**2. Genomic actions of estrogens**

The genomic effects of steroid estrogens are mediated through two subtypes of nuclear and β-receptors (ER and ERβ). In mid 1980s. studies on the cloning of DNA encoding the steroid estrogen receptors were initiated (**Walter et al., 1986; Green, G.L. et al., 1986; Koike et al., 1987).** These studies led to the determination of their primary structure. Later they were assigned to ERα. In 1996, a number researchers discovered new members of the superfamily of nuclear estrogen receptors in rat prostate and ovaries (**Kuipper et al, 1996**) in human and mouse (**Mosselman et al., 1996**; **Tremblay et al., 1997)** and named it ERβ. Very soon the complete amino acid sequence of human ERβ (hERβ) was determinated (**Ogawa et al., 1998**). It was established that ERβ is significantly shorter than ERα. The comparison of amino acid sequences of these receptors and the investigation of affinity of various ligands to mutant forms of receptors allowed to establish that these receptors have 6 domains.

The N-terminal domains (A/B) have variable length and amino acid sequences. Usually, they exhibit a hormone-independent transactivation function (AF-1) that interacts with the elements of the transcription machinery and activate genes in target organs (**Tzukerman et al., 1994).** Domain C is responsible for the DNA recognition and receptor dimerization. This is a DNA-binding region, consisting approximately 50 amino acids. Domain E (HBD) contains approximately 250 amino acids and ensures the binding of hormones. This domain is the «hinge» - after the binding of ligand with receptors the following conformational rearrangement takes place with the participation of domain D. Clone 29 protein is highly homologous to rat ER, particularly in DNA-binding domain (95%) and ligand-binding domain (55%) (**Kuipper et al., 1996**). High degree of conservation of DNA-binding domain (96%) and ligand-binding domain (58%) was registered by other authors (**Mosselman et al., 1996).**

the treatment of oncological estrogen-sensitive diseases, cardio-vascular system, osteoporosis, neuroendocrinal diseases. The division of agents on such groups has relative character, because the activity of steroid estrogens has multifunctional character and one compound may be effectively used in various fields. This is a main reason why modified steroid estrogens have the advantage in comparison with huge number of heterocyclic

Obviously that most perspective search of such compounds is done on the basis of knowledge about estrogen action mechanism, particularity of its structure and metabolism. Several authors consider that main fast non genomic effects of estrogens are mediated by their membrane receptors (**Levin, 2002; Sak & Evaraus, 2004).** Experimental data about the identity of nuclear and membrane ERα in MCF7 cell are presented in the publication (**Pedram et al., 2009**). Obviously, ligand specificity may be significantly different. Mechanisms of membrane receptors action are still unclear, very often it is not proved that one or another effect is mediated by namely this group of receptors (**Warner & Gustafsson, 2006**). The absence of data about ligand specificity to membrane receptors does not allow to plan the synthesis of modified steroids with selective action, especially taking into account that their activity is realizing on many ways. During the consideration of osteoprotective action of estrogens and its influence on processes of cardio-vascular system we restrict ourselves to state the facts, because the decision concerning the synthesis of potential agents

The genomic effects of steroid estrogens are mediated through two subtypes of nuclear and β-receptors (ER and ERβ). In mid 1980s. studies on the cloning of DNA encoding the steroid estrogen receptors were initiated (**Walter et al., 1986; Green, G.L. et al., 1986; Koike et al., 1987).** These studies led to the determination of their primary structure. Later they were assigned to ERα. In 1996, a number researchers discovered new members of the superfamily of nuclear estrogen receptors in rat prostate and ovaries (**Kuipper et al, 1996**) in human and mouse (**Mosselman et al., 1996**; **Tremblay et al., 1997)** and named it ERβ. Very soon the complete amino acid sequence of human ERβ (hERβ) was determinated (**Ogawa et al., 1998**). It was established that ERβ is significantly shorter than ERα. The comparison of amino acid sequences of these receptors and the investigation of affinity of various ligands to mutant forms of receptors allowed to establish that these receptors have 6 domains. The N-terminal domains (A/B) have variable length and amino acid sequences. Usually, they exhibit a hormone-independent transactivation function (AF-1) that interacts with the elements of the transcription machinery and activate genes in target organs (**Tzukerman et al., 1994).** Domain C is responsible for the DNA recognition and receptor dimerization. This is a DNA-binding region, consisting approximately 50 amino acids. Domain E (HBD) contains approximately 250 amino acids and ensures the binding of hormones. This domain is the «hinge» - after the binding of ligand with receptors the following conformational rearrangement takes place with the participation of domain D. Clone 29 protein is highly homologous to rat ER, particularly in DNA-binding domain (95%) and ligand-binding domain (55%) (**Kuipper et al., 1996**). High degree of conservation of DNA-binding domain (96%) and ligand-binding domain (58%) was registered by other authors (**Mosselman et al.,** 

compounds, having more selective action.

