It is suggested that estrogen hormones recruit mechanisms controlling histone acetylation to bring about their effects in the uterus. However, it is not known how the level of histone acetylation affects estrogen-dependent processes in the uterus, especially proliferation and morphogenetic changes. Therefore, this study examined the effects of histone deacetylase blockers, trichostatin A and sodium butyrate, on proliferative and morphogenetic reactions in the uterus under long-term estrogen treatment. Ovari-ectomized mice were treated with estradiol dipropionate (4 μg per 100 g; s.c., once a week) or vehicle and trichostatin A (0.008 mg per 100 g; s.c., once a day) or sodium butyrate (1% in drinking water), or with no additional treatments for a month. In animals treated with estradiol and trichostatin A or sodium butyrate, uterine mass was increased, and abnormal uterine glands and atypical endometrial hyperplasia were found more often. Both histone deacetylase inhibitors produced an increase in the numbers of mitotic and bromodeoxyuridine-labelled cells in luminal and glandular epithelia, in stromal and myometrial cells. Levels of estrogen receptor-α and progesterone receptors in uterine epithelia, stromal and myometrial cells were decreased in mice treated with estradiol and trichostatin A or sodium butyrate. Expression of β-catenin in luminal and glandular epithelia was attenuated in mice treated with estradiol with trichostatin A or sodium butyrate. Both histone deacetylase inhibitors have similar unilateral effects; however the action of trichostatin A was more expressed than that of sodium butyrate. Thus, histone deacetylase inhibitors exert proliferative and morphogenetic effects of estradiol. The effects of trichostatin A and sodium butyrate are associated with changes in expression of estrogen receptor-α, progesterone receptors and β-catenin in the uterus.
Estrogen hormones have a variety of effects in the uterus. Proliferation and changes in structure and architecture of uterine tissues are estrogen-dependent events (Martin et al. 1973, Bigsby 2002). Endometrial cancer is also an estrogen-dependent disease, and it can be regarded as a consequence of estrogen-induced alteration in proliferation and morphogenesis (Emons et al. 2000, Archer 2004). Endometrial hyperplasia and cancer can easily be induced in laboratory rodents by continuous estrogen exposure (Akhmedkhanov et al. 2001). Administration of estrogen hormones can lead to endometrial cancer formation in women (Deligdisch 2000). Numerous cases of endometrial cancer are registered each year in every country (Archer 2004). Therefore, regulation of estrogen action and the interactions between estrogens and target tissues must be investigated more intensively in order to achieve more effective prevention and treatment of estrogen-dependent pathology.
It is known that the action of estrogen is regulated by a variety of extracellular factors – such as hormones, growth factors and others (Couse & Korach 1999, Gunin et al. 2001). Recent data showed that estrogen also recruits intracellular regulatory systems to bring about its action (Couse & Korach 1999, Deroo et al. 2004, Gunin et al. 2004b). The scientific literature gives some direct and indirect evidence that the system that maintains the status of histone acetylation interacts with estrogen signalling (Alao et al. 2004, Kurtev et al. 2004, Margueron et al. 2004).
Several types of histone deacetylases are present in the nucleus of cells, and together with histone acetyltrans-ferases they determine the acetylation status of histone proteins (Margueron et al. 2004). Shifts in the level of histone acetylation are accompanied by changes in the activity of the transcription process (Riester et al. 2004). Histone deacetylase activity keeps chromatin in a transcriptionally inactive state (Riester et al. 2004). Data have been published showing interactions between estrogen signalling and acetylation of histones (Jang et al. 2004, Margueron et al. 2004). It has also been reported that estrogen hormones affect the level of histone acetylation in target tissues (Sun et al. 2001). There is some evidence that histone acetylation is involved in processes that are also regulated by estrogens, for example proliferation and cell differentiation (Sakai et al. 2003). It is therefore supposed that shifts in histone acetylation can affect the action of estrogen on the uterus, such as proliferation and morphogenetic alterations. However, it is not known how the histone acetylation status influences estrogen-dependent processes in the uterus, such as proliferation and changes in structure of tissues. Therefore, the aim of this research was to examine estrogen-induced processes in the uterus and their response to substances that change the level of histone acetylation.
