Differential biochemical and cellular actions of Premarin estrogens: distinct pharmacology of bazedoxifene-conjugated estrogens combination

Abstract

The use of estrogen-based therapies and the selective estrogen receptor (ER) modulator (SERM), raloxifene, which are approved for postmenopausal osteoporosis, is associated with side effects such as uterine/breast hyperproliferation, thromboembolism, and hot flashes. A combination of a new SERM, bazedoxifene (BZA), and Premarin (conjugated estrogens; CE) is under investigation to mitigate the estrogen/SERM side effects with promising results in Phase III clinical trials. To explore the mechanism of BZA/CE action, we investigated the recruitment of cofactor peptides to ERalpha by components of CE and a mixture containing the 10 major components of CE with or without three different SERMs. Here, we demonstrate differential recruitment of cofactor peptides to ERalpha by the individual CE components using a multiplex nuclear receptor-cofactor peptide interaction assay. We show that estrone and equilin are partial agonists in comparison with 17beta-estradiol in recruiting cofactor peptides to ERalpha. Further, CE was more potent than 17beta-estradiol in mediating ERalpha interaction with cofactor peptides. Interestingly, BZA was less potent than other SERMs in antagonizing the CE-mediated cofactor peptide recruitment to ERalpha. Finally, in accordance with these biochemical findings, 17beta-estradiol and CE, as well as SERM/CE combinations, showed differential gene regulation patterns in MCF-7 cells. In addition, BZA showed antagonism of a unique set of CE-regulated genes and did not down-regulate the expression of a number of CE-regulated genes, the expression of which was effectively antagonized by the other two SERMs. These results indicate that SERMs in combination with CE exhibit differential pharmacology, and therefore, combinations of other SERMs and estrogen preparations may not yield the same beneficial effects that are observed in clinic by pairing BZA with CE.

Fig. 1.

Premarin CE components mediate differential ERα LBD-cofactor interactions. The 10 major Premarin estrogenic components were characterized individually at 10 μm in the ERα-LBD multiplex assay. Hierarchical clustering of the CE components was performed in Spotfire DecisionSite 8.1.1 using the UPGMA clustering method with Euclidian distance similarity measure and average value ordering. The 20 peptides showing the greatest interaction with the individual CE component-bound ERα-LBD are included in the heat map.

Fig. 2.

Individual CE components induce selective peptide interactions. To illustrate the selective ERα-peptide interactions induced by the 10 major Premarin estrogenic components, the mean fluorescence intensities were plotted for five of the peptides: SRC1 IV (A), DRIP205 II (B), CBP I (C), p300 I (D), and MNAR II (E). Each of the 10 CE components was used at a saturating concentration of 10 μm.

Fig. 3.

A comparison of 17β-estradiol and CE mix in multiplex ERα-cofactor peptide recruitment assay. The ERα-LBD multiplex assay was performed in a dose-response mode using 300 nm to 0.1 nm 17β-estradiol (A) or 1000 nm to 0.3 nm of the CE mix (B). The mean fluorescence intensity for each dose was plotted for 35 of the peptides tested. The EC₅₀ values and associated sem values for 29 of the peptides fitted to a nonlinear regression model are summarized in Table 3. E2, Estradiol.

Fig. 4.

Effect of SERMs on CE mix-mediated recruitment of cofactor peptides on to ERα LBD protein. The ERα multiplex assay was performed with 10 nm CE alone or in combination with 30 nm BZA or the SERMs, RAL and LAS. The mean fluorescence intensity for each treatment was plotted for 35 of the peptides. Tam, Tamoxifen.

Fig. 5.

Differential global gene expression profiles of estrogens, SERMs, and their combinations. The visualization of the hierarchical clustering heat map shows records in a dendrogram based on the similarity among different treatments. The clustering method is described in Materials and Methods. The calculation settings were done using total number of selected records (8374, all genes that were significantly regulated by at least one treatment) that used z-score normalized mean of signal intensities for each treatment condition. E2, Estradiol.

Fig. 6.

SERMs with and without CE combination exhibit differences in their gene expression patterns. Venn diagram analysis showed overlaps of gene lists differentially regulated by either 17β-estradiol or CE (A, >3-fold; P < 0.01), by the three SERMs (B, >3-fold; P < 0.01; C, >2-fold; P < 0.01) and by the three SERM plus CE (D, >3-fold; P < 0.01).

Fig. 7.

SERM/CE combinations (TSECs) exhibit differential gene regulation patterns. A heat map using total number of selected records (1375, from selected clusters of interest) that used z-score normalized mean of signal intensities for each treatment condition is shown. Total number of genes shown in the heat map (1375) were again sorted by a filter of greater than 0.25 (indicated by red) or less than −0.25 (green) of their corresponding z-scores for each treatment condition to generate significantly differentially regulated genes among the four treatments. A–D, A subset of genes obtained after the four treatments.