Outcome of the screening exercise

An overview of the screening exercise is shown in Table 4. Following the ECHA/EFSA ED guidance (ECHA/EFSA 2018), sufficient investigation of endocrine adversity, which is the first criterion for identifying a substance as an ED according to WHO’s definition (WHO/IPCS 2002), was demonstrated in all of the 12 substances identified as ED and included in the screening exercise.

Table 4 Summary of the screening exercise of the identified EDs (bold indicates substance selected as a case study)

However, sufficient investigation of endocrine activity, the second criterion of an ED according to the WHO, was only achieved for a few substances. This is mainly due to lack of validated methods for determining thyroid activity or because the respective substances were approved prior to standardized ED-testing and assessment. Nevertheless, for all 12 substances, there is an indication of potential endocrine activity.

Notably, biological plausibility of the link between endocrine activity and adversity, the third and last criterion of an ED according to the WHO, could likewise be established only in some of the substances. For the other substances the respective MoA analysis was not conclusive, albeit not to an extent that a biological link could have been excluded (see Table 4 for more details).

The targeted screening resulted in the identification of three substances for which there was sufficient data as to investigate if and what threshold level could be observed. Namely, this were dimethomorph, metiram and propiconazole.

Case studiesDimethomorph—an anti-androgen

Dimethomorph (CAS 110488–70-5) has been approved as an active substance for the use on potatoes in plant protection products (PPPs) within the European Union (EU) since 2007. It belongs to the morpholine fungicides with respective PPPs being authorized in all EU Member States. For the purpose of renewal of approval an initial draft Renewal Assessment Report (RAR) was prepared in 2017 by the Netherlands as the Rapporteur Member State (RMS) and Germany as Co-Rapporteur Member State (Co-RMS) (Netherlands 2017). Dimethomorph has been under the EU’s Pesticides Peer Review procedure since 2018 with an amendment on the assessment of the endocrine disrupting potential in accordance to the ECHA/EFSA ED guidance added to the RAR in 2019 (ECHA/EFSA 2018).

Under the harmonized classification and labeling (CLH) Dimethomorph is classified as Repro 1B (H360F) and as such meets the cut-off criteria for renewal of approval as set out in Regulation (EC) No 1107/2009. It further fulfills the endocrine disrupting criteria according to the “Overview of the endocrine disrupting (ED) assessment of pesticide active substances in line with the criteria introduced by Commission Regulation 2018/605” (EFSA 2022a).

Together this makes dimethomorph an ideal candidate for a case study as it is subject to a comprehensive dataset spanning multiple studies in two different species (dog and rat) in conjunction with a plausible endocrine mode of action for the observed adversity.

In line with the regulatory process the initial ED assessment for dimethomorph was performed by the RMS with the result subsequently being put forward for peer-review and discussion by the Pesticide Peer Review Expert Panel at EFSA.

The underlying dataset predominantly consists of standard OECD guideline studies together with supplementary data from the ToxCast program (EPA 2022). Briefly, dimethomorph was found to fulfill the criteria of an endocrine disruptor with sufficient experimental animal data to indicate anti-androgenic adversity from dimethomorph exposure. Observed effects include decreased prostate weight and concomitant histopathological findings (fibrosis and prostatitis), increased incidence of testicular effects (e.g., interstitial cell proliferation, hyperplasia and adenoma) and altered sexual development (delayed preputial separation and decreased anogenital distance in males) in dogs or rats, respectively. The effects considered relevant for the identification of dimethomorph as ED are listed in Table 5.

Table 5 Reported findings on dimethomorph related to male reproduction/sexual development—effects considered relevant for ED identification are marked in bold

In addition to the evidence in vivo, anti-androgenic activity of dimethomorph was also observed in the yeast androgen screening (YAS) assay (Table 5) and there was an indication of antagonistic activity in the ToxCast androgen receptor (AR) pathway model (Table 6). Together this was found to sufficiently warrant a biological plausible link between adversity and activity according to ECHA/EFSA’s ED guidance (ECHA/EFSA 2018), ED criteria for EAS-modalities were met (EFSA 2023a).

