1 Introduction
The topic of dose level setting is intrinsically linked to the basic principle of toxicology. When in 1538, Paracelsus expressed the classic toxicology dictum “ Alle Dinge sind Gift, und nichts ist ohne Gift; allein die Dosis macht, dass ein Ding kein Gift ist” - which is often condensed to “ The dose makes the poison” - it became already recognized that all chemicals, even water, oxygen and kitchen salt can be toxic if too much is drunk, inhaled or consumed, thereby acknowledging the combination of the intrinsic toxic property of the chemical and the level of (human) exposure, or consumption. Also, the classification of many human health endpoints is based on a combination of the intrinsic toxic properties of a substance (hazard identification, and the amount administered (hazard quantification) most often resulting from data from animal studies. However, whereas regulatory authorities apply this concept to most human health endpoints they make an exception for carcinogenicity and reproductive toxicity where classification is based on the level of evidence and thus hazard identification only, and not on the quantification or potency level, viz. the amount of chemical administered causing the effects.
Recently, the European Chemicals Agency (ECHA) has emphasized the importance of dose level selection in animal studies specifically targeted to investigate reproduction and/or developmental toxicity [ 1], which, in our view, is aiming at a classification for these endpoints by default. In their recent report ‘Review of 55 extended-one generation reproductive toxicity (EOGRT) studies under REACH’, ECHA considered insufficient dose level setting as one of the critical issues hampering hazard identification, and thus classification, in 20% (11 out of 55) of the EOGRT studies reviewed [ 2].
This discussion paper is meant to highlight the challenges faced by industry and test labs as a result of ECHA’s advice to select high dose levels when conducting EOGRT (OECD TG 443) studies. The same applies to postnatal developmental toxicity studies (OECD TG 414).
2 What level of toxicity should be accepted for dose level setting?
In the REACH Annexes on information requirements, it is explicitly mentioned in the amendment of 17 th June 2021 (Commission Regulation 2021/979) that: “ Where a test method offers flexibility in the study design, for example in relation to the choice of dose-levels, the chosen study design shall ensure that the data generated are adequate for hazard identification and risk assessment. To this end, testing shall be performed at appropriately high dose levels” [ 3]. The question then is: how should ‘appropriately high’ be defined?
ECHA’s advice on dose level selection mentions that “ Irrespective of the specifications in the OECD TGs regarding the selection of the highest dose, for classification and labelling, it is critical that the tested doses are sufficiently high to also be able to conclude on a lack of clear evidence on reproductive toxic properties warranting a classification as Repr. 1B for the tested parameters” [ 1]. With this statement, it appears as if ECHA is mainly (if not exclusively) interested in substance classification, thereby neglecting the fundamental principle of toxicology that the ultimate toxicity is determined by a combination of the intrinsic properties of a substance in combination with the dose level. Or in other words, studies showing no reproductive findings at the highest dose tested – using doses far in excess of any likely human exposure - could still be relevant for ECHA to use in human risk assessment.
3 Selection of too high dose levels leads to unacceptable animal suffering
OECD TG 443 for EOGRT studies mentions: “ If dose levels are based on toxicity, the highest dose should be chosen with the aim to induce some systemic toxicity, but not death or severe suffering of the animals ” [ 4]. However, in another ECHA document on dose level selection for EOGRT studies it was indicated that: “ To be compliant and not rejected due to too low dose-levels, the highest dose level must induce clear evidence of an adverse effect on sexual function and fertility […]” [ 5]. This does not only mean that reproductive toxicity studies may even be rejected when dose levels are considered too low, it also seems to imply that the highest dose level of any substance must induce clear evidence of an adverse effect on sexual function and fertility, which in fact would mean that – as already concluded by Paracelsus – any chemical could induce reproductive toxicity as long as the dose would be sufficiently high. Indeed, it is quite likely that in case of severe toxicity (including death) mating will be limited if occurring at all. If so, reproductive toxicity testing would even become redundant, and any chemical could actually be classified for reproductive toxicity by default.
As a result of the required high dose level testing, we are currently facing profound discussions with our contract labs on dose level selection. Together with the labs we feel forced to increase the doses to unethically high levels leading to too much animal suffering, toxicity that is too severe and unnecessary morbidity and mortality, especially in dose-range finding (DRF) studies. However, even this does not provide any guarantee for the main studies because DRF studies are generally of a much shorter duration, viz. 14-28 days versus at least 90 days in an EOGRT study.
We had a recent case of a DRF study in preparation of an EOGRT study where dose levels were selected according to the abovementioned ECHA requirements. In this DRF study F0 time-mated females were dosed from gestational day 6 until lactation day 20 and F1 selected offspring were dosed between day 21 until 34 of age. During DRF testing it became evident that the test item had a very steep dose-response curve and pups, not dosed before, appeared to be substantially more sensitive than their mothers. In fact, the toxicity observed in the animals was exceeding the acceptable limits under the lab’s Animal Welfare License after which the lab was forced to prematurely sacrifice the animals. This resulted in additional DRF testing with several other (lower) dose levels to select the hopefully appropriate dose levels for the scheduled EOGRT study. In our view, the current trend to request higher dose levels leading to more animal suffering cannot be aligned with one of the basic principles of the REACH regulation stating that (Article 25): “ ….testing on vertebrate animals for the purposes of this Regulation shall be undertaken only as a last resort ” [ 6]. Moreover, this is also fully in contradiction with the fundamental toxicology 3R’s principle and the request from society to ban animal tests. It will also lead to more and probably endless discussions whether findings are the result of the intrinsic property of a substance or the result of general (maternal) toxicity as a consequence of too high dosing.
