Pesticides Research No. 111 2007 – Effects of azole fungicides on the function of sex and thyroid hormones

Effects of azole fungicides on the function of sex and thyroid hormones

4 Discussion

Several pesticides have been demonstrated to be ER agonists or AR antagonists in various in vitro test systems (Andersen et al., 2002; Kojima et al., 2004). However, the affinities for the steroid receptors are very low relative to endogenous hormones and therefore it could be questioned whether these mechanisms alone can induce effects in vivo at environmentally relevant concentrations. However, it is conceivable that chemicals with the ability to act via several endocrine pathways are more likely to cause effects in vivo due to the combined response from different mechanisms (Sanderson, 2006; Sharpe, 2006). Furthermore, as humans are rarely, if ever, exposed to one single chemical at a time, the combined exposure to several chemicals acting via the same pathway may cause adverse effects in humans.

The results from the in vitro investigations in this study demonstrate that the three azole fungicides: propiconazole, tebuconazole, and epoxiconazole all had the ability, like prochloraz, to act via several different mechanisms. The three triazoles were less potent than prochloraz to act as ER antagonists (prochloraz > tebuconazole » epoxiconazole » propiconazole) and AhR agonists (prochloraz >>> tebuconazole » epoxiconazole » propiconazole) while the potency as AR antagonists was similar. Like prochloraz, the triazoles were aromatase inhibitors, although again, with lower potency than prochloraz (prochloraz >epoxiconazole » tebuconazole > propiconazole).

While several pesticides have been reported to be ER agonists in vitro only prochloraz (Andersen et al., 2002) and some pyrethroid insecticides (Kim et al., 2004) have, to our knowledge, been reported to be ER antagonists. The impact of this mechanism for the integrated response in vivo is unknown at present and no standardized method for investigating this effect in vivo have been established. The results obtained in this study on aromatase (CYP19) inhibition is in accordance with previous studies (Trosken et al., 2004;Trosken et al., 2006;Vinggaard et al., 2000) in which prochloraz was also identified as a stronger aromatase inhibitor than the other fungicides tested.

The higher potency of prochloraz towards the ER, and especially the AhR, indicates that the molecular structure of this compound favour the ability to bind to these receptors. In general, high-affinity AhR ligands are planar aromatic polycyclic molecules such as the prototypical ligand 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (Rifkind, 2006). The number and places of halogens affect the planarity of the molecule and computer modelling could be used for comparison of the conformation of the different azole molecules but such analysis have not been performed. Since the affinities of the triazoles for the ER and AhR were similar in spite of their varied number and placement of halogens, it seems more likely that the differences in affinities between prochloraz and the triazoles are related to the imidazole structure.

The results from the H295R steroid hormone synthesis assay showed that the triazoles, as well as prochloraz, inhibited testosterone and estradiol production, and increased progesterone production. This indicates that enzymes involved in the conversion of progesterone to testosterone are inhibited. Azole fungicides and azole antifungal drugs are designed to inhibit the biosynthesis of ergosterol in fungi by inhibiting the cytochrome P450 enzyme 14a-lanosterol demethylase (CYP51). In fungi the lack of ergosterol leads to collapse of the cell membrane. In humans CYP51 is important for the conversion of lanosterol to cholesterol and in the production of meiosis activating steroids. Thus, inhibition of human CYP51 may disturb steroid synthesis. Prochloraz, propiconazol, tebuconazol, and epoxiconazol have all been reported to inhibit human CYP51 (Trosken et al., 2006) but with lower potency than the antifungal drugs miconazole and ketoconazole. The azole compounds are not selective for CYP51 and a range of other cytochrome P450 enzymes, including aromatase and other CYP enzymes involved in steroidgenesis, are also affected by azole compounds. Hence, CYP17 (17a-hydroxylase and/or 17,20-lyase activity), which are involved in the conversion of cholesterol to testosterone (see Figure 3), has been demonstrated to be inhibited by several azole compounds including the imidazole fungicides imazalil and prochloraz as well as and some imidazole antifungal drugs (e.g., ketoconazol) (Ayub and Levell, 1987;Ayub and Levell, 1989;Mason et al., 1987). Hence, the reduced testosterone and estradiol synthesis and increased progesterone production observed for all four azole fungicides in the H295R steroid synthesis assay might, at least partly, be due to inhibition of CYP17.