**2. Genomic actions of estrogens** 

**1996).**

is usually reached on the analogy with known compounds.

Ligand-independent transactivational function AF1 is in zone А/В of both sub-types of receptors, zone E/F of -receptor of estrogens has additional ligand-independent activation function (**Tora et al., 1989**). ER has ligand-dependent trans-activation function AF-2.

Without ligand estrogen receptors are in complexes with heat shock proteins (Hsp90, Hsp70, Hsp56, Hsp60, Hsp48, Hsp23), this binding takes place in domain С. Probably, being in complexes with heat shock proteins estrogens are protected against the action of proteases. After binding its ligand conformational rearrangement of LBD of nuclear receptor and dissociation of complexes ER – heat shock proteins takes place. Except of the potential possibility to activate transcription, this rearrangement causes the change of topography of receptor regions, sensitive to proteolysis (**Ramsey & Klinge, 2001).** Phosphorylation of estrogen receptors is observed after their binding with hormone (**Kato et al., 1995)** that enhances the binding.

Transformed receptor forms dimers and in this form binds with DNA and may activate the transcription. Effective dimerization of ER requires a weak constitutive activity of sequences in domain C (**Kumar & Chambon, 1988).** It is necessary to note, that AF1 and AF2 exhibit relatively weak activity, whereas the maximum of transcription induction is observed when they act together (**Tora et al., 1989; Pham et al., 1992)**. From other side, in some cases the activation of AF-1 is enough to activate the transcription. Thus, 4 hydroxytamoxifen **4** is unable to induce AF-2 activity, but it is a strong agonist in cellular and promoter context where AF-1 is effective transcriptional activator (**Berry et al., 1990**).

Transcriptional process is modulated by receptors' ligands and various co-regulators **(Cheskis et al., 1997; An et al., 1999; 2001; Tcherepanova et al., 2000; Wong et al., 2001; Liu, J. et al., 2003; Bai & Gugière, 2003; Xu & Li, 2003)**. Conformation of ER changes upon interaction with coactivator proteins **(Tamrazi et al., 2005)**, as result the activity of one ligand may change in depending on cell nature. Thus, tamoxifen **5** and raloxifene **6** are agonists of estradiol in cardiovascular system, whereas they show antagonistic properties in breast and endometrium **(Jordan, 2007)**. Many others SERMs have similar properties.

Nowadays more 10 estrogen receptor-β isoforms are known, which have been identified starting from 2000 (**Lu et al., 2000).**

To study the roles of each receptor *in vivo*, a series of the mice were generated lacking either a functional ER and ERβ or both (ERKO, βERKO, βERKO) (**Emmel & Korach, 2001).** ERβ may modulate the functions of ER, if these receptors are in the same cells (**Matthews et al., 2006**). It became crucial during the diagnostics and the treatment of oncological

Approaches for Searching

of ERs (**Celic et al., 2007**).

correspondingly.

(**Zhu et al., 2006**).

**OH**

**H H H**

**OH**

**O**

**OH**

**H H H** **OH**

**OH**

of Modified Steroid Estrogen Analogues with Improved Biological Properties 175

**Chakravarti, 2003; Wolohan & Reichert, 2004; Pasha et al., 2005)**. In spite of the fact that such approaches have moderate predicted force, they do not provide the deeper understanding of estrogen action mechanism. From our point of view, using the multimodal approach, based on the application of methods of comparative molecular field analysis and

The knowledge about conformational mobility of different groups of potential ligands of ERs has great importance (**Kym et al., 1993; Grese et al., 1998; Selivanov & Shavva, 2002).** These data have interest in connection with data about modeling of conformations of ligandreceptor complexes (**Kraichely et al., 2000; Egner et al., 2001**) and conformational dynamic

Let's consider the examples of successful searching of modified estrogens which have

