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase and γ-glutamyltransferase (see, e.g. Hall et al. 2012). Molecular methods, such as receptor binding assays and newer toxicogenomics technologies, may also provide supportive evidence for liver enzyme induction-mediated liver toxicity (Hall et al. 2012). Further, Yamada et al. (2013) reported that substances with direct vs indirect thyroid-related MoAs could be discerned by the ratio of relative thyroid weight increase vs relative liver weight increase (≥1.7 but generally ≥3.2 for substances with direct thyroid-related MoA and ≤1.2 for substances with indirect thyroid-related MoA; Yamada et al. 2013).

The available in vitro and/or in silico data on UGT induction and/or substance interaction with serum binding proteins are used as supporting evidence to determine whether an indirect thyroid-related MoA is likely. Further, if in vitro data from both rat and human test systems (cells, tissues, etc.) are already available during Step 1 of Tier 1, a comparative assessment of the findings can help identify species differences in effects.

The pattern of thyroid effects is considered to determine whether the liver enzyme induction coincides with serum T4 decrements and, possibly, increased TSH and follicular cell hypertrophy/hyperplasia of the thyroid gland. After at least 28-day exposure of adult rats, liver enzyme inducers generally not only cause serum T4 decrements, but these decrements also trigger a response of the hypothalamus-pituitary-thyroid axis, which is recognisable by serum TSH increases and altered thyroid weight/histopathology (see, e.g. Smith et al. 1991; Strupp et al. 2020; Parmentier et al. 2022).

However, for perfluorooctane sulphonates, Chang et al. (2007, 2008) hypothesised that these liver enzyme inducers do not trigger TSH increases since they also compete with T4 for serum binding proteins in rats. The resulting increase in free T4 would increase thyroid hormone availability to peripheral tissues for utilisation, metabolic conversion and excretion without direct interference with regulatory functions of the rat hypothalamus-pituitary-thyroid axis. Thereby, the rats would maintain an euthyroid state (Chang et al. 2007, 2008). This hypothesis further supports the relevance of thyroid histopathology as parameter for sustained and physiologically relevant thyroid hormone disruption in rats. Also, this hypothesis may help explain why Marty et al. (2022) observed offspring (and maternal) serum T4 decrements, but no TSH increases, upon gestational and lactational exposure of rats to substances that induce enhanced thyroid hormone metabolism and clearance and/or that affect T4 binding to serum carrier proteins [Marty et al. 2022, citing studies assessing perfluorohexane sulphonates (Ramhøj et al. 2018, 2020; Gilbert et al. 2021), the polychlorinated biphenyl Aroclor 1254 (Morse et al. 1996; Goldey et al. 1995) or DE-71, a mixture of polybrominated diphenyl ethers (Kodavanti et al. 2010; Bowers et al. 2015)]. In these studies, either no hypothyroid-mediated effects on neurodevelopment were observed, or such effects were less clear (e.g. absence of heterotopia) than observed for the substances with a direct thyroid MoA (Marty et al. 2022).

2.3.3. Outcome of Tier 1, Step 1: does substance have indirect thyroid-related MoA? If so, is it non-relevant to humans?

All data collected for the Tier 1, Step 1, initial MoA and human relevance assessment are submitted to a WoE evaluation to determine whether the available evidence is sufficient to conclude that the substance of interest has (or does not have) an indirect thyroid-related MoA. If liver effects are observed at lower doses and earlier timepoints than the thyroid effects, enhanced thyroid hormone clearance via liver enzyme induction is likely to be a relevant MoA (as the modified Bradford Hill considerations temporal concordance and dose-response concordance are fulfilled; Section 2.3.1.2). The sensitivity of the investigated liver parameters vs. thyroid parameters should always be considered before a final assessment of the temporal concordance and dose-response concordance of the events can be made.

If the outcome of Tier 1, Step 1, indicates that the substance likely has an indirect thyroid-related MoA, it is assessed whether this MoA is (not) relevant for humans. As noted in Section 2.3.1.1, the ECETOC T4 TF and CLE recommend that this MoA and human relevance assessment is structured following the sequence of events of the AOP(s) that reflect the hypothesised MoA(s). For the assessment of an indirect thyroid-related MoA, Figure 6 illustrates the sequences of events that lead from (1) hepatic nuclear receptor activation and liver enzyme induction or (2) substance interaction with serum binding proteins to adverse neurodevelopmental outcomes, i.e. the sequences of events that reflect AOP 8 and AOP 152, respectively (Section 2.3.1.1). Figure 6 also shows which in vivo or in vitro data may be useful to determine whether, and at which dose level, the respective event was triggered by the substance of interest. When applying test systems using cells or tissues from both rat and human origin or when conducting PBK modelling to compare estimated thyroid hormone levels in rat vs human (Leonard et al. 2016), the in vitro assessments/PBK models can also be used to determine species concordances or differences of effects (see Table 3; Column “Test system, from rat & human = facilitates H.R. assessment”).