Several other cytochrom P450 enzymes have also been reported to be affected by azole compounds. Enzymes of the CYP1, CYP2 and CYP3 gene families were induced by some azoles (Juberg et al., 2006;Mason et al., 1987;Sanderson et al., 2002;Sun et al., 2005) and inhibited by others (Zhang et al., 2002).The triazole fungicide, fenbucazole, was recently reported to be a phenobarbital-type inducer of mouse liver cytochrom P450 and to cause a dose-dependent increase in liver weight associated with hepatocellular hypertrophy and vacuolation (Juberg et al., 2006). In a recent study in adult male rats, the four triazole fungicides: fluconazole, myclobutanil, propiconazole, and triadimefon all affected the expression of several CYP genes in liver and testis including genes involved in steroid metabolism (Tully et al., 2006). The induction of CYP1 enzymes is mediated by the Ah receptor. The Ah receptor is a transcription factor that after binding of ligand causes increased gene transcription of Ah receptor responsive genes such as the CYP1 gene family. CYP1A1 and 1A2 metabolize estradiol to 2- and 4-hydroxy estradiol (Rifkind, 2006) and iIInduction of these enzymes could therefore affect the hormone concentration in the organism. All four azole compounds in this study acted as AhR agonist and hence, they might induce the activity of CYP1 enzymes. Prochloraz was by far the most potent of the four azoles.

No significant and convincing effect on thyroid-receptor mediated proliferation was found for any of the azole fungicides. However, in a similar assay prochloraz inhibited the T₃-induced growth of GH3 cells in a concentration dependent way to 55 % of the reference level (Ghisari and Bonefeld-Jorgensen, 2005) and in animals both prochloraz and tebuconazole have shown effects on thyroid function (Pesticidkontoret, 1996b;Vinggaard et al., 2002). However, this effect could be due to other mechanisms than direct interaction with the TR.

The anti-androgenic effects observed in the AR assay in vitro were not confirmed in vivo when tebuconazole and propiconazole were tested in the short-term Hershberger assay. In contrast to prochloraz that induced marked anti-androgenic effects at doses between 50 and 150 mg/kg bw/day, no anti-androgenic effects were observed for tebuconazole or epoxiconazole at doses below150 mg/kg bw/day. At 150 mg/kg bw/day, FSH levels were increased by propiconazole, and at this dose level both compounds caused a down-regulation of ODC mRNA in ventral prostates. ODC is an androgen-regulated gene but is regulated by other pathways as well including regulation by estrogens (Nellemann et al., 2005). Since none of the other end points (weight of androgen dependent organs, expression of androgen responsive genes or serum concentration of LH) were affected we believe that the effects observed on ODC mRNA expression are due to a non-androgen-regulated pathway. The design of the Hershberger assay is optimised to detect effects on organ weights and not for detecting changes in gene expression and our previous experience supports that changes in weights of reproductive organs is a more sensitive parameter for anti-androgenicity in this animal model than gene expression in prostates. Liver weights were increased with the highest doses of both tebuconazol and propiconazole. A stimulation of liver growth is often observed with chemicals that induce microsomal liver enzymes including the cytochrom P450 enzyme system (Juberg et al., 2006) and this has been observed previously for prochloraz.

No other studies on anti-androgenic effects of azole fungicides in adult animals were found, but among the pharmaceuticals, the imidazole ketoconazole exhibited anti-androgenic effects in rats by reducing serum androgen level and increasing estradiol, FSH and LH levels. The weights of androgen sensitive organs (epididymides, prostate, and seminal vesicles) were reduced (Marty et al., 2001;O'Connor et al., 2002). Ketoconazole also reduced testosterone synthesis in male human volunteers after a single oral dose of 400 mg and ketoconazole as well as other imidazole antifungal drugs (isoconazole, miconazole, econazole, and clotrimazole) inhibited testosterone production in mouse Leydig cells in vitro (Schurmeyer and Nieschlag, 1984). The triazole compound fluconazole did not reduce testosterone levels in men after oral doses of 25 or 50 mg/kg lgv/day for 28 days (Hanger et al., 1988).