Structure analysis of LBD of these receptors shows that the nearest surrounding of bound estradiol in these receptors differs from each other on two amino acid residues: Leu384 and Met421 in ERα are substituted by Met336 and Ile373 in ERβ. Volumes of side chains of these amino acid residues are 85.9 Ǻ (Met), 82.6 Ǻ (Leu) and 82.3 Ǻ (Ile). It was predicted that the increased flexibility of the linear Met side chain would allow larger substituent to be accommodated. Met336 ERβ is situated closely to position 8β, and the introduction of small lipophilic substituent into this position must lead to the increasing of RBA of modified analogue to ERβ. Мet421 of ERα is at region of D-ring, and analogue with substituent at

Experiments have proved the stated predictions. As reference compounds steroids **10** and **12**  have been used. First one has selectivity hERα/hERβ is 20, and for second one this value hERβ/hERα is 22.5. Modified analogues **11** and **13** have the mentioned ratios as 70 and 180

Unfortunately, there is no correlation between RBA values for hERα and hERβ with values

The investigation of RBA for 71 compounds (mainly, metabolites of estrogens) to hERα and hERβ have shown that the differences in experimental values are not more then one order

**O**

**OH**

 **10 11 12 13** 

**OH**

**H H H** **OH**

**OH**

**O**

**H H**

**OH**

**O**

**OH**

**H H H** **OH**

binding energy calculations is most perspective (**Wolohan & Reichert, 2004).** 

selective binding affinity to hERα or hERβ (**Hillisch et al, 2004**).

positions 16α and/or 17α will have the increased affinity to ERα.

obtained in the experiments with the corresponding receptors of animals. In some cases the ratio between these values are opposite (**Hillisch et al., 2004**).

> **H H H**

> > **OH**

 **14 15 16** 

diseases (**Leygue et al., 1998; Pujol et al., 1998; Hall & McDonnel, 1999; Lazennec et al., 2001; Monroe et al., 2005)**. Compounds having the preferential binding to ERβ have great perspectives to be used for the treatment of autoimmune diseases, prostate disease, depression and ovulation disorders (**Gustafsson, 2005**). The noted above is the evidence for actuality of the search for modulators with preferable affinity to ER or ERβ.

Obviously, determination of complexes structure of estrogen receptors with various ligands with known biological properties was supposed to contribute to the development of models for mechanisms of estrogens action and understanding of the connection between structure and hormonal activity of ligands and thus to solve the abovementioned problem.

X-rays data of ligand-binding domain of ER with estradiol **1** and raloxifene **6** (**Brzozowski et al., 1997**), 4-hydroxytamoxifen **4** and diethylstilbestrole **7** (**Shiau et al., 1998**), and ligandbinding domain of ER with raloxifene **6** and genistein **8** (**Pike et al., 1999**) have been obtained. Estradiol **1** and diethylstilbestrol **7** are full agonists to ER and ER, 4 hydroxytamoxifen **4,** raloxifene **6**, and genistein **8** are agonists/antagonists. Interestingly, synthetic compounds of type **9** bind with ER as effectively as genistein **8** (**Miller et al., 2003**). Removal of ligands **1, 4**, **6**, **7, 8** from its complexes in crystal with LBD of ER and ERβ, docking of other potential ligands and the following optimization of its position in complex and binding energy by molecular modeling methods allow to evaluate the properties of new compounds independently from their belonging to steroid series. Some vagueness of evaluation of affinity to receptors is connected with the fact that the X-Ray data for complexes with different ligands are obtained only with LBD, but not with fulllength receptors.

The last of utmost importance, because during the realization of transcriptional action the formation of complexes between the DNA binding domain of estrogen receptors and estrogen response element (ERE) of duplex DNA is necessary, which requires the exact disposition of these elements of structure. Earlier structure of the DNA-binding domain of the ER has been investigated (**Schwabe et al., 1990**), however three-dimensional model of the DNA binding domain of ER was proposed significantly later (**Deegan et al., 2010**). On account of the above mentioned the differences in RBA values of model compounds to fulllength receptor and to its LBD could be significant. Thus, the affinity of (5S,11S)-5,11 diethyl-5,6,11,12-tetrahydrochrisene-2,8-diol **9** to full-length ERβ is in 10 times higher, than to its LBD (**Meyers et al., 1999**).