Figure 6. Sequence of events that may lead from liver enzyme induction (AOP 8) and interaction with serum binding proteins (AOP 152) to adverse neurodevelopmental outcomes in mammals and opportunities for their further investigation.

AhR: aryl hydrocarbon receptor; AO(P): adverse outcome (pathway); BROD: benzoxyresorufin; CAR: constitutive androstane receptor; Cyp: cytochrome p-450; fT4: free thyroxine; H.R.: human relevance; KE: key event; LDG: lower-dose groups; MIE: molecular initiating event; NIS: sodium – iodide symporter; PBK: physiologically based kinetic; PPAR: peroxisome proliferator-activated receptor; PROD: pentoxyresorufin; PXR: pregnane X receptor; SBP: serum binding protein; T3: triiodothyronine; T4: thyroxine; TDG: top-dose group; TSH: thyroid stimulating hormone; TTR: transthyretin; UGT: uridine diphosphate glucuronyltransferase.

Colour legend: yellow arrow: available data indicating that AOP 8/AOP 152 may be relevant for the MoA and human relevance assessment. Boxes with blue/red-brown shading: MIE, early KEs and AOs that relate to AOP 8/AOP 152. White boxes with blue/red-brown text: Supportive in vivo or in vitro evidence that may inform on the MIE or specific KEs for AOP 8/AOP 152 (linked by blue/red-brown arrows; black arrow for enhanced traceability). Boxes with green shading: KEs relating to serum/brain T4 decrements; these KEs are central to five of the six potentially relevant AOPs included in the OECD AOP Wiki (exception: AOP 300 on thyroid receptor antagonism; Table Appendix 3). White box with green text: In vivo data or PBK modelling to inform on thyroid-related events. Boxes with grey shadings: KEs relating to the hippocampus; these KEs are central to five of the six potentially relevant AOPs (exception: AOP 54 on NIS inhibition leading to impaired learning and memory). Dotted black arrow: T4 measurements in relevant tissues, if available.

[a] In the OECD AOP Wiki, the MIE of AOP 8 is recorded as PXR activation. Noyes et al. (2019) indicated CAR, AhR and PPAR activation as further MIEs leading to liver enzyme induction. The ECETOC T4 TF and CLE contend that all these MIEs are not indispensable to trigger UGT upregulation. Therefore, assessments addressing hepatic nuclear receptor activation may not be needed for the MoA assessment of the substance of interest.

[b] While AOP 152 only refers to TTR, the available in vitro assays generally allow measuring all three major serum binding proteins, i.e. TTR, albumin and thyroid binding globulin. Since thyroid hormone distribution across these three serum binding proteins and their binding affinities differ considerably between rats and humans, the ECETOC T4 TF and CLE recommend considering in vitro substance interaction with all three serum binding proteins, as relevant.

[c] See Tinwell and Bars (2022) for details on the indirect assessment of CAR/PXR activation in rat studies via induction of transcript level and corresponding enzyme activity associated with each receptor (Cyp2b/PROD and Cyp3a/BROD for CAR and PXR, respectively).

[d] PBK modelling: Estimate serum/brain T4 levels in rat vs human considering relevant parameters, such as binding constants, potencies of MIEs and/or liver enzyme inductions in rat vs human tissue.

[e] The AOPs in the OECD AOP Wiki (Table Appendix 3) only generally refer to “T4 in serum, decrease” without distinction between maternal and offspring serum T4 levels; also, none of the AOPs considers serum T3 (or TSH). Following the observations by Marty et al. (2022), maternal serum T4 levels do not appear predictive of neurodevelopmental effects. However, there seems to be some association between ≥ 60%/≥ 50% offspring serum T4 decrements in the TDG/LDGs (and ≥ 20% and statistically significant offspring serum T3 decrements) and the occurrence of statistically significant neurodevelopmental effects. Therefore, the ECETOC T4 TF and CLE recommend considering offspring serum T4 as predominant parameter related to serum thyroid hormone levels. In addition, information on maternal serum T4, maternal and/or offspring serum T3 and offspring brain T4/T3 should be considered, if available (see Section 2.1.2.3 for further discussion).