Trichostatin A, an antifungal antibiotic, has a potent and specific inhibitory effect on histone deacetylase activity (Riester et al. 2004). Butyric acid and its salts are also well-known blockers of the activity of histone deacetylases (Riester et al. 2004). Therefore two reagents, trichostatin A and sodium butyrate, were chosen for use in our experiments to block histone deacetylases; this was followed by induction of a hyperacetylated histone state.
Material and Methods
All procedures were performed in accordance with the UFAW Handbook on the Care and Management of Laboratory Animals and with the Chuvash State University Rules for work with laboratory animals. White out-bred CFW female mice (19.1±0.3 g, mean±s.e.m.) were used. Animals were obtained from the Animal Department of Chuvash State University (Cheboksary, Russia) and were housed with free access to water and food. Mice were ovariectomized 2 weeks before the experiments were started. All surgical procedures were performed under anesthesia with ketamine and diazepam (75 and 0.12 mg/kg respectively, i.p.; Gedeon-Richter, Budapest, Hungary).
Ovariectomized mice were divided into several groups according to the treatments, as follows. The first group (n=15) was treated with s.c. injections of estradiol dipropionate in olive oil (Minmedprom, Rostov-Don, Russia) at a dose of 4 μg per 100 g of body mass once a week and received s.c. injections of saline (0.15 M sodium chloride) at a dose of 0.1 ml per mouse once a day for 30 days. The second group (n=15) was treated with s.c. injections of estradiol once a week and received s.c. injections of trichostatin A (Sigma) at a dose 0.008 mg per 100 g once a day for 30 days. Trichostatin A was dissolved in 0.15 M sodium chloride. The third group of mice (n=15) was treated with injections with estradiol once a week and allowed to drink tap water with 1% (w/v) sodium butyrate (Sigma) for 30 days. These groups also received s.c. injections of saline (0.1 ml per mouse) once a day for 30 days.
The fourth (n=5), fifth (n=5) and sixth (n=5) groups received s.c. injections of the vehicle of estradiol (olive oil; 0.1 ml per mouse) once a week and saline or trichostatin A or sodium butyrate, respectively, once a day for 30 days.
Our previous observations clearly showed that treatment with estradiol for 30 days was quite enough to produce expressed estrogen-dependent changes in uterine morphology and to induce hyperplastic changes in the uterus (Gunin et al. 2001, 2002, 2004a,Gunin et al.b). This estradiol dose and treatment regime produced estradiol levels in the blood that were close to the normal physiologic values (Gunin et al. 2004a).
The uteri were removed 48 h after the last estradiol or vehicle injection. All animals were injected i.p. with bromodeoxyuridine (BrdU; 5 mg per 100 g of body mass; Sigma) dissolved in 0.15 M sodium chloride 2 h before the tissues were removed. Organs were removed under deep ether anesthesia. Uteri were weighed and middle segments of uterine horns were then placed in modified Bouin’s fixative (Gunin et al. 2000) for 6 h at room temperature, and were then dehydrated and embedded in paraffin. Uteri were transversely oriented and cut at 5–7 μm.
Histological changes in the uterus were analyzed and diagnosed according to Scully et al.(1994). To estimate the extent of any hyperplastic or neoplastic changes in the endometrium, uterine glands were subdivided into four morphological types: (1) normal glands (simple tubular glands which can appear in section as round, oval, or elongated with a narrow lumen; this type has no branches or daughter glands); (2) cystic glands (round-shaped glands of more than average or large size); (3) glands with daughter glands (these glands have various shapes (round, elongate, tortuous) and sizes and have forming or formed daughter gland or glands inside the epithelium or inside the mother gland lumen, or on the outer surface of the mother gland); (4) conglomerate of glands (this type has a very complex architecture in which individual glands are closely disposed to each other almost without intervening stroma and have multiple interconnecting lumens – this type may develop from glands with daughter glands), as described in Gunin et al. 2001). The numbers of each type of gland were calculated in randomly selected sections. No less than three sections from each animal were examined. Results were expressed as the percentage of each type of gland. The epithelium of all glands in randomly selected sections was examined and typed as simple, pseudostratified or stratified (multilayered) epithelia. The percentage of glands with each type of epithelium was calculated.
The number of mitotic and BrdU-labelled cells
Proliferative processes were assessed from the number of mitotic and BrdU-labelled cells