Bekæmpelsesmiddelforskning fra Miljøstyrelsen, 88 – Hormonforstyrrende effekter af kombinationer af pesticider

Endocrine Disrupting Effects of Combinations of Pesticides

Summary and conclusions

The human population is in most situations exposed to mixtures of chemicals, some of which are suspected of having endocrine disrupting properties. The effect of many of these chemicals is weak compared to the effects of our natural sex hormones, and it is difficult to explain potential adverse effects on humans on the basis of effects observed for single compounds. However, it is conceivable that the endocrine disrupting chemicals interact in a way that enhances their hormonal effect, and, therefore, it is important to elucidate the mechanisms of interaction, that may come into play when compounds are combined. That is the reason why this investigation that specifically aims at elucidating mixture effects of endocrine disrupting chemicals has been initiated.

The overall purpose is also to contribute with input to the discussion on principles of risk assessment of chemicals, that, traditionally, does not take account for the fact that humans are exposed to mixtures of chemicals.

The major part of the work that has been done to study the mixture effects of endocrine disrupting chemicals is based on in vitro experiments of varying quality, and only very few experiments on mixtures of endocrine disrupting chemicals have been performed so far. Investigations of mixture effects of antiandrogenic chemicals have not previously been performed. It has been shown that mixture effects can vary in an experimental system depending on the selected end-point. Compounds that act synergistically at the molecular level (e.g. receptor binding) may show additivity when end-points at more complex functional molecular levels such as cell growth are measured. Because of this, it has been considered important to analyse mixture effects of endocrine disrupting pesticides at different levels of molecular complexity including in vivo studies.

The mechanism of interaction of pesticides with estrogenic and antiandrogenic activity, respectively, was studied in in vitro studies and in vivo short-term animal experiments. As model compounds for the estrogenic pesticides o,p-DDT and methoxychlor were chosen and the test systems comprised a breast cancer cell proliferation assay and a uterus weight gain test in ovariectomized female rats. In the in vitro experiments good agreement between observed effects and predicted effects for additivity based on the concentration addition principle was found. In the in vivo experiments the following end-points were measured: uterus weight, serum levels of FSH, LH, and prolactin, mRNA levels of estrogen receptor (ER) and , and lactoferrin in uterus tissue, and selected neurotransmitters and acetylcholinesterase activity in brain tissues. Briefly, both pesticides gave rise to an increase of uterus weights and prolactin level, whereas LH, FSH level and ER, ER and lactoferin mRNA levels decreased. No effect on neurotransmitter levels and acetylcholinesterase activity was seen. On the basis of the results obtained for organ weights, hormone levels and ER mRNA levels, no deviations from additive effects could be detected, i.e there is no evidence for assuming that synergistic or antagonistic mixture effects take place between the compounds. This is further supported by the in vitro data showing a similar additivity.

The estrogenic activity in serum from rats was analysed using a new biomarker method, in which the physiologically formed estrogens are removed from the sample and the remaining serum fraction is tested in vitro for estrogenic activity. Using this method, an integrated measure for the total estrogenic activity of the xeno-estrogens in the serum sample is obtained – in this case the total activity of the administered pesticides and their break-down products is measured. In the pesticide-exposed animals a clear dose-dependent increase of the estrogenic response was seen, indicating that this method is valuable for elucidating the extent of exposure of animals and humans to estrogenic chemicals.

As model compounds for pesticides with antiandrogenic effects, vinclozolin and procymidon were chosen, and the mechanism of their interaction was studied in an androgen receptor reporter gene assay and in vivo in an extended Hershberger test, that involves the use of castrated male rats. In the in vitro experiments an excellent agreement between observed effects and the predicted values for additive action was found. In the in vivo experiments the following effects were analysed: weights of reproductive organs, serum FSH and LH levels, TRPM-2 and PBP C3 mRNA levels in prostates, and selected neurotransmitter levels and acetylcholinesterase activity in brain tissue. Briefly, both pesticides cause a decreased weight of all reproductive organs, an increase of LH and FSH serum levels and a decreased PBP C3 and an increased TRPM-2 mRNA level in prostates. The sensitivity of the three selected parameters was in these experiments found to be: reproductive organ weights > serum LH levels > gene expression in prostates.

Neither norepinephrine, dopamine concentration, acetyl- or butylcholinesterase activity in brain tissue was affected, indicating that the involved parts of CNS were not affected by the pesticides. However, vinclozolin and procymidon separately and in combination gave rise to a significant increase of the serotonin concentration in the total brain.

In general, it is estimated that the selected experimental design gives rise to a biological variation between the animals on organ weight level that has a magnitude which allows for a relatively precise and detailed investigation of mixture effects. The results obtained give a clear picture of mixture effects for vinclozolin and procymidon. On basis of the present results on organ weights, hormone levels and mRNA levels, no deviations from additivity can be detected that is there is no evidence that any synergistic or antagonistic mixture effects take place. This is further supported by data from the in vitro experiments in which the same additivity is found.