[f] For AOP 152 (as well as AOP 42 and AOP 134), “cochlear function, decreased/loss” was indicated as adverse outcome in the OECD AOP Wiki as per 13 September 2019, whereas it was indicated as “cognitive function, decreased” as per 15 October 2019.

For each hypothesised MoA, the (rat and/or human) evidence indicating that the MIE and subsequent key events were triggered (and at which dose level) should be recorded. Tinwell and Bars (2022) and Parmentier et al. (2022) have shown how relevant information can be sorted by MIE and key event to support the MoA and human relevance assessment for substances that exhibit an indirect thyroid-related MoA in rats.

2.3.4. Tier 1, Step 2: generation of supportive in vitro/in silico evidence

Tier 1, Step 2, of the Thyroid-NDT-TAS serves to complete the in vitro/in silico database, as relevant for the substance of interest (see Section 2.1.3 for definition of “sufficient” in vitro database). Such supportive in vitro and/or in silico evidence may be relevant (1) to inform on the MoA(s) of the substance of interest, and (2) to establish whether thyroid-related MoAs observed in rats are (not) relevant for humans. Generally, expert judgement is required to select suitable in vitro assays, in silico models and/or PBK models (Table 3; Section 2.1.3) and to evaluate the corresponding findings.

If (additional) evidence is required to establish or rule out a direct thyroid-related MoA, in vitro/in silico data on TPO and NIS inhibition are helpful; however, the high false positive rate of the corresponding assays needs to be considered during the evaluation of any positive findings (Section 2.1.3). As relevant, orthogonal assays using different endpoint detection methods can be used to support preliminary evidence that the substance may trigger a specific MIE. For example, if an Amplex UltraRed TPO assay indicates that the substance may inhibit TPO, this concern may be followed up by assessing tyrosine iodination via liquid chromatography. Similarly, if a radioactive-iodide uptake assay indicates that the substance may inhibit the NIS, this concern may be followed up by assessing NIS activation based on the Sandell–Kolthoff reaction (see Table 3 for assay references).

If (additional) evidence is required to establish or rule out an indirect thyroid-related MoA (e.g. to follow up on adverse liver effects observed in the available rat studies), comparative liver enzyme induction assays and/or liquid chromatography (coupled with mass spectrometry) assays measuring inhibition of thyroid hormone glucuronidation and/or suitable PBK modelling may be relevant (see Table 3 for assay references).

2.3.5. Outcome of Tier 1, Step 2: does substance have direct or indirect thyroid-related MoA? If so, is it non-relevant to humans?

As an outcome of Tier 1, Step 2, a further WoE evaluation is conducted that generally follows the same outline as the WoE evaluation conducted as an outcome of Tier 1, Step 1, with the exception that the in vitro/in silico database is now complete so that the overall evaluation can address both indirect and direct thyroid-related MoAs. If the findings indicate that more than one (thyroid-related or non-thyroid-related) MoA may be triggered, the WoE evaluation should consider which link between thyroid activity and adverse outcome appears most plausible – and thus which MoA is likely predominant. For example, if the substance of interest causes in vitro NIS inhibition at very high concentrations, but thyroid effects are not observed at comparable in vivo dose levels, the modified Bradford Hill criterion of dose-response concordance is not met, and the MIE NIS inhibition is assessed as not biologically relevant. Similarly, in vitro effects that are recorded at concentrations ranging close to the cytotoxic concentrations are weighted lower than in vitro effects recorded at lower concentrations.

2.4. Tier 2 and Tier 3: higher-tier testing and final MoA and human relevance assessment and final conclusion regarding EDC-T

If Tier 1 of the Thyroid-NDT-TAS is inconclusive, Tier 2 of the Thyroid-NDT-TAS can be entered to generate and evaluate relevant higher-tier data. If the higher-tier data indicate that the substance of interest does not elicit both offspring serum T4 (T3 and TSH) changes and NDT, the final conclusion is drawn that the substance of interest does not meet the EDC-T (since it does not have endocrine activity in the offspring and/or does not cause an adverse effect). If, however, both thyroid perturbations in the offspring and adverse neurodevelopmental effects are recorded, Tier 3 of the Thyroid-NDT-TAS is entered, and the final MoA and human relevance assessment is performed. The overarching WoE evaluation of all available and new data is used to answer the questions whether a (direct or indirect) thyroid-related MoA is predominant and whether the MoA is (not) relevant in humans. If this overall evaluation indicates that the substance likely has a thyroid-related MoA as predominant MoA and human relevance of this MoA cannot be ruled out, the final conclusion is drawn that the EDC-T are met.