Examination of reproductive development in offspring after exposure during pregnancy and lactation is presently considered to be the most sensitive animal model system to detect endocrine disrupting effects. In mammals sex determination and sexual differentiation are under strictly hormonal control and even minor disturbances in steroid hormone concentration or function during critical periods of development may affect the process (Sharpe, 2006). In this study, effects on reproductive development in offspring (fetuses and pups) were investigated for two of the triazoles: tebuconazole and epoxiconazole. Epoxiconazole was not included in the original study plan, but during 2005 we were informed that the use of epoxiconazole had increased dramatically and now was one of the most widely used azole fungicides in Denmark. In light of the negative results for propiconazole in the Hershberger test and of the more limited use of this pesticide in Denmark, we decided to include epoxiconazole instead of propiconazole.

Previous studies have demonstrated that perinatal exposure to prochloraz feminized male offspring and virilized female offspring (Laier et al., 2006; Noriega et al., 2005; Vinggaard et al., 2005a). The results from this study, demonstrate that also tebuconazole cause feminization of male offspring as both doses (50 and 100 mg/kg bw/day) caused increased number of nipples and at the highest dose, the testosterone concentration in fetal testis was reduced. The AGD was unaffected at birth but in male fetuses (GD21) an increased AGD, an indicator of masculinization, was observed for the highest dose of tebuconazole. However, this effect on AGD had disappeared after birth and since the other end-points mentioned indicate feminization, our conclusion is that the dominating effect of tebuconazole in male offspring is a feminization. In female fetuses, both doses of tebuconazole caused increased AGD and, for the highest dose, this effect was also evident at birth. The virilizing effect in female offspring is also supported by a decreased estradiol concentration in the ovaries at PND16, although this effect was not statistically significant. The feminizing effect of males and virilizing effect of females are in agreement with another study demonstrating reduced weight of epididymis in adult male offspring and reduced uterus weight in adult female offspring after perinatal tebuconazole exposure (Moser et al., 2001) although the weight of epididymis and uterus was not affected in the offspring at PND16 in our study.

The testicular progesterone level in male fetuses was elevated after tebuconazole exposure and this, together with the lowered testosterone level, indicates a direct impact on the steroid synthesis pathway in the Leydig cells. This is also supported by the in vitro results on steroid hormone synthesis obtained for tebuconazole. The effects observed on hormone concentrations and nipple retention for tebuconazole are comparable to our previous results for prochloraz (Vinggaard et al., 2005a). The changes in hormone level may (as mentioned earlier) be an indication of CYP 17 inhibition.

Epoxiconazole acts differently in vivo than the other tested azoles, as a marked fetotoxic effect was observed. The dams dosed with 50 mg/kg were in general unable to deliver their pups. Only two litters were born normally while the other litters were included in the caesarian sections at GD21. Thus, the data for this dose of epoxiconazole on the live born pups is based on a very limited number of animals. Like tebuconazole, epoxiconazole seemed to increase AGD in male fetuses but the effect is not persistent at birth. No feminizing effects on male fetuses were seen. In contrast an increased birth weight was seen with epoxiconazole which may be related to the marked up-regulated levels of testosterone in the dams. We suggest that the increased androgen exposure during pregnancy may have had a growth promoting effect on the pups. Overall, epoxiconazole seems to alter sex hormone levels in the dams, but not in the fetuses. In female offspring, AGD was increased in fetuses as well as in pups at birth indicating a virilizing effect. As for tebuconazole, the virilizing effect of epoxiconazole is supported by a lower (non-significantly) estradiol concentration in the ovaries at PND16.

Our findings on epoxiconazole are in accordance with the studies reported when epoxiconazole was approved for use in Denmark (Pesticidkontoret, 2003). In these studies, the fertility of male rats was reduced, gestational length was increased, the number of live born rat pups was reduced, and the weight of the adrenal glands in male offspring was reduced in a 2-generation study in rats using doses between 30 ppm (approx. 3 mg/kg bw/day) and 1500 ppm (approx. 125 mg/kg bw/day) in the feed. To our knowledge, no other studies on reproductive or endocrine disrupting effects of epoxiconazole have been published.