The results of quantum chemical calculations of ligand-receptor complexes allowed to synthesize a huge number of compounds having more then in 100-times higher effective binding to -ER in comparison to -ER (**Manas et al., 2004).** 

There are other ways for evaluation of properties of new compounds. For example some of them are traditional QSAR methods (**Gantchev et al., 1994; Tong et. al., 1997; Wiese et. al, 1997; Azzaoui et al., 1998; Gao et al., 1999; Wurtz et al., 1998; Sippl, 2002; Klopman &** 

diseases (**Leygue et al., 1998; Pujol et al., 1998; Hall & McDonnel, 1999; Lazennec et al., 2001; Monroe et al., 2005)**. Compounds having the preferential binding to ERβ have great perspectives to be used for the treatment of autoimmune diseases, prostate disease, depression and ovulation disorders (**Gustafsson, 2005**). The noted above is the evidence for

Obviously, determination of complexes structure of estrogen receptors with various ligands with known biological properties was supposed to contribute to the development of models for mechanisms of estrogens action and understanding of the connection between structure

**O**

**OH**

**OH**

**Et**

**Et**

**OH OH**

**O**

X-rays data of ligand-binding domain of ER with estradiol **1** and raloxifene **6** (**Brzozowski et al., 1997**), 4-hydroxytamoxifen **4** and diethylstilbestrole **7** (**Shiau et al., 1998**), and ligandbinding domain of ER with raloxifene **6** and genistein **8** (**Pike et al., 1999**) have been obtained. Estradiol **1** and diethylstilbestrol **7** are full agonists to ER and ER, 4 hydroxytamoxifen **4,** raloxifene **6**, and genistein **8** are agonists/antagonists. Interestingly, synthetic compounds of type **9** bind with ER as effectively as genistein **8** (**Miller et al., 2003**). Removal of ligands **1, 4**, **6**, **7, 8** from its complexes in crystal with LBD of ER and ERβ, docking of other potential ligands and the following optimization of its position in complex and binding energy by molecular modeling methods allow to evaluate the properties of new compounds independently from their belonging to steroid series. Some vagueness of evaluation of affinity to receptors is connected with the fact that the X-Ray data for complexes with different ligands are obtained only with LBD, but not with full-

The last of utmost importance, because during the realization of transcriptional action the formation of complexes between the DNA binding domain of estrogen receptors and estrogen response element (ERE) of duplex DNA is necessary, which requires the exact disposition of these elements of structure. Earlier structure of the DNA-binding domain of the ER has been investigated (**Schwabe et al., 1990**), however three-dimensional model of the DNA binding domain of ER was proposed significantly later (**Deegan et al., 2010**). On account of the above mentioned the differences in RBA values of model compounds to fulllength receptor and to its LBD could be significant. Thus, the affinity of (5S,11S)-5,11 diethyl-5,6,11,12-tetrahydrochrisene-2,8-diol **9** to full-length ERβ is in 10 times higher, than

The results of quantum chemical calculations of ligand-receptor complexes allowed to synthesize a huge number of compounds having more then in 100-times higher effective

There are other ways for evaluation of properties of new compounds. For example some of them are traditional QSAR methods (**Gantchev et al., 1994; Tong et. al., 1997; Wiese et. al, 1997; Azzaoui et al., 1998; Gao et al., 1999; Wurtz et al., 1998; Sippl, 2002; Klopman &** 

actuality of the search for modulators with preferable affinity to ER or ERβ.

and hormonal activity of ligands and thus to solve the abovementioned problem.

**OH**

**OH**

 **7 8 9** 

**OH**

length receptors.

to its LBD (**Meyers et al., 1999**).

binding to -ER in comparison to -ER (**Manas et al., 2004).** 

**Chakravarti, 2003; Wolohan & Reichert, 2004; Pasha et al., 2005)**. In spite of the fact that such approaches have moderate predicted force, they do not provide the deeper understanding of estrogen action mechanism. From our point of view, using the multimodal approach, based on the application of methods of comparative molecular field analysis and binding energy calculations is most perspective (**Wolohan & Reichert, 2004).** 

The knowledge about conformational mobility of different groups of potential ligands of ERs has great importance (**Kym et al., 1993; Grese et al., 1998; Selivanov & Shavva, 2002).** These data have interest in connection with data about modeling of conformations of ligandreceptor complexes (**Kraichely et al., 2000; Egner et al., 2001**) and conformational dynamic of ERs (**Celic et al., 2007**).