3. Conclusions and outlook

A comprehensive and structured approach to assess whether active substances in plant protection products, biocidal products and REACH substances meet the EDC-T is currently unavailable. To address this shortcoming, the ECETOC T4 TF and CLE have now proposed a Thyroid Function-Related Neurodevelopmental Toxicity Testing and Assessment Scheme (Thyroid-NDT-TAS). The Thyroid-NDT-TAS provides a structured, tiered approach to determine (1) whether substances elicit adverse neurodevelopmental effects and (2) have thyroid activity and (3) whether the two are linked by an indirect or direct thyroid-related MoA, and thus meet the EDC-T, unless it can be shown that the MoA is not relevant for humans. The Thyroid-NDT-TAS is based on the state-of-the-science, and it has been developed to comply with the European Commission (2017, 2018) EDC and the EFSA and ECHA (2018) Endocrine Disruptor Guidance. Further, the Thyroid-NDT-TAS takes into consideration knowledge gaps that have been identified in the earlier ECETOC T4 TF reviews (Sauer et al. 2020; Marty et al. 2021, 2022); see Table Appendix 2 for details on research needs. In all tiers of the Thyroid-NDT-TAS, prevailing uncertainties may be addressed by a WoE approach if multiple data are available or, if possible, by the generation of new data. If the database is complete (in accordance with the state-of-the-art) and the overall WoE evaluation is inconclusive (e.g. because findings are equivocal), the entire hazard profile of the compound should be considered on a case-by-case basis before risk management is conducted. To make human safety assessments more accurate while at the same time minimising animal testing, it is recommended that registrants and regulators apply the Thyroid-NDT-TAS for future regulatory assessments.

Abbreviations
ADME	=	absorption, distribution, metabolism, elimination
AE	=	adverse effect (Table Appendix 1)
AhR	=	aryl hydrocarbon receptor
AO(P)	=	adverse outcome (pathway)
BROD	=	benzoxyresorufin (Figure 6)
CAR	=	constitutive androstane receptor
Cefic LRI	=	European Chemical Industry Council Long-Range Research Initiative
CHO	=	Chinese hamster ovary (cells) (Table 3)
CLE	=	CropLife Europe
CTA	=	comparative thyroid assay (Figure 3)
Cyp	=	cytochrome p-450 (Figure 6)
DIO	=	deiodinase
DNT	=	developmental neurotoxicity
EA	=	endocrine activity (Table Appendix 1)
ECETOC	=	European Centre for Ecotoxicology and Toxicology of Chemicals
ECHA	=	European Chemicals Agency
EDC(-T)	=	Endocrine Disruptor Criteria (for the thyroid modality)
EFSA	=	European Food Safety Authority
EOGRTS	=	extended one-generation reproductive toxicity study
EP	=	European Parliament
EPA	=	Environmental Protection Agency
EU	=	European Union
EU NETVAL	=	EU Network of Laboratories for the Validation of Alternative Methods
EURL ECVAM	=	European Union Reference Laboratory for alternatives to animal testing
FRTL	=	Fischer rat thyroid follicular (cell line) (Table 3)
fT3	=	free triiodothyronine
fT4	=	free thyroxine
GABA	=	gamma amino-butyric acid (Table Appendix 3)
GD	=	gestational day (Table 1, Table 2)
H.R.	=	human relevance (Figures 1–6)
HP	=	histopathology (Figure 2)
HPT	=	hypothalamic–pituitary–thyroid (Table Appendix 2)
HTS	=	high-throughput screening
ILSI/HESI	=	International Life Sciences Institute / Health and Environmental Sciences Institute
KE(R)	=	key event (relationship) (Figure 6, Table Appendix 3)
LC	=	liquid chromatography (Table 3)
LD	=	lactational day (Table 1)
LDG	=	lower-dose groups (Figure 6)
LEI	=	liver enzyme induction (Figure 5)
MCT	=	monocarboxylate transporter (Table 3)
MDCK	=	Madin Darby Canine Kidney (Table 3)
MIE	=	molecular initiating event
MoA	=	mode-of-action
MS	=	Mass spectrometry (Table 3)
MTD	=	maximum tolerated dose (Figure 2)
NA	=	not addressed (Table 3)
NDT	=	neurodevelopmental toxicity
NIS	=	sodium – iodide symporter
OCSPP	=	Office of Chemical Safety and Pollution Prevention
OECD	=	Organisation for Economic Co-operation and Development
PBK	=	physiologically based kinetic
PFHxS	=	perfluorohexane sulphonates (Table Appendix 2)
PND	=	postnatal day
PPAR	=	peroxisome proliferator-activated receptor
PROD	=	pentoxyresorufin (Figure 6)
PXR	=	pregnane X receptor
QSAR	=	quantitative structure activity relationship (Table 3)
REACH	=	Registration, Evaluation, Authorisation and Restriction of Chemicals
SBP	=	serum binding protein (Figure 5, Figure 6)
SC	=	Scientific Committee
T-modality	=	thyroid modality for endocrine disruption
T3	=	triiodothyronine
T4	=	thyroxine
TBG	=	thyroid binding globulin (Table 3)
TDG	=	top-dose group (Figure 6)
TF	=	Task Force
TG	=	Test Guideline
TH	=	thyroid hormone
ThyM	=	thyroid method (Table 3)
Thyroid-NDT-TAS	=	Thyroid Function-Related Neurodevelopmental Toxicity Testing and Assessment Scheme
TK	=	toxicokinetics (Figure 3)
TPO	=	thyroid peroxidase
TR	=	thyroid receptor (nuclear) (Table 3, Figure 5)
TRH	=	thyrotropin-releasing hormone (Table 3)
TSAR	=	Tracking System for Alternative Methods towards Regulatory Acceptance
TSH	=	thyroid stimulating hormone
TTR	=	transthyretin (Table 3, Figure 6, Table Appendix 3)
UGT	=	uridine diphosphate glucuronyltransferase
Val	=	validation (Table 3)
WHO / IPCS	=	World Health Organisation / International Programme on Chemical Safety
WoE	=	weight-of-evidence