Table 6 Predictions from the ToxCast models for dimethomorph

The corresponding MoA analysis presented a probable AOP where the molecular initiating event of antagonistic binding to the AR ultimately results in feminization or incomplete development of primary and accessory male organs as well as to impairment of reproductive capacity ((AOP-192021) (Manibusan and Touart 2017)). Substance triggered altered sexual development due to binding to the AR is also supported by other related AOPs (e.g., “androgen receptor antagonism leading to short anogenital distance (AGD) in male (mammalian) offspring”) (AOP-3062021). Based on the weight of evidence dimethomorph has hence to be considered an endocrine disruptor as laid down in Regulation (EU) No 2018/605 (EU 2018b) and according to the ED guidance (ECHA/EFSA 2018).

The studies for the ED assessment of dimethomorph are all compliant in accordance with the principles of Good Laboratory Practice (GLP) and with respective OECD Test Guidelines (see Table 5 for the guidelines used). Data cover multiple species and various exposure scenarios that comprise subchronic as well as chronic toxicity. For each study with reported endocrine adversity there were 3 doses along with a control group. The intervals between doses were appropriately spaced between 3- and fivefold. Therefore, the studies are considered reliable and suitable for a threshold assessment. Table 5 presents an overview of the findings. The no- or lowest-observed-adverse-effect-levels (NOAEL/LOAEL) for each of the relevant endocrine adverse effects can be seen in Fig. 1.

Fig. 1

Overview of the NOAELs/LOAELs for the endocrine adverse effects of dimethomorph relevant for the identification

A clear dose–response relationship with statistical significance at the LOAEL could be observed for nearly all of the adverse effects (e.g., organ weights, anogenital distance). For example, male dogs exposed to the LOAEL of 43 and 44.6 mg/kg bw/d for 13 and 52 weeks exhibited a significant decrease in absolute and relative prostate weight, whereas at the NOAEL of 14.7–15 mg/kg bw/d a decrease was allegedly noticeable but not statistically significant anymore. Likewise, the histopathological findings (e.g., prostatitis and fibrosis) accompanying the decreased prostate weight at the LOAEL at 13 and 52 weeks were not observed at all or only to a much lesser degree at the NOAEL. In dog the respective histopathological findings were less clear. However, given the sample size of 4 animals per dose any in-depth interpretation beyond a qualitative analysis appears inadequate due to the very limited significance. Again, a clear dose–response relationship between exposure to dimethomorph and altered sexual development was observed with regard to altered sexual development in the extended one-generation reproductive toxicity study (EOGRTS; OECD TG 443) in rat. In male pups this comprised anogenital distance and day of preputial separation as well as decrease of weight of prostates and seminal vesicles. Also, there was an increasing trend between exposure to dimethomorph and testicular hyperplasia and/or adenoma.

For dimethomorph the overall evidence therefore strongly supports a threshold for dimethomorph-induced adversity via antagonism of androgen signaling.

Propiconazole and epoxiconazole—multiple putative endocrine modes of action

Triazoles belong to the most frequently used groups of agricultural fungicides because of their efficacy as well as their curative and eradicative properties. The fungicidal activity of azole fungicides roots in their inhibition of the fungal cytochrome P-450-containing (CYP)-enzyme ergosterol synthetase.

Propiconazole (CAS 60207–90-1) is classified as Repro 1B due to its effects on developmental toxicity. Further it has been identified as endocrine disruptor based on an assessment according to the ECHA/EFSA ED guidance (ECHA/EFSA 2018). As active substance it hence fulfills the exclusion criteria of the PPPR and the BPR resulting in non-renewal and listing as candidate for substitution, respectively.

Similarly, epoxiconazole (CAS 106325–08-0) is classified as Repro 1B due to developmental toxicity. It is also classified as carcinogenic Carc. 2. Consequently, its approval as active substance for the use in PPPs has been suspended. Although a formal ED assessment according to the ECHA/EFSA ED guidance (ECHA/EFSA 2018) was never performed, the data on its effects on the endocrine system are so extensive that they were included into this study.

