Binding affinities of various endogenous estrogen metabolites for human estrogen receptor α and β subtypes: Insights into the precise structural requirements favoring a preferential binding affinity for the β subtype

2876

To search for selective endogenous agonists for human estrogen receptor (ER) α and β subtypes, and also to gain insights into the precise structural determinants required for the selective activation of each subtype, we systematically compared the binding affinities of over 50 endogenous metabolites of 17β-estradiol (E₂) and estrone (E₁) for human ERα and ERβ. Among ∼20 A-ring estrogen metabolites (mostly catechol estrogens and their methyl ethers) tested, 4-hydroxy-E₂ and 2-hydroxy-E₂ displayed similarly high binding affinity for both ERα and ERβ, whereas 2-hydroxy-E₁ had a slightly preferential binding affinity for ERα over ERβ. While most of the methylated catechol estrogens did not have appreciable binding affinity for ERα and ERβ at the highest concentration (1000 nM) tested, a few of them (2-methoxy-E₂ and 4-methoxy-E₂) retained rather high binding affinity for both ER subtypes. Among over 10 B-ring metabolites tested, 6α-hydroxy-E₂ and 6β-hydroxy-E₂ were found to have a reduced affinity for ERα and ERβ, but 6-keto-E₂ retained a rather high affinity for these two ER subtypes. Interestingly, various 11-position metabolites (C-ring metabolites) did not have any appreciable binding affinity for either subtype. Most of the D-ring metabolites retained considerable binding affinity for both ERα and ERβ, but several of them had distinct preferential binding affinity for ERβ over ERα. Also, several of the dehydroestrogen metabolites displayed slightly higher binding affinity for ERβ than for ERα. According to the X-ray crystal structures of human ERα and ERβ, we also performed computational modeling analyses of the binding affinities of various endogenous estrogen analogs for ERα and ERβ. These analyses yielded useful information on the subtle structural determinants required for the preferential activation of human ERβ subtype. The knowledge derived from our present study will aid greatly in our on-going efforts to rationally design selective ligands for human ERβ. [Supported, in part, by an NIH grant (CA 97109).]