Acknowledgements

We would like to thank the members of the ECETOC Scientific Committee as well as the in-house reviewers (see Declaration of competing interests) for their critical comments. Also, we would like to thank the former ECETOC T4 TF chair Nina Hallmark (Bayer AG, Germany) as well as the former TF members Davy Guignard (Evotec France, France), Bethany R. Hannas (Eli Lilly & Co., USA) and Daniel Urbisch (Syngenta, Switzerland) for valuable discussions that contributed to the basis for this review. We are indebted to Andreea Cuciureanu, Andrea Salvadori, Francesca Ugucchioni and the entire ECETOC secretariat for organisational and technical assistance to the ECETOC T4 TF. We thank the reviewers selected by the Editor that were anonymous to us. The comprehensive and helpful comments served to improve the manuscript.

Declaration of interest

The co-authors of this manuscript consist of members of the ECETOC T4 TF and/or of CLE (SMK, KB, AC, BG, RG, SJ, SM, HAM, EFM, LPS, CS, HT, ChW), the scientific writer (UGS) and Stewards from the ECETOC Scientific Committee (PAB, BvR). The views expressed in this article are solely those of the co-authors and may not represent those of the sponsoring organisations. None of the co-authors received funding in cash or kind for their contribution to this manuscript, with the exception of UGS, a freelance scientific writer.

Stephanie Melching-Kollmuss (SMK), Heike-Antje Marxfeld (HAM) and Christiane Wiemann (ChW) are employed by BASF SE, Limburgerhof, Germany, BASF SE, Ludwigshafen, Germany, and BASF Oesterreich GmbH, Vienna, Austria, respectively. BASF produces a very wide range of chemicals including some of those mentioned in this paper and/or substances that may have to be tested for their potential to cause maternal thyroid disruption and subsequent developmental neurotoxicity. An in-house review yielded few requests for amendments, which served clarification. SMK’s responsibilities within BASF SE include being regulatory toxicologist for agrochemicals. SMK is the current Chair of the ECETOC T4 TF. HAM’s responsibilities within BASF include pathological evaluation and assessment of a broad variety of regulatory studies conducted by BASF. ChW’s responsibilities within BASF include being a regulatory toxicologist and human risk assessor for agrochemicals. Also, ChW is contributing to the CLE ED Thyroid Subgroup.

