As expected the nortestosterone derived progestins

As expected, the 19-nortestosterone-derived progestins tested were moderate agonists of hAR (C), with EC50s ranging from 2 to 10nM, however all were less potent than the endogenous ligand DHT by at least a factor of 6, with relative potencies ranging from 0.0301 to 0.164. This is in agreement with previous in vitro studies showing that although members of this class of progestins exhibit androgenic activity, they activate hPR with greater potency than hAR (, , ). In contrast, the potency of 19-nortestosterone-derived progestins against rainbowfish ARα was approximately equivalent to that of DHT (D), with EC50s ranging from 0.36 to 0.83nM and relative potencies ranging from 0.74 for 19-NE to 1.73 for LNG. Similarly, the majority of progestins tested were strong agonists of rainbowfish ARβ, exhibiting EC50s at least one-third that of DHT (E). A notable exception was NGM, which only weakly activated rainbowfish ARβ at relatively high concentrations but was a strong agonist of rfARα. The spironolactone-derived progestin, DPN, did not activate rainbowfish ARα or ARβ at the concentrations tested. Overall, our observations are in concordance with the recent study by , which showed that fmAR, an androgen receptor typical of the single AR subtype expressed by otomorphs (), was activated strongly by 19-nortestosterone-derived progestins and only weakly activated by DPN. Our study shows that in general, rainbowfish ARα and ARβ, androgen receptors typical of the two subtypes expressed in members of the acanthomorpha, are also activated by 19-nortestosterone-derived progestins. The notable exception was NGM, which activated rfARα strongly but was a very weak agonist of rfARβ. The structure of NGM differs to that of other ethinylated gonanes (e.g., LNG, GSD and ETG) by the presence of an acetate group at position 17, which is likely to be rapidly cleaved by intracellular esterases (), and an oxime group at position 3, thought to be responsible for the observed reduction in androgenic activity in mammals relative to LNG (, , ). Since the amino FK866 identity between rfARα and rfARβ in the ligand-binding domain is only approximately 70% (), it is perhaps not surprising that differential responses to some compounds are observed. Indeed, we have previously shown that a synthetic androgen, 17α-trenbolone, preferentially activates rfARα (). In the case of NGM, structural differences between the ligand-binding domains of the two rainbowfish AR subtypes probably underlie the specificity for ARα observed in the present study.This specificity is likely to involve interactions between the receptor and the oxime group at position 3. It should also be noted from studies in mammals that although the major metabolite of NGM is levonorgestrel-oxime, a proportion of NGM is metabolised further to LNG (, ), suggesting that in vivo effects in fish would also be influenced by how the compound is metabolised. Our findings suggest that androgen-regulated gene expression would be modulated differentially in percomorph teleosts by exposure to compounds such as NGM depending on whether ARα or ARβ is the main driver of gene expression in a given tissue, organ or cell type. Identifying differential tissue distribution and determining which target genes are preferentially regulated by ARα and ARβ would help to predict the likely effects of this progestin in vivo. The findings presented here confirm that 19-nortestosterone-derived progestins do not activate fish PRs but are potent agonists of fish ARs. Consequences for the general applicability of the read-across hypothesis () are clear – for certain pharmaceuticals, predicting the effects in any given species based on those observed in a model organism may not be feasible, despite the evolutionary conservation of the pharmacological target in terms of primary amino acid sequence. In vitro transactivation experiments can aid in determining the potency of steroidal compounds against different receptors, thereby predicting off-target effects in various aquatic organisms that may be exposed to pharmaceuticals in the environment. While testosterone-derived compounds were the focus of the present study, the latest generation of P4-, pregnane- and norpregnane-derived progestins could also be readily characterized using similar experiments. have shown that one such compound, nomegestrol acetate, did not activate either PR or AR from fathead minnow, however the most likely biological off-target activity for this group of compounds is anti-androgenic activity (), which was not investigated in the study.