Let's consider the examples of successful searching of modified estrogens which have selective binding affinity to hERα or hERβ (**Hillisch et al, 2004**).

Structure analysis of LBD of these receptors shows that the nearest surrounding of bound estradiol in these receptors differs from each other on two amino acid residues: Leu384 and Met421 in ERα are substituted by Met336 and Ile373 in ERβ. Volumes of side chains of these amino acid residues are 85.9 Ǻ (Met), 82.6 Ǻ (Leu) and 82.3 Ǻ (Ile). It was predicted that the increased flexibility of the linear Met side chain would allow larger substituent to be accommodated. Met336 ERβ is situated closely to position 8β, and the introduction of small lipophilic substituent into this position must lead to the increasing of RBA of modified analogue to ERβ. Мet421 of ERα is at region of D-ring, and analogue with substituent at positions 16α and/or 17α will have the increased affinity to ERα.

Experiments have proved the stated predictions. As reference compounds steroids **10** and **12**  have been used. First one has selectivity hERα/hERβ is 20, and for second one this value hERβ/hERα is 22.5. Modified analogues **11** and **13** have the mentioned ratios as 70 and 180 correspondingly.

Unfortunately, there is no correlation between RBA values for hERα and hERβ with values obtained in the experiments with the corresponding receptors of animals.

In some cases the ratio between these values are opposite (**Hillisch et al., 2004**).

The investigation of RBA for 71 compounds (mainly, metabolites of estrogens) to hERα and hERβ have shown that the differences in experimental values are not more then one order (**Zhu et al., 2006**).

Approaches for Searching

**3. Carcinogenicity causes of ER ligands** 

of Modified Steroid Estrogen Analogues with Improved Biological Properties 177

The establishment of at least major reasons of carcinogenicity of estrogens, their agonists/antagonists has significant importance, since this understandings lies in the

At the present time two main types of carcinogenesis are known - promotive and genotoxic (mutagenic), action of them may add up. Existence of the fist one is confirmed by wellknown facts of tumor formation under the estrogen action (**Russo et al., 2002**) and promotion of tumor induction under the action of various agents, for example, nitrosobutylurea (**Sumi et al., 1984**). The same results have been obtained in other models. For example, 17α-ethynylestradiol **24** enhances the tumor formation in rat males under the action of diethylnitrosamine (**Yoshida & Fukunishi, 1981).** Estradiol increased dysplasia of

The explanation of this negative influence of estrogens is quite clear – when in active state

**OH**

**O**

**+ OH**

**2-OHE OH 4-OHE**

**H H H**

**O**

**2 26 27**

**+**

**O**

**<sup>O</sup> 2-OHE-o-QUINONE 4-OHE-o-QUINONE**

**28 29**

**H H H**

Lately using 2-fluoroestradiol **25** as an example it was shown that there is no correlation between carcinogenicity and hormonal action of estrogens (**Liehr, 1983**). It is important that 17-ethynylestradiol **24** possesses high hormonal activity and lowered carcinogenicity (**Li, J.J. et al., 1995**). These investigations (and earlier ones) became the basis for searching of

**F**

**H H H** **OH**

**O**

**H H H**

**O**

**H H H**

fundament of the creation of medications with the improved properties.

breast on androgenized rats, induced by 7,12-dimethylbenz[a]anthracene.

**OH**

**H H H**

 **24 25** 

**P450**

**OH**

**O**

**O**

**OH**

the cells are vulnerable to attacks by reactive compounds.

**OH**

other mechanisms of estrogens' carcinogenicity.

**O**

**H H H**

**estrone**

**OH**

The introduction of relatively small linear substituents (4-5 carbon atoms) at position 11 of steroid skeleton may lead to the appearance of antagonistic activity to ER at agonistic properties to ER (steroids **14** and **15**), whereas analogue **16** is agonist to both receptors (**Loosen, 1999**).

Steroids with small substituents at position 11 without aromatic ring possess significantly higher affinity and transcriptional activity of ER (**Loosen еt al., 2000**), for example, analogues **17**-**19** have been found**.**

Finally, there are compounds, which fully block the binding of estradiol **1** with known ERs, for example steroids **20-23** (**Jordan, 2003**). A number of compounds with large substituents at С-7 () and С-11 () have been synthesized, such analogues are of great interest for the treatment of estrogen-dependent oncological diseases and hormones imbalance, caused by the increased formation of estrogens in the body.