Kathrin Bothe (KB), Babunilayam Gangadharan (BG), Larry P. Sheets (LPS) and Helen Tinwell (HT) are employed by the Crop Science Division of Bayer AG (Germany/USA) or Bayer SAS (France). The Bayer portfolio includes substances that may have to be tested for their potential to cause maternal thyroid disruption and subsequent developmental neurotoxicity, including some of those mentioned in this paper. An in-house review of this manuscript yielded no requests for amendments. KB’s responsibilities include being a regulatory toxicologist at the Bayer Crop Science Division (Germany). BG’s responsibilities include pathological evaluation and assessment of a broad variety of regulatory studies conducted on Bayer compounds at the Bayer Crop Science Division (France). LPS’s responsibilities include being Senior Fellow Regulatory Toxicology at the Bayer Crop Science Division. HT’s responsibilities include being Regulatory Toxicology Team Leader at the Bayer Crop Science Division (France).

Alex Charlton (AC), Elizabeth F. McInnes (EFM) and Philip A. Botham (PAB) are employed by Syngenta, an international agribusiness that markets crop protection chemicals and seeds. The Syngenta portfolio includes substances that may have to be tested for their potential to cause maternal thyroid disruption and subsequent developmental neurotoxicity. An in-house review of this manuscript yielded no requests for amendments. AC’s responsibilities within Syngenta include providing scientific support to research and development activities and to regulatory toxicology projects. EFM’s responsibilities within Syngenta include providing peer review of all laboratory animal pathology data generated by Syngenta and providing expert advice on the possible adversity of all pathology findings to the global toxicology platform within Product Safety. PAB’s responsibilities within Syngenta are to provide strategic scientific advice on product safety issues to the company’s Product Safety, Business Sustainability and Crop Protection Development organisations.

Rashin Ghaffari (RG) is employed by Corteva Agriscience, USA. Corteva markets products (or previously marketed products) containing some of the chemicals included in this paper. Further, the Corteva portfolio includes substances that may have to be tested for their potential to cause maternal thyroid disruption and subsequent developmental neurotoxicity. An in-house review of this manuscript yielded no requests for amendments. RG’s responsibilities within Corteva include providing scientific support to developmental and reproductive toxicology testing strategy and to research and development projects.

Sylvia Jacobi (SJ) is a freelance consultant and was employed by Albemarle Europe SRL as Corporate Toxicology Director until July 2022. SJ still provides scientific advice to this company. Albemarle is a chemical company that markets or previously marketed products containing some of the chemicals included in this paper. Its portfolio also includes substances that may have to be tested for thyroid disruption and subsequent neurotoxicity.

Sue Marty (SM) is employed by Dow, Inc. The issues of hazard identification and risk assessment of thyroid-active compounds, and how these are assessed by regulatory/other agencies, impact substances of interest to the corporation. An in-house review of this manuscript yielded no requests for amendments. SM’s role at Dow is focused on Dow’s science strategy and testing, which includes the assessment of endocrine-active compounds.

Ursula G. Sauer (UGS), a freelance scientific writer, was hired by ECETOC to assist in the preparation of this manuscript.

Christian Strupp (CS) is employed by Gowan Crop Protection Ltd., Reading, United Kingdom. Gowan is a company manufacturing and marketing plant protection products. An in-house review of this manuscript yielded few requests for amendments, which served clarification. CS’s responsibilities include all aspects of human safety testing of Gowan’s portfolio. CS was also the chair-elect of CLE’s Human Health Expert Group (H2EG) at the time of manuscript compilation. CLE’s H2EG contributes to improving the process of plant protection product safety assessment by scientific contributions and constructive commenting to consultations on regulatory processes.

BvR has been an independent consultant (Environmental Sciences Consulting) since the beginning of 2022. BvR is an Associate Professor of Reproduction Toxicity of the University of Wageningen, Netherlands, and the Chairman of the ECETOC Scientific Committee.

This manuscript was reviewed by the ECETOC Scientific Committee consisting of representatives of academia, regulatory agencies and industry (https://www.ecetoc.org/about-ecetoc/structure/scientific-committee/). This review yielded few requests for amendment, which served clarification.

Finally, all authors suggest that the proposed Thyroid-NDT-TAS, which was developed for regulatory purposes, should be used in future regulatory actions – both within the EU and in other jurisdictions.