Notably the aforementioned classification of propiconazole and epoxiconazole as Repro 1B comes from the observed ability to cause malformations such as cleft palate and other cranio-facial defects. Biologically such defects are often caused by aberrant retinoic acid (RA) metabolism, for example increased concentrations due to inhibition of CYPs 26 or 3A4 (Robinson et al. 2012). In the case of epoxiconazole also further possible modes of action are discussed (ECHA 2012). Without elucidation of the mode(s) of action for these effects and classification of RA signaling as endocrine or non-endocrine still being subject to scientific debate an assessment of these effects following the ECHA/EFSA ED guidance was not attempted (ECHA/EFSA 2018).

However, apart from effects on RA metabolism the varying ability of azole fungicides to inhibit CYP enzymes can also lead to side effects on hormone production. Unsurprisingly, ED properties of azoles have hence been discussed for quite some time (Marx-Stoelting et al. 2014; Taxvig et al. 2013). The CYP-mediated effects (i.e., via CYP19A1 inhibition) of azole fungicides potentially affect a range of endocrine pathways and comprise substance-induced decrease in estradiol or increase in testosterone with subsequent changes of levels of follicle stimulating hormone (FSH) or adrenocorticotropic hormone (ACTH), respectively. The resulting hormonal imbalances can affect fertility or induce tumor formation in endocrine organs (e.g., ovaries and adrenals for epoxiconazole). Apparently female animals seem more sensitive, which is in line with predominant promiscuous interference with sex hormone metabolism and signaling.

Concomitantly effects reported for propiconazole are estrous cycle disturbance and altered testis weights (Tables 7, 8), however there is no affection of thyroid-related modalities. Epoxiconazole, likewise, shows a number of adverse effects on reproduction. In females these include, among others, ovarian cysts, ovarian theca–granulosa tumors, adrenal cortex adenoma, vaginal hemorrhages during pregnancy and increased placenta weight, while in males sex hormone levels for testosterone, ACTH and FSH can be critically affected (Tables 9 and 10). However, as before propiconazole, epoxiconazole does not show primary thyroid toxicity.

Table 7 Reported findings on propiconazole related to female reproduction - effects considered relevant for ED identification are marked in boldTable 8 Reported findings on propiconazole related to male reproduction - effects considered relevant for ED identification are marked in boldTable 9 Reported findings of epoxiconazole related to female reproduction - effects considered relevant for ED identification are marked in boldTable 10 Reported findings of epoxiconazole related to male reproduction - effects considered relevant for ED identification are marked in bold

Both propiconazole and epoxiconazole have clear effects on steroidogenesis with aromatase (CYP19a1) being one of the primary molecular targets. Yet, other enzymes relevant for steroidogenic metabolism are affected too—one being steroid-17alpha-hydroxylase (CYP17a1). The underlying multi-enzyme inhibition has been confirmed in vivo as well as in vitro and explains the apparent variety of effects observed for azoles such as altered levels of testosterone and/or estradiol or changes in progesterone, prolactin, aldosterone, ACTH, cortisol and/or androstenedione (Draskau and Svingen 2022).

Furthermore, some triazoles have been reported to be potent AR inhibitors (Kjærstad et al. 2010) and so were epoxiconazole and propiconazole with the first being more effective than the latter.

The overlapping actions of this class of fungicides make it difficult to attribute observed adverse endocrine effects to a specific MoA, although a link to a general underlying disruption of sex-hormone metabolism is plausible (Draskau and Svingen 2022; Marx-Stoelting et al. 2014; Taxvig et al. 2013).

Discussion of threshold for adversity relevant for the ED identification

The studies evaluated for the ED assessment of epoxiconazole as well as propiconazole were mainly GLP-compliant studies performed in accordance with specific OECD Test Guidelines (see Tables 7, 8, 9, 10). Similar to dimethomorph, for each study (with one exception) with reported endocrine adversity, three doses were tested along with a proper control group, and the interval between doses was appropriately spaced between 3- and tenfold. Therefore, the studies were considered sufficiently reliable for a threshold assessment regarding respective endocrine adversity (Tables 7, 8, 9, 10). For propiconazole, the most sensitive ED-related effects observed in vivo were estrous cycle disturbances and a decrease in absolute testis weights (LOAEL: appr. 192 mg/kg bw/d, NOAEL: appr. 42 mg/kg bw/d) in a two-generation reproduction study in rats (Anonymous 1985). The most sensitive relevant effects observed in epoxiconazole studies were vaginal hemorrhages, a decreased gestation index and a prolonged pregnancy (LOAEL 23 mg/kg bw/d (Anonymous 1992f)) and ovarian cysts (LOAEL 8 mg/kg bw/d (Anonymous 1992g) in females as well as a reduced fertility index and a prolonged cohabitation time in males at a LOAEL of 23 mg/kg bw/d (Anonymous 1992f). The NOAEL for these effects in both studies (Anonymous 1992f, 1992g) was 2.3 and 2 mg/kg bw/d, respectively.

