Registration Dossier

Toxicological information

Specific investigations: other studies

Currently viewing:

Administrative data

Link to relevant study record(s)

Description of key information

See above.

Additional information

Effects on the endocrine system

Several studies are available that specifically assess the potential of MTBE to interact with the endocrine system in vitro and in vivo. A weight-of-evidenceapproach used to evaluate the toxicological and ecotoxicological MTBE database is also available.

Three in vitro studies were conducted under GLP and to standardised guidelines. The overall assessment that MTBE and its metabolite TBA, were classified as non-binders to the androgen receptor, were classified negative for effects on testosterone and estradiol in the steroidogenesis assay, and were classified as non-inhibitors of aromatase activity. 

The first study (CeeTox, Inc., 2013a) investigated the ability of MTBE and its metabolite tert-butyl alcohol (TBA) to interact with the androgen receptors (ARs) isolated from rat prostates following U.S. EPA OPPTS 890.1150. Under the conditions of the assay, MTBE and TBA were classified as“non-binders” to the androgen receptor in all three independent runs and thus have a final classification of “non-binder” according to the test guideline.  

The second study (CeeTox, Inc., 2013b) investigated the ability of MTBE and TBA to affect the steroidogenic pathway, specifically by inhibiting catalytic activity of aromatase, the enzyme responsible for the conversion of androgens to estrogens, using a human recombinant test system following U.S. EPA guideline OPPTS 890.1200. Under the conditions of the assay, the test substance and its primary metabolite were classified as non-inhibitors, with mean aromatase activities of 101.6 % (± 1.7 % SD) and 102.3 % (± 1.7 % SD), respectively, at the highest test concentrations at 10-3 M.

The third study (CeeTox, Inc., 2013c) investigated the ability of the test substance and its main metabolite to affect the steroidogenic pathway, affecting the production of testosterone or estradiol was investigated in a H295R steroidogenesis assay following U.S. EPA guideline OPPTS 890.1550. The OECD 456 (H295R Steroidogenesis Assay) guideline was used to provide additional guidance in evaluation of the results. Under the conditions of this assay, the test substance and its main metabolite do not cause changes compared to the controls in the production of testosterone or estradiol in accordance with the US EPA guideline. Furthermore, based on the OECD guidelines, both substances should be classified as negative for effects on testosterone or estradiol in this H295R steroidogenesis assay. While statistically significant effects were observed, they were observed in only one run of the assay for each test substance and were not reproducible.

A non-guideline study is available that investigated the potential for MTBE to bind with the estrogen receptor in a Yeast Estrogen Screen (YES) assay. Because the YES assay has not been validated at a national or international level, and without full details available on the test system and methods used, there is limited potential utility for the study. However, several in vitro investigations on potential for MTBE biding with the estrogen receptor are available from Moser et al. (1998). Specifically, MTBE did not displace or compete with bound estradiol from purifed human estrogen receptor, MTBE did not bind to human estrogen receptor present in transected HepG2 cells (luciferase assay), and MTBE did not antagonise the activity of estradiol in the HepG2 system. Together the studies indicate that MTBE does not interact with the estrogen receptor.


A specific investigation on steroid hormone levels and measurements of both aromatase activity and aromatase mRNA in liver and testis microsomes was included as part of three 14-day in vivo experiments in male Sprague-Dawley rats with doses of MTBE ranging from 400 to 1500 mg/kg bw/day (de Peyster & Mihaich, 2014). Serum testosterone and estradiol did not dramatically change in the experiments although the general pattern was a decrease in testosterone and either an increase or no change in estradiol.


Across the experiments, there was a lack of definitive and consistent supporting statistically significant findings in steroid hormone measurements and aromatase activity and mRNA measured in liver and testis microsomes (de Peyster et al, 2013). Evidence of other underling systemic effects were also seen, including reduced body weight gain, increased adrenal weights, and elevated corticosterone suggestive of a more general stress response. When considered together with the results of the results of the three guideline in vitro assays, and a general literature review relating to MTBE and potential effects on the steroidogenic pathway, the authors conclude from these studies suggest that MTBE and TBA do not directly impact the steroidogenic pathway (de Peyster et al, 2013).

The investigation of the steroidogenic pathway included examining P450, as the cytochrome P450-dependent mixed function oxidases (CYP) and hydroxysteroid dehydrogenases have a key role in relevant enzymatic conversions. There were no statistically significant changes in either total P450 levels in liver and testis microsomes at two test exposure concentrations of 600 and 1200 mg/kg bw/day (de Peyster & Mihaich, 2014). These results are consistent with an earlier study in male Sprague-Dawley rats, discussed in section 5.9.3 on reproductive toxicity. which showed total liver P450 to only be increased at the highest dose group of 1500 mg/kg bw/day over 15 days but not at 250, 500 or 1000 mg/kg bw/day and also not after 28 day exposures (Williams & Burghoff, 2000). When measuring indicators of CYP activity, increases were evident at the higher two exposure concentrations, although statistically differences mostly limited to the highest dose of 1500 mg/kg/bw day (Williams & Burghoff, 2000).