Additional information

Funding

This manuscript relates to work undertaken by members of the European Centre for Ecotoxicology and Toxicology of Chemicals (ECETOC) T4 TF and/or CropLife Europe (CLE). ECETOC (www.ecetoc.org) is a scientific organisation which provides a collaborative space for scientists from industry, academia and governments. Its mission is to develop and promote practical, trusted and sustainable solutions to scientific challenges which are valuable to industry, as well as to the regulatory community and society in general. ECETOC is financed by its membership, which are the leading companies with interests in the manufacture and use of chemicals, biomaterials and pharmaceuticals (http://www.ecetoc.org/ecetoc-membership/member-companies/). ECETOC Task Force members work within their regular working hours but do not receive compensation by ECETOC. CLE (https://croplifeeurope.eu/) represents company members that develop and supply pesticides and biopesticides and/or invest in precision applications and plant biotechnology traits, as well as national organisations related to these areas. CLE is financed by its membership.

Appendices

Table Appendix 1. Possible scenarios for the Tier 0 in vivo thyroid- and neurodevelopment-related database.

AE: adverse effects; DNT: developmental neurotoxicity; EA: endocrine activity; EDC-T: endocrine disruptor criteria for the thyroid modality; MoA: mode-of-action; WoE: weight-of-evidence. Note: “Findings” relates to statistically significant and biologically relevant (thyroid- or neurodevelopment-related) findings.

Colour legend: dark grey shading: no effect/database not sufficient/no EA/no AE. Light grey shading: database inconclusive regarding EA and/or AE.

Scenario I (Section 2.2.1): Enter Thyroid-NDT-TAS to conduct (Tier 1, Step 1) initial MoA and human relevance assessment and to complete database, beginning with (Tier 1, Step 2) in vitro testing/in silico modelling, as relevant, followed by (Tier 2) higher-tier testing, if necessary for (Tier 3) final MoA and human relevance assessment and final WoE evaluation. The completion of the database addresses the generation of:  • Variant A: thyroid- and neurodevelopment-related data,  • Variant B: thyroid-related data,  • Variant C: neurodevelopment-related data.

Scenario II (Section 2.2.2): Enter Thyroid-NDT-TAS to conduct (Tier 1, Step 1) initial MoA and human relevance assessment and to perform (Tier 1, Step 2) in vitro testing/in silico modelling, as relevant, followed by (Tier 2) non-standard higher-tier testing, if relevant for (Tier 3) final MoA and human relevance assessment and final WoE evaluation.

Table Appendix 2. Research recommendations to enhance assessments of thyroid function and neurodevelopmental impairment (adapted from Marty et al. 2022).

Overarching topic	Recommended research topic/research question
General	Substantiate and refine the insight from Marty et al. (2022) to support the biological relevance of the observational findings and to robustly establish which thyroid-related parameters and threshold(s) is/are useful for a toxicity testing strategy to assess maternal/offspring thyroid hormone imbalance and, ultimately, the substance’s potential to also cause neurodevelopmental impairment
MoA: TPO inhibition	Can differences in the extent of hypothyroxinaemia caused by different TPO inhibitors (that determine whether neurodevelopmental effects will occur) be explained by e.g. differences in toxicokinetics, the presence of further MoAs or other yet unknown reasons?
MoA: DIO inhibition	Under which conditions does DIO inhibition in the liver or brain lead to a reduction or an increase in serum and/or brain T4/T3 levels?
Offspring serum T4 and T3 measurement	Enhance understanding of best suitable timepoints for offspring serum thyroid hormone measurement; investigate whether information on attenuation vs aggravation of offspring T4 decrements during exposure period provides added value to predict neurodevelopmental effects Follow up on the empirically set offspring serum T4 and T3 reduction thresholds to determine their usefulness for a toxicity testing strategy Enhance the reliability of serum T3 measurements
Offspring serum fT4/fT3 measurement	Establish the relevance of fT4/fT3 when assessing substances causing maternal/offspring thyroid hormone imbalance by different MoAs and ultimately possibly also effects on neurodevelopment Address technical difficulties in determining serum fT4 and fT3 levels in rodents
Offspring serum TSH measurement	For substances with direct thyroid-related MoA: Do prolonged offspring serum TSH increases exceeding a certain level indicate that the HPT axis is "overwhelmed" so that adaptation for substance-mediated thyroid hormone imbalance no longer occurs?
Brain T4/brain T3	Expand the database on substance-mediated changes in brain T4 and T3 levels to enhance the understanding of the implications of different patterns of thyroid hormone imbalance on neurodevelopment; aim to establish thresholds of altered brain T4 /T3 that reflect dose levels at which neurodevelopmental impairment will occur Establish the best suitable brain tissues and the best suitable timepoints of measurement of brain T4/T3 levels
Brain-related assessments (general)	Enhance knowledge base on (patterns of) brain-related effects by collecting multiple neurodevelopment-related endpoints in the same study
Motor activity	Do different levels of thyroid hormone imbalance during lactation indicate whether altered motor activity is transient or persistent?
Cognitive function	Enhance understanding of sensitivity of the different test methods use to assess cognitive function
Periventricular heterotopia	Standardise methodology, encourage cross-laboratory validation of this endpoint in view of potential uptake into formal test guidelines
Brain gene expression	Determine suitable sets of genes to assess how thyroid hormone imbalance is linked to neurodevelopment impairment; link these sets of genes to relevant phenotypic alterations Standardise methodologies for regulatory use
Sensitive neurodevelopmental endpoints	Establish whether more appropriate and sensitive sets of neurodevelopmental endpoints to be measured in rats are needed
Magnetic resonance imaging	Determine how this technique “can be combined with functional assessments of the brain to enable comprehensive evaluations of child neurodevelopmental outcomes” and establish its applicability for toxicological assessments using rodents (Sauer et al. 2020)
Toxicokinetics	Develop framework for the integration of toxicokinetics into assessments of substance-mediated thyroid hormone imbalance