Hence, both triazoles caused virilization of the female offspring after perinatal exposure and these effects resemble those previously observed for prochloraz (Laier et al., 2006). Male offspring was feminized by tebuconazole but not by epoxiconazole. Epoxiconazole was a clear reproductive toxicant and this effect may overshadow any possible endocrine disrupting effect in male pups.The effects induced by tebuconazole were comparable to effects induced by prochloraz (her så vi effekter ved 30 mg/kg) although prochloraz had a higher potency compared to tebuconazole.

Both tebuconazole and epoxiconazole induced a high plasma concentration of progesterone in the mothers. This is probably the reason for the increased gestational length as also previously seen for prochloraz (Vinggaard et al., 2005a). In utero exposure to natural or synthetic progesterons can induce hypospadia in male mice and the synthetic progesterone medroxyprogesterone acetate feminize male and virilizes female genitala (Willingham et al., 2006). Thus, the high maternal progesterone concentration is likely to be involved in the virilizing effect of the female offspring. Whether any potential aromatase inhibition and/or ER antagonism in vivo are also involved cannot be ruled out from these data. The limited hormone analyses performed on the female pups indicate a tendency towards reduced estradiol concentration in the ovary, while effects on fetal progesterone levels were not investigated. For epoxiconazole, the maternal testosterone concentration in plasma was markedly increased and this may also have contributed to the effects observed in the offspring. Besides, it indicates a virilizing effect in the mothers. Since epoxiconazole was not included in the Hershberger assay, we have no hormone measurements from adult male rats but in studies reported as a part of the approval (Pesticidkontoret, 2003) increased serum concentrations of testosterone, androstendion, FSH, and ACTH as well as reduced estradiol, corticosterone, and aldosterone was reported in adult male rats. In females increased serum concentrations of dihydroepiandrosterone, androstendion, LH, FSH, and ACTH as well as reduced estradiol, prolactin, corticosterone, and aldosterone was reported after oral administration of approximately 250 mg/kg bw/day (3000 ppm in the feed) for 28 days.

The endocrine disrupting properties demonstrated for many pesticides in vitro have only been confirmed in vivo for a limited number of pesticides. These include some of the ‘old’ persistent organochlorine insecticides such as DDT and methoxychlor (Shelby et al., 1996), and some fungicides: procymidon (Gray, Jr. et al., 1999b;Lambright et al., 2000;Ostby et al., 1999), vinclozolin (Gray, Jr. et al., 1999a;Kelce et al., 1994), fenarimol (Andersen et al., 2006;Vinggaard et al., 2005b), and prochloraz (Vinggaard et al., 2002;Vinggaard et al., 2005a). General traits for these pesticides are their ability to react with more than one steroid hormone receptor (Andersen et al., 2002;Gaido et al., 2000;Molina-Molina et al., 2006;Radice et al., 2006;Sohoni and Sumpter, 1998) and to affect the expression of enzymes involved in steroid synthesis and/or metabolism (Delescluse et al., 1998;Sanderson, 2006;Vinggaard et al., 2006;Zachow and Uzumcu, 2006). Whether the ability to act via different mechanisms is a major contributing factor in inducing endocrine disrupting effects in vivo, is not known at this time. The importance of the different mechanisms for the integrated in vivo response may differ between compounds and may also depend on the developmental stage of the exposed organism.

We have previously demonstrated that another fungicide, fenarimol, had estrogenic properties in vitro in the MCF-7 cell proliferation assay as well as in vivo where the compound increased uterine weight in ovariectomized female rats (Andersen et al., 2006). In addition, fenarimol is an anti-androgen in vitro (Vinggaard et al., 1999) and causes anti-androgenic effects in the Hersberger assay in vivo (Vinggaard et al., 2005b). Fenarimol inhibits aromatase activity (Sanderson et al., 2002; Vinggaard et al., 2000) and affects a range of other cytochrom P450 enzymes including key enzymes involved in steroid hormone biosynthesis (Paolini et al., 1996). Like the azole fungicides, fenarimol was reported to increase progesterone concentration in pregnant rats leading to delayed parturition (WHO, 1995). In some older studies, a reduced fertility of male rats was observed and suggested to be related to an effect in the central nervous system controlling sexual behavior (Hirsch et al., 1987) but the exact mechanism were not explored. No studies on reproductive development in offspring after exposure of pregnant animals to fenarimol have been reported.