A detailed review of the potential for MTBE to induce its own metabolism through increased P450 activity is available in the ECETOC risk assessment report on MTBE (ECETOC, 2003). This report concludes that the level of P450 induction following prolonged exposure to MTBE at high test concentration is low and that the status of P450 in animals receiving such repeated high doses of MTBE is unlikely to be predictive of the cytochrome P450 status in animals or humans at lower exposures. Consequently, NOELs relating to total P450 or P450 activity have a limited relevance in human health risk assessment of MTBE


Overall, the reported effects on the endocrine system caused by exposure to MTBE only occur at high doses (typically above limit doses used in guideline studies) and the available evidence indicates that MTBE does not directly interact with the endocrine system, hence MTBE is a low concern for endocrine effects at occupational and environmental levels. Recent studies have examined potential for effects on the endocrine system resulting from investigations of Leydig cell tumours observed in male rats, which have a very limited or no potential relevance to humans. Moreover, these tumours appear mostly at high systemically toxic concentrations and above metabolic saturation. In turn, high dose levels have been used in the investigative studies in order to elucidate a potential mode of action for previously reported effects.

The significant database on MTBE, which includes several carcinogenicity and reproductive toxicology studies, is consistent in reported signs of clinical toxicity and other adverse effects (e.g. kidney, liver) in rodents exposed to high doses of MTBE. The most consistent findings across studies suggest a role of high dose stress and/or general toxicity, particularly as these can impact hormone homeostasis and the function/structure of endocrine relevant organs such as adrenals and testes (e.g. Everds et al, 2013; Pellegrini et al., 2008). In particular, the findings on the impact on corticosterone levels in the most statistically powerful studies must be considered, which are the carcinogenicity studies in rats and mice (n=100/dose). In addition,liver P450 induction that also occurs with high dose exposure of MTBE may potentially be causing some enhanced metabolism of steroid hormones at exposure levels above which lesions to the liver are evident in histopathological analyses (eg Williams & Burghout 2000, Williams et al 2000). It is therefore highly likely that effects on the endocrine system from high dose MTBE exposures are a result of stress-related conditions of exposure or experimentation, general toxicity, specific toxicities, or a combination thereof.

A weight-of-evidence approach used to evaluate the toxicological and ecotoxicological MTBE database has been performed (de Peyster & Mihaich, 2014). This review followed a method for weight-of-evidence referenced in OECD guidance for evaluating chemicals for endocrine disruption (OECD, 2012). 

The authors note that studies with many non-specific apical endpoints present challenges in the identification of endocrine mode of action and that even some seemingly specific endpoints such as estrogen and testosterone levels can be impacted by non-endocrine modes of action (de Peyster & Mihaich, 2014). As most relevant mammalian studies reviewed were conducted with high doses which appear to be in excess of a maximally tolerated dose and resulting in overt toxicity. The authors explain that this confounds the determination of primary endocrine system involvement if the studies are evaluated in isolation (de Peyster & Mihaich, 2014). Therefore, the weight-of-evidence approach as referenced in regulatory guidance was used to evaluate the in vitro and in vivo literature for MTBE according to eight hypotheses: estrogen agonist/antagonist, androgen agonist/antagonist, thyroid agonist/antagonist, and steroidogenic pathway inducer/inhibitor.

The method involves ranking study endpoints based on their relevance for a particular hypothesis. Evidence for/against each hypothesis based on endpoint relevance and inherent quality (reliability) of the study was determined. Endpoints were assigned one of four ranks: Rank 1 are endpoints are specific and sensitive for the hypothesis and rarely confounded by non-specific activity; Rank 2 are sensitive and specific for the hypothesis but can be confounded by non-specific activity; Rank 3 are primarily apical endpoints that can be affected by systemic toxicity and non-hormonal activity; and Rank 4 are not relevant for the hypothesis being evaluated.

Results of the weight-of-evidence are summarised below (de Peyter & Mihaich, 2014):

Estrogen Agonist and Antagonist Pathway

Vitellogenin (VTG) induction in male fish and an increase in uterine weight in an uterotrophic study are two rank 1 endpoints for the estrogen agonist hypothesis. No guideline uterotrophic assays are available although a one dose, uterotrophic-like study performed at or near the maximum tolerated dose and above the limit dose suggested in the EPA 890.1600 guideline resulted in no change in uterine weight. There are also rank 2 studies measuring the endpoint. Uterine weight changes are species-specific (reduced in mice but no change in rats), and there are conflicting studies evaluating VTG. Two guideline fish studies (OECD 229) were conducted. One study with fathead minnow was negative for male VTG induction while the other with zebrafish was equivocal given that the statistically significant increase in VTG was within the normal historical control range. One non-standard study reported an increase in male VTG; however, methodological problems render the finding suspect. With only minor and non-consistent ranks 2 and 3 responses in high dose rodent studies there is no clear supportive evidence of a direct effect on the estrogen pathway.

