-
“Nearly 300 natural product compounds have been evaluated for their ability to inhibit aromatase, in noncellular, cell-based, and in vivo aromatase inhibition assays. Flavonoids have been tested most frequently and generally found to be the most active class of natural product AI compounds. Some of the more active flavonoids included apigenin (
, chrysin (11), 7-hydroxyflavone (26), isolicoflavonol (27), (2S)-abyssinone II (45), (2S)-2′,4′-dihydroxy-2?-(1-hydroxy-1-methylethyl)dihydrofuro[2,3-h]flavanone (49), eriodictyol (50), 8-prenylnaringenin (62), 3′-[?-hydroxymethyl-(E)-?-methylallyl]-2,4,2′,4′-tetrahydroxychalcone 11′-O-coumarate (75), isoliquiritigenin (77), and rotenone (132). Other very active AI compounds included the xanthone, ?-mangostin (239), the sesquiterpene lactone, 11?H,13-dihydro-10-epi-8-deoxycumambrin (211), and the anthraquinone, benzanthraquinone I (249).
Since natural product drug discovery efforts frequently utilize non-cellular screening assays, several of the compounds reported to be active in non-cellular assays should be avoided by future AI drug discovery endeavors. This is exemplified by the unsaturated fatty acids, which are commonly found in natural product extracts and have been shown to interfere with non-cellular AI assays [152]. Several flavonoids have also been found to be active in non-cellular screening and inactive in cell-based assays. In natural product AI screening efforts it is recommended that extracts active in non-cellular bioassays be dereplicated for the presence of known aromatase inhibitors prior to expensive and time-consuming bioassay-guided fractionation.
All of the most active compounds were of the flavonoid or xanthone compound classes, with the exception of the active sesquiterpene lactone and the active anthraquinone. The ability of flavonoids to inhibit aromatase has been well established [169, 170] and some flavonoids have continued into in vivo studies with various results [125, 148]. Interestingly, Saarinen et al. [125] have shown that apigenin (
, chrysin (11), and naringenin (59) were all inactive using an in vivo AI mouse model. The AI activity of flavonoids needs further in vivo testing to substantiate the extensive and potent in vitro results. Various AI mouse models are currently available or in development, including a transgenic model that overexpresses aromatase [171], an aromatase-knockout mouse model [172], and a MCF-7 xenograft model [173].”