27-hydroxycholesterol binds directly to ESR1 and ESR2 to modulate estrogen signaling in a cell-, tissue-, and gene-specific manner, making it a physiological selective ER modulator (SERM) (Umetani et al, 2007; Nelson et al, 2013; Nguyen et al, 2015). In the context of breast cancer, 27-HC acts as an estrogen agonist, promoting ER-dependent cellular proliferation. The development of resistance to aromatase inhibitors in breast cancer can arise in part through epigenetic reprogramming that activates the cholesterol biosynthetic pathway, elevating 27-HC levels and resulting in constitutive ER alpha activation (Nelson et al, 2013; Nguyen et al, 2015). Note that 27-HC binding to the estrogen receptors likely occurs in the context of a chaperone complex as is the case for estrogens, however this has not been explicitly demonstrated.

In the nucleus, estrogens bind to estrogen receptors, members of the nuclear receptor superfamily. Human cells have 2 estrogen receptors, ER alpha and ER beta, encoded by two genes (ESR1 and ESR2 respectively). Expression of the two genes varies by tissue: both are expressed in the central nervous system, the cardiovascular system, the urogenital tract and in the breast and bone; ER alpha expression predominates in the uterus, mammary gland, and liver, and the gastrointestinal tract expresses only ER beta (Pearce and Jordan, 2004; Gustafsson et al, 1999; Pfaffl et al, 2001; reviewed in Bai and Gust, 2009). The receptors show 47% identity overall and share a common organization consisting of 6 domains: an N-terminal A/B domain with ligand-independent activation function, a C domain containing the 2 DNA-binding zinc fingers, a hinge region (D) with a nuclear localization signal, an E domain that contains the ligand binding and dimerization domains as well as a ligand-dependent transactivation function, and a C-terminal F domain of poorly characterized function. The DNA-binding domain is the most highly conserved (97% identity) while the ligand-bindind domain is more variable (47% identity) (reviewed in Ruff et al, 2000; Bai and Gust, 2009). ER alpha and beta can homo- and heterodimerize, and recognize a common estrogen-response element due to their shared DNA-binding domains (reviewed in Bai and Gust, 2009). Functional studies suggest that ER alpha and beta have overlapping but distinct roles in estrogen-responsive transcription (Harrington et al, 2003; Katzenellenbogen and Katzenellenbogen, 2000; Pearce and Jordan, 2004; Pfaffl et al, 2001)
In the unliganded state, estrogen receptors are part of a multi-subunit complex containing HSP90, p23 (also known as PTGES3) and other chaperone-associated proteins (Joab et al, 1984; Segnitz et al, 1995; Knoblauch et al, 1999; Bouhouche-Chatelier et al, 2001; Fliss et al, 2000; Oxelmart et al, 2006; reviewed in Smith and Toft, 2008; Bai and Gust, 2009). This complex is part of a chaperone binding and release cycle shared by many nuclear receptors (described in more detail in the pathway "HSP90 chaperone cycle for steroid hormone receptors") that governs receptor folding and activity and may contribute to a high-affinity conformation of the ligand-binding domain (reviewed in Pratt and Toft, 1997; Smith and Toft, 2008). HSP90 release from the receptor complex requires ATP hydrolysis (Smith et al, 1993; Grenert et al, 1997; Panaretou et al, 1998; Obermann et al, 1998; Smith et al, 1992; reviewed in Smith and Toft, 2008).

In addition to nuclear signaling, estrogen receptors also promote a rapid, transcription- and translation-independent signaling response from the plasma membrane (reviewed in Levin, 2005; Schwartz et al, 2016). In contrast to their classical nuclear counterparts, estrogen receptors that signal from the plasma membrane are lipid-modified by palmitoylation. The amino-acid sequence encompassing an exposed cysteine residue (Cys447 for ESR1 and Cys399 for ESR2) is conserved between other steroid hormone receptors (Marino and Ascenzi, 2006). Mutation of the target cysteine in this motif abrogates membrane localization and the rapid response to estrogen (Acconcia et al, 2004; Acconcia et al, 2005; Pedram et al, 2007; Pedram et al, 2011). Golgi-localized palmitoyltransferases ZDHHC7 and ZDHHC21 catalyze the transfer of palmitoyl to the cysteine residue in the estrogen receptor (Pedram et al, 2012).

Estrogens mediate their transcriptional effects through interaction with the estrogen receptors, ESR1 (also known as ER alpha) and ESR2 (ER beta). ESR1 and ESR2 share overlapping but distinct functions, with ESR1 playing the primary role in transcriptional activation in most cell types (Hah and Krauss, 2014; Haldosén et al, 2014. The receptors function as ligand-dependent dimers and can activate target genes either through direct binding to an estrogen responsive element (ERE) in the target gene promoter, or indirectly through interaction with another DNA-binding protein such as RUNX1, SP1, AP1 or NF-kappa beta (reviewed in Bai and Gust, 2009; Hah and Krause, 2014). Binding of estrogen receptors to the DNA promotes the assembly of higher order transcriptional complexes containing methyltransferases, histone acetyltransferases and other transcriptional activators, which promote transcription by establishing active chromatin marks and by recruiting general transcription factors and RNA polymerase II. ESR1- and estrogen-dependent recruitment of up to hundreds of coregulators has been demonstrated by varied co-immunoprecipitation and proteomic approaches (Kittler et al, 2013; Mohammed et al, 2013; Foulds et al, 2013; Mohammed et al, 2015; Liu et al, 2014; reviewed in Magnani and Lupien, 2014; Arnal, 2017). In some circumstances, ligand-bound receptors can also promote the assembly of a repression complex at a target gene, and in some cases, heterodimers of ESR1 and ESR2 serve as repressors of ESR1-mediated target gene activation (reviewed in Hah and Kraus, 2014; Arnal et al, 2017). Phosphorylation of the estrogen receptor also modulates its activity, and provides cross-talk between nuclear estrogen-dependent signaling and non-genomic estrogen signaling from the plasma membrane (reviewed in Anbalagan and Rowan, 2015; Halodsèn et al, 2014; Schwartz et al, 2016)

A number of recent genome wide studies highlight the breadth of the transcriptional response to estrogen. The number of predicted estrogen-dependent target genes ranges from a couple of hundred (based on microarray studies) to upwards of 10000, based on ChIP-chip or ChIP-seq (Cheung and Kraus, 2010; Kinnis and Kraus, 2008; Lin et al, 2004; Welboren et al, 2009; Ikeda et al, 2015; Lin et al, 2007; Carroll et al, 2006). Many of these predicted sites may not represent transcriptionally productive binding events, however. A study examining ESR1 binding by ChIP-seq in 20 primary breast cancers identified a core of 484 ESR-binding events that were conserved in at least 75% of ER+ tumors, which may represent a more realistic estimate (Ross-Innes et al, 2012). These studies also highlight the long-range effect of estrogen receptor-binding, with distal enhancer or promoter elements regulating the expression of many target genes, often through looping or other higher order chromatin structures (Kittler et al, 2013; reviewed in Dietz and Carroll, 2008; Liu and Cheung, 2014; Magnani and Lupien, 2014). Transcription from a number of estrogen-responsive target genes also appears to be primed by the binding of pioneering transcription factors such as FOXA1, GATA3, PBX1 among others. These factors bind to heterochromatin by virtue of their winged helix domains and promote chromatin opening, allowing subsequent recruitment of other transcription factors (reviewed in Zaret and Carroll, 2011; Fiorito et al, 2013; Arnal et al, 2017; Magnani et al, 2011)