The dicarboximide fungicide vinclozolin was one of the first pesticides identified as a potent in vitro and in vivo androgen antagonist and to disturb sexual development in male pups after perinatal exposure exhibiting reduced AGD, nipple retention and hypospadia (Colbert et al., 2005; Gray, Jr. et al., 1994; Gray, Jr. et al., 1999a; Kelce et al., 1994; Nellemann et al., 2003). Recently, vinclozolin was also reported to virilize female fetuses in mice after prenatal exposure and to up-regulate progesterone receptor mRNA in offspring of both sexes while ER mRNA was down-regulated in females and up-regulated in males (Buckley et al., 2006). Vinclozolin has also been reported to affect the activity of a range of cytochrom P450 enzymes in mice by increasing the activity of some and inhibiting others depending on dose, sex and organ (Hrelia et al., 1996). In addition, vinclozolin was reported to increase the activity of aromatase and mRNA expression in H295R human adrenocortical cells identified (Sanderson et al., 2002). A recent study demonstrated that vinclozolin and its two major metabolites were not only AR antagonists, but also ER agonists and that vinclozolin and one of the metabolites (M²) were progesterone and mineralocorticoid receptor antagonists and M² was also a glucocorticoid receptor antagonist (Molina-Molina et al., 2006). Hence, also vinclozolin possess the ability to induce endocrine disruption via a range of different mechanisms. Recently, vinclozolin was shown to alter the spermatogenic capacity of male germ cells and sperm viability via its effects on DNA methylation and fetal exposure to vinclozolin during gonadal sex determination in rats caused reduced fertility and sperm development in the adult testis (Uzumcu et al., 2004). Alarmingly, this phenotype was transmitted through the male germ line to at least generation F4 without any further exposure (Anway et al., 2005). Testis from the F2 generation at PND20 had normal morphology but an increase in spermatogenic cell apoptosis. At PND60 (adult) the morphology was still predominantly normal but the germ cell apoptosis was significantly increased (Anway et al., 2006). A similar transgenerational transmission of endocrine disruption was also demonstrated for methoxychlor (another well known endocrine disrupting pesticide) in the same study (Anway et al., 2005) indicating that this phenomenon also might apply to other endocrine disrupters.

In conclusion, all three triazole fungicides included in this study possessed similar properties as prochloraz to act via several endocrine disrupting pathways in vitro although the potency in some of the assays were lower than for prochloraz. However, the potency of the triazole fungicides in vivo might be higher than predicted from the in vitro studies since both tebuconazole and epoxiconazole were capable of inducing effects on reproductive development in the offspring after exposure in utero. These effects are likely due to several different mechanisms operating simultaneously and thereby enhancing the integrated biological response. The lack of effect in the Hershberger assay (where effects on steroid synthesis is omitted by using testosterone supplemented castrated male rats) combined with the effects on reproductive developmental after perinatal exposure strongly indicate that one of the main responsible mechanisms is disturbance of key-enzymes involved in synthesis of steroid hormones. Subtle effects on hormone levels might be without clinical effects in adults because of a tightly controlled homeostasis but could have detrimental effects if they occur under vulnerable stages of reproductive development in the fetus (Sharpe, 2006; Toppari et al., 2006).

Although the endocrine disrupting potency for the individual triazole compounds seems rather low compared to the concentrations of these compounds in the environment and diet, exposure to several azole fungicides (both triazoles and imidazoles) simultaneously is likely due to the wide use of these compounds. Since, the azole fungicides share several mechanisms (e.g., ER and AR antagonism, Ah receptor agonism, and effects on steroid hormone synthesis including inhibition of aromatase) the combined effects induced by these fungicides might be additive. Likewise, they can also add to effects induced by other environmental endocrine disruptors sharing similar mechanisms (Birkhoj et al., 2004; Rajapakse et al., 2002; Silva et al., 2002). Thus, it is important to survey and minimize the exposure of the human population to azole fungicides.