As examples of commonly used pesticides in Denmark were chosen: the carbamate insecticide methiocarb, that has weak estrogenic and antiandrogenic effects in vitro, the herbicide tribenuron-methyl and the pyrethroid insecticide deltamethrin, that both are weak estrogens in vitro, the fungicide prochloraz that has antiestrogenic and antiandrogenic effects in vitro and the herbicide simazin, that is an aromatase inducer. These five pesticides were separately and in combination tested for estrogenic activity in vitro. Methiocarb and deltamethrin induced an estrogenic response measured as an increase of breast cancer cell proliferation, whereas tribenuron-methyl, prochloraz and simazin had no effect. The 1:1:1:1:1 equimolear mixture of the five pesticides did not give any measurable estrogenic response and the calculation of an isobole coefficient of 7 for the mixture indicates an antagonistic effect between the compounds, probably caused by prochloraz that reacts as an antiestrogen when tested together with 17β-estradiol. Furthermore, the pesticides were tested separately and in combination for antiandrogenic effects in vitro and in vivo in an extended Hershberger test. Neither tribenuron-methyl or simazin had any antiandrogenic effect in vitro, whereas both prochloraz, methiocarb and deltamethrin blocked the androgen receptor. A 1:1:1:1:1 equimolear mixture of the five pesticides gave an antiandrogenic response, that was similar to the predicted response under assumption of additivity, and thus the results indicate that the pesticides in combination act additively in this assay. The Hershberger experiment was complicated by the fact that it was not possible to administer the pesticides at the same dose level because of other toxic effects caused by some of the compounds. The dose of deltamethrin and methiocarb therefore had to be lowered to 2.5 mg/kg/day, whereas prochloraz, simazin and tribenuron-methyl each was administered with 25 mg/kg/day. The mixture group got a dose of a total of 80 mg/kg (25+25+2.5+25+2.5 mg/kg). Regarding effects on weights of reproductive organs, no effects of single compounds was seen, whereas the mixture group showed a significantly reduced weight of musc. levator ani/bulbocavernosus. A tendency towards decreased prostate and seminal vesicles weights was seen but no statistical significance was obtained. In general, no effect on hormone levels was found. In contrast, a marked effect of all pesticides dosed separately and in combination on the expression of ODC mRNA and PBP C3 mRNA was seen, with the exception of deltamethrin and tribenuronmethyl that did not affect PBP C3 mRNA.

In conclusion, the five pesticides showed clear additive effects in vitro, and, further, an accumulating effect on weights of reproductive organs in vivo was found. It was surprising that the single pesticides induced marked effects on expression of PBP C3 and ODC mRNA. Gene expression in these experiments may either be a more sensitive parameter for antiandrogenic effects or the genes are regulated by a signal pathway not related to the androgen receptor. The meaning of these changes is not clear at present, but may indicate a very weak antiandrogenic action that does not manifest itself in functional changes.

Investigation of pesticide-induced effects in the offspring after in utero exposure is considered the most sensitive in vivo test for determination of endocrine disrupting effects. The mixture (20 mg/kg/day: 15 mg/kg prochloraz and 1.25 mg/kg of the four other pesticides) and prochloraz in itself (30 mg/kg/day) had a reversible effect on maternal weight during gestation, but no clear foetotoxicity was seen. The originally planned dose of the mixture (40 mg/kg/day: 30 mg/kg prochloraz and 2.5 mg/kg of the four other pesticides) appeared to be too high, as clear visible maternal toxicity was seen after a few doses, and the dose was therefore halved. This toxicity must be caused either by deltamethrin, methiocarb, simazin or tribenuron-methyl in the mixture or by interaction of one or more compounds.

Both prochloraz and the mixture caused a markedly reduced testosterone level and an increased progesterone level in male foetuses at gestational day 21. The reduced testosterone level reflects an antiandrogenic effect comparable to that observed for several phthalates. The effect had disappeared at postnatal day 16. It is conceivable that this effect is the cause of the increased number of nipples seen in males after dosing of prochloraz and the mixture, thereby feminising the males. The increased progesterone level in male foetuses is probably related to the increased lenght of gestation that was significantly increased after prochloraz exposure.

It is conceivable that prochloraz in the mixture (15 mg/kg) is responsible for the antiandrogenic effects on testosterone level and number of nipples, as the compound given alone at a dose of 30 mg/kg gave the same effects.

The mixture caused no effects on neither behaviour nor semen quality in the males, whereas prochloraz in itself affected behaviour in males, as the activity level and behaviour in the sweet preference test was changed, indicating a feminising effect on the males.

The overall conclusion of this project is that endocrine disrupting pesticides that act via the same mechanism of action, have additive effects. This is the clear conclusion from the in vitro experiments and is supported by the observations that have been made in the animal experiments.

The investigation of deltamethrin, methiocarb, prochloraz, simazin and tribenuron-methyl for antiandrogenic effects indicates that there are clear antiandrogenic effects both in vitro and in vivo of prochloraz, whereas no functional endocrine disrupting effects of the four other pesticides in the animal experiments was found. In contrast a number of changes were found at the gene expression level after exposure to even low doses of the four pesticides, but the meaning - if any - of these changes is unknown.

This project illustrates the great advantage of in vitro experiments for evaluation of mixture effects, because of the great capacity of the assays, that allows detailed dose-response curves for many mixture ratios of the compounds. The project also illustrates that investigation of mixture effects in animal experiments is possible, but demands more resources and can turn out to be a difficult task because of differing general toxicity of compounds, meaning that the doses in a mixture that - from a theoretical point of view are optimal,- cannot always be chosen.