Besides the OECD-compliant studies there are results from a number of mechanistic assays available (Goetz et al. 2009; Laville et al. 2006; Ohno et al. 2004; Sanderson et al. 2002; Taxvig et al. 2013; Vinggaard et al. 2000). In several cases these assays were conducted in vitro with five or more dose levels tested. As observed with the in vivo data these tests generally demonstrate a clear dose–response relationship, indicating a no effect level (see Tables 7, 8, 9, 10). The exception are those few cases where the respective lowest dose measured was still in the effect range (e.g., 1 µM in aromatase assays; Table 7).

As for dimethomorph the overall evidence strongly supports a threshold for the endocrine effects of propiconazole and epoxiconazole. However, it should be noted that for substances with such a plethora of overlapping MoAs any threshold will inherently be an effect threshold free of any further mechanistic implications.

Metiram—thyroid-disruption

Metiram (CAS 9006–42-2; zinc ammoniate ethylenebis(dithiocarbamate)-poly[ethylenebis(thiuramdisulphide)] belongs to the dithiocarbamates and is a non-systemic fungicide. It was first discovered in the 1960s by BASF and is registered in Europe since the late 80ies. Its predominant use is in PPPs for foliar spray in grapes and potatoes against Plasmopara viticola, Guignardia bidwellii, Pseudopezicula tracheiphila, Phytophthora infestans and Alternaria spp (Italy 2019). As active substance metiram is currently under renewal evaluation, providing it with an extensive up-to-date in vivo data set consisting of several studies from multiple species (rat and dog). The ongoing assessment identified ethylenethiourea (ETU), a metabolite and degradation product of metiram, together with several other metabolites as likely to cause T-mediated adverse effects. Given the updated dataset and modality metiram hence was selected as a third for this study.

The RAR for metiram was prepared in accordance with Regulation (EU) No 2018/1659 (EU 2018a) and put forward for peer review by EFSA’s Pesticide Peer Review Experts´ Meeting (EFSA 2023b). The dataset available for metiram mainly consists of the standard OECD test guideline studies. In particular, metiram was assessed in two short-term (90-day) and a 2-generation reproduction toxicity studies in rats, a 19-week and a 1-year dog study, and a 90-day mouse study. With regard to the relevant metabolite ETU the dataset includes studies with ETU only as well as spiking studies with metiram with 0.2, 2 or 2.2% of ETU added, respectively. ETU is a metabolite, contaminant, and an impurity of metiram having a direct effect on the synthesis of thyroid hormones by thyroid peroxidase (TPO) inhibition (Doerge and Takazawa 1990; Freyberger and Ahr 2006 apud (Italy 2019). Data from bile cannulation (Wenker and Krebbers 2015 apud (Italy 2019)) indicate approximately 7.5% of orally administered to be converted to ETU. Therefore, although generally not quantified further exposure to ETU can be anticipated in all in vivo studies with metiram (Italy 2019).

Other metabolites identified during metiram metabolism studies are ethylenebisisothiocyanatesulfide (EBIS), ethyleneurea (EU), ethylenediamine (EDA), N-(2-aminoethyl)acetamide (n-acetyl EDA), 1-(4,5-dihydro-1H-imidazol-2-yl)-2-imidazolidinethione (Jaffe’s Base) and 2,3,7,8-tetrahydrodiimidazo[2,1-b:1’,2’-e][1,3,5]thiadiazine-5-thione (TDIT) (EFSA 2020; Italy 2019). The underlying pathway sees metiram converted to ETU presumably directly or indirectly via TDIT and EBIS. ETU is then transformed into EU with the generation of other minor metabolites via EBIS methylation being postulated. While there is no toxicological information available for all metabolites repeated dose studies with EBIS as well as with EU show that they can also induce thyroid hypertrophy and hyperplasia (EFSA 2020; Italy 2019).

Substance-induced histopathological changes in the thyroid and changes in thyroid hormones (THs) and thyroid-stimulating hormone (TSH) were observed in three species (rat, mouse and dog) at doses of metiram that were considered to be at or below the maximum tolerated dose (MTD) (Table 11). In rat dose–response relationships were (partially) observed with regard to hormone measurements in a 2-generation study. It should be noted in this context that while the corresponding study failed to record a significant effect on PND 21 the magnitude of the effect was unambiguously higher on PND 21 than on PND 4. Given the statistical significance of the latter it therefore seems questionable to discard the observations made on PND 21. The more so as the same study saw thyroid follicular cell hypertrophy/hyperplasia incidences dose dependently increased in F0 males and F1 adult males and females, respectively. Further in vivo data indicate an effect of ETU on brain morphometry (EFSA 2023b). In contrast to the extensive data in vivo mechanistic studies in vitro are very limited, most likely due to the rapid transformation and degradation of metiram in such systems (EPA 2005).

Table 11 Thyroid effects identified at metiram studies - effects considered relevant for ED identification are marked in bold

Even so the overall data are sufficiently conclusive as to indicate the adverse effects of metiram to be T-mediated (EFSA 2023b). There is empirical support that exposure to metiram causes dose-dependent inhibition of TPO due to the formation of ETU. The resulting inhibition of thyroid hormone synthesis then leads to hypothyroidism, i.e., an increased pituitary TSH release, eventually manifesting in follicular hypertrophy and hyperplasia both of which are known neoplastic precursors in the thyroid.

Discussion of threshold for adversity relevant for the ED identification

Despite its limitations regarding an assessment according to the ECHA/EFSA ED guidance (ECHA/EFSA 2018) the data set on metiram is sufficiently comprehensive to conclude on the existence of endocrine effect thresholds. However, at least with regard to metiram the underlying toxicology does not allow conclusions on single substances as any thyroid effects observed could be either related to metiram and/or its metabolite ETU (EFSA 2023b).

For ETU the lowest recorded NOAEL for thyroid histological endpoints is 0.2 mg/kg bw/d based on a 1-year dog study (thyroid histopathology). This is supported in rat by an extended one-generation study (thyroid histopathology and hormones) and a two-generation study with a parental NOAEL of 0.2 mg/kg bw day 0.27 mg/kg bw day, respectively. However, another study reported thyroid hyperplasia in rat at 0.25 mg/kg bw day after 24 months, indicating some remaining uncertainties for the use of 0.2 mg/kg bw day as a NOAEL (EFSA 2023b).

Contrastingly the lowest recorded NOAEL for metiram is 2.6 mg/kg bw. Again, this is based on the observed thyroid effects (thyroid hormones, thyroid weight and histopathological changes) from the 1-year dog study (EFSA 2023b). In this study, the substance-induced decrease of T4 in males was dose-dependent (13, 26 and 52 weeks). In females the respective measurements for T4 were not always statistically significant at the high dose of 84.8 mg/kg bw for all different time point but were nevertheless clearly reduced at the LOAEL (29.9 mg/kg bw). In rat the lowest LOAEL where thyroid histopathology (thyroid follicular cell hypertrophy and hyperplasia in parental animals, i.e., F0 males and F1 adult males and females) was observed is 9 mg/kg bw day in adult animals, based on a two-generation study. Considering hypothyroidism as the adverse outcome (AO) for metiram, a threshold for each individual Key Event (KE) in vivo can be derived from the 1-year dog metiram study (T4 changes, relative thyroid weight changes and increase severity of thyroid follicular cell hyperplasia) at male and female animals. As for ETU, it causes similar effects on the thyroid hormonal system in rats, mice and dogs, although quantitative differences are seen in dose–response. However, ETU T4 effects in rats are in the same range as for the lowest NOAEL for metiram in dogs. Overall, therefore there is strong evidence that metiram thyroid toxicity (hypothyroidism) follows a dose-dependent threshold behavior.