DIO: deiodinase; fT3: free triiodothyronine; fT4: free thyroxine; HPT: hypothalamic–pituitary–thyroid; MoA: mode-of-action; PFHxS: perfluorohexane sulphonates; PND: postnatal day; T3: triiodothyronine; T4: thyroxine; TPO: thyroid peroxidase.

Table Appendix 3. Thyroid-related AOPs including neurodevelopmental outcomes in mammals listed in the OECD AOP Wiki in May 2023 (adapted from Marty et al. 2021).

AO(P): adverse outcome (pathway); fT4: free thyroxine; GABA: gamma amino-butyric acid; KE(R): Key event (relationship); MIE: molecular initiating event; NIS: sodium – iodide symporter; PXR: pregnane X receptor; T4: thyroxine; TPO: thyroid peroxidase; TR: Thyroid receptor; TTR: transthyretin; UGT: uridine diphosphate glucuronyltransferase.

Colour legend:  • Text colour = strength of evidence of KER: Green = strong; brown: moderate; blue: weak; black: MIE (since 2nd KE of respective KER is marked) OR: information unavailable.  • Shading of cell = quantitative evidence for KER (always marking 2nd KE of the respective KER): Green = strong; brown: moderate; blue: weak; white: information unavailable.  • The bold boxes highlight that the KEs “T4 in serum; decrease” and “T4 in neuronal tissue; decrease” are central to five of the six AOPs.

AOP 8 refers to rodent studies addressing exposure to polychlorinated biphenyls to support the described sequence of key events (Crofton and Zoeller 2005), and postnatal lactation exposure is indicated as the critical period of exposure (Crofton et al. 2000). For AOP 42, the supporting evidence is mostly derived from in vitro studies and rodent studies, whereas AOP 54, AOP 152 and AOP 300 refer to human studies, rodent studies (and in vitro studies) as providing the supportive evidence. AOP 134 does not summarise the underlying evidence.

[a] Evidence assessment as presented in the AOP Wiki for AOP 8: “Concordance of dose-response relationships: Multiple studies provide limited (2–3 doses) dose-response data for many of the key events. These studies demonstrate similar magnitudes of effect on circulating hormones for doses of PCBs [polychlorinated biphenyls] that are within an order of magnitude (3–25 mg/kg/day for Aroclor 1254)… Very limited data are available correlating any of the key events. One exception is the relationship between circulating serum T4 concentrations during development and the magnitude of hearing loss… There is a very good correlation between total serum T4 concentrations on postnatal day (PND) 14 and hearing loss assessed in adult offspring of PCB exposed dams… All of these events occur within a 2–3 fold dose range….”

[b] Adverse outcome “cochlear function, decreased/loss” indicated in AOP Wiki on 13 September 2019, and replaced by “cognitive function, decreased” by 15 October 2019.