Androgen Agonist and Antagonist Pathway

Accessory sex organ weight in rodent studies and secondary sex characteristics (e.g. tubercles) in fish are rank 1 endpoints for the androgen pathway hypotheses. MTBE did not impact either of these two rank 1 endpoints. No effect in the androgen receptor binding assay and inconsistent responses in other in vivo rank 2 endpoints such as testis weight and histopathology and rank 3 endpoints (e.g. uterus and ovary weight, histopathology, testosterone levels) are also not supportive of a direct androgen pathway involvement.

Thyroid Agonist and Antagonist Pathway

Metamorphosis and thyroid histopathology in amphibians and thyroid weight and histopathology in rodents are rank 1 endpoints for the thyroid hypothesis. No relevant amphibian study was available. No differences from controls for the female rodent endpoints were noted. In male rodent studies, thyroid weight was primarily unchanged although there are two studies with conflicting results (e.g., an increase in thyroid weight in one and a decrease in the other). Similar conflicting results were seen in rank 2 endpoints where thyroid hormone measurements were primarily unchanged although there were some increases and/or decreases in a few studies that did not fit a pattern indicative of a direct effect on the thyroid.

Steroidogenic Pathway Induction and Inhibition

Uterus weight in an uterotrophic study is the only identified rank 1 endpoint for the steroidogenic inhibition hypothesis. As noted above, the uterotrophic study had only one dose and it was performed above the limit dose suggested in the EPA 890.1600 guideline. Results of this study and other rank 2 studies show that uterine weight is routinely negative. There were also no effects on female VTG, fecundity or fertilization success in the OECD 229 fish studies. In the OECD 229 fish studies, an effect on fecundity was only seen in one study at an exposure above that recommended in the test guideline and was not linked to an endocrine mode of action. The lack of consistent supporting statistically significant findings in steroid hormone measurements from rank 2 studies, as well as the absence of impacts on aromatase and steroidogenesis enzyme activity in vitro and the fish study in vivo, suggest that it is unlikely that MTBE is interacting with the steroidogenic pathway.

The weight-of-evidence assessment concludes that the presently available results from a wide variety of mammalian and fish studies suggest that MTBE is not directly or primarily impacting endocrine pathways. (Reference: de Peyster & Mihaich, 2014. Hypothesis-driven weight of evidence analysis to determine potential endocrine activity of MTBE. Regul Toxicol Pharmacol 69, 348-370).

The following information is taken into account for any hazard / risk assessment:

Due to confounding effects of general and specific toxicities of MTBE in rodents, as well as potential for stress to cause changes in hormonal homeostasis, studies performed in rodents at high exposures to MTBE limit the potential for understanding how exposure to high doses of MTBE can affect the endocrine system. The lack of effects on endocrine relevant tissues and organs at low exposures to MTBE, as well as no changes to hormone (with exception of corticosterone) or related-enzyme levels at low exposures to MTBE, support a robust conclusion that effects on the endocrine system evident at high exposures to MTBE are not relevant to human health hazard and risk assessment. These high exposure levels (e.g. 1000 mg/kg bw/day or 3000 ppm inhalation) are above metabolic saturation in rodents and are at levels significantly above which other adverse effects (e.g. liver, kidneys), as well as clinical signs of toxicity such as ataxia, are occurring in rodents.

The more recent in vivo toxicological studies on MTBE can be viewed as supporting the previous findings and conclusion on MTBE from the EU Risk Assessment Report (European Commission, 2000). In particular, the potential for endocrine disruption of MTBE based on the findings of Williams et al. (2000) and Moser et al. (1998) were reviewed by the EU Scientific Committee on Toxicity, Ecotoxicity and the Environment (CSTEE) which concluded that the “effects are judged not to be critical” (CSTEE, 2001). This is in line with recent discussions on endocrine disruption classifications, where potential endocrine effects are not deemed as relevant when observed with other toxic effects and/or may arise from secondary non-specific toxicity.

The most recent in vitro assays on endocrine interactions (CeeTox, 2013a, 2013b, 2013c) and two OECD 229 fish studies (Wildlife International Ltd., 2012, 2013) were conducted with a specific aim of investigating potential for MTBE to cause effects on the endocrine system. From these studies, there was no evidence of endocrine interactions, which is important to consider when evaluating mammalian in vivo toxicological data under the OECD framework for evaluating substances for potential endocrine disruption (OECD, 2012).

While the MTBE database continues to grow and its evaluation will continue, a systematic and transparent weight-of-evidence assessment of the presently available results from a wide variety of mammalian and fish studies suggest that MTBE is not directly or primarily impacting endocrine pathways (de Peyster & Mihaich, 2014).