In a recent paper, van Berlo et al. [ 7] suggested that a body weight decrement greater than 10% compared to controls (thus at least a 10% lower body weight than controls) should be the MTD criterion for non-carcinogenicity studies. However, according to OECD Guidance Document (GD) 19 a body weight decrement of 20% compared to controls already qualifies for humane sacrifice due to excessive toxicity [ 8, 9] which was acknowledged in the rebuttal by van Berlo et al. [ 10].
But reduced mean BW gain over time can consist of individual days of BW gains and losses and application of the MTD criterion based on BW alone becomes especially challenging for pregnant animals in Developmental and Reproductive Toxicology (DART) studies where even single day(s) of maternal toxicity can have potential consequences on the developing fetuses [ 9]. Thus, if a high dose with 10% lower mean BW compared to controls would not be considered sufficiently excessive or severe, and a high dose with 20% lower mean BW compared to controls would be candidate for humane euthanasia, not much space is left to select the adequate high dose based on a DRF study of shorter duration. Notwithstanding the fact that 10% BW decrement compared to controls is already very large, to end somewhere in the 10-20% range as van Berlo et al. [ 10] seem to suggest would in practice require one or more quite extensive DRF studies consisting of several doses and a study duration equal to the duration of an EOGRT study.
4 Focus on fertility in an EOGRT study by ECHA impairs investigation of other aspects of toxicity such as developmental toxicity
In their clarification on dose level selection for EOGRT studies ECHA claims that “ The focus of the OECD TG 443 study in the REACH annexes is on sexual function and fertility…’ and “ As the study should be designed to ensure adequate assessment of the effects on sexual function and fertility, the dose levels should not be reduced to get enough offspring for the assessment of developmental toxicity” [ 1]. This is a very typical statement, in fact an unfounded opinion, and an intrinsic conflict of the different aims of the EOGRT study – especially when developmental (neuro/immuno)toxicity (DNT/DIT) cohorts need to be added. In our view, addressing the different aims of an EOGRT study cannot be solved without a compromise. While prioritizing for the selection of (higher) dose levels that would allow the identification of potential effects on sexual function and fertility (although according to ECHA “ the highest dose level must induce clear evidence of an adverse effect on sexual function and fertility”; see previous section), this may lead to a significant (but according to ECHA acceptable) reduction in the number of offspring. However, in OECD GD 151 the following is stated regarding the evaluation of DNT: “ …..Interpretation of TG 443 DNT test results should take into account available information on mechanisms of action, toxicokinetics, maternal toxicity and potential indirect effects on offspring, as well as any available data on neurotoxic effects of the specific test chemical” [ 11] indicating that evaluation of pups should be part of the investigations in an EOGRT study which is impossible in cases where the number of offspring is insufficient for this kind of evaluation. The lack of sufficient animals is even more an issue in the extensive EOGRT studies where additional cohorts and/or testing is requested. Insufficient offspring would mean that under the current REACH requirements there would be no assessment of toxicity possible in the developing offspring. The only other study addressing fetal development is in an OECD TG 414 study but here the development of fetuses is only studied until the end of gestation.
In our view, testing at dose levels that are too high is in clear conflict with the initial requirement of EOGRT studies being able to also capture (subtle) neurobehavioral changes to identify a developmental neurotoxicity hazard. Moreover, with a focus on classification for reproductive toxicity and its requirement for high dose testing, any findings in pups (if still available) may be mistakenly interpreted as (neuro)developmental toxicity although being the mere consequence of high maternal toxicity resulting in e.g. lack of care by the mothers. Also, as indicated in our earlier example, pups may be more sensitive than adult animals requiring the introduction of asymmetric dosing, whereby adult animals and pups are exposed to different doses. But even more, it should be questioned whether the neurodevelopmental cohorts (as well as the immuno-developmental cohorts) should be dosed at all; not only to be in line with the OECD test guideline to investigate neurodevelopmental toxicity (OECD TG 426) but also to be able to distinguish between developmental neurotoxicity potentially developed during gestation and/or lactation or neurotoxicity caused by direct exposure of the pups after weaning at PND 21, see also our recent paper [ 12]. Also, other agencies may not accept the high dose levels used because of nonlinear kinetics that could occur at high (irrelevant) maternally toxic doses, and could request for additional studies.
5 Contact with authorities on dose level selection is highly welcomed
At this moment it is uncommon to be in direct contact with the authorities before a REACH dossier is submitted or updated. In fact, all test proposals for higher tier (more advanced) studies as well as any request for studies as part of a compliance check are processed via IUCLID and REACH-IT. In the early days before REACH came into force it was possible to meet with the competent authority to go through testing requirement(s) and discuss any details of the test(s) or substance. In fact, in the case of our earlier example, we asked ECHA for a consult which was actually accepted, and appreciated by both parties. We are in favour of more of such interactions.
6 Conclusion
In our view, the current mindset at ECHA – backed up by member state competent authorities - is focused on testing at dose levels that are too high for repeat dose toxicity studies. This may lead to an unethical increase in animal use and animal suffering, an inability to correctly interpret the results of such studies and an unacceptable request for the repetition of studies leading to an even higher demand for experimental animals. This does not align with the current trend to promote 3R’s and non-animal testing. We also would like to suggest that ECHA re-evaluates the current requirements of an EOGRT study which has as a target to investigate both fertility and sexual behavior as well as development of offspring until sexual maturity, as there are no other REACH Annex VII-X studies in which these endpoints are examined.
We also plea for ECHA to have more interaction with registrants on, among others, dose level selection for repeat dose toxicity Annex IX and X studies. This would create a better understanding of each other’s viewpoints and will ultimately result in improved registration dossiers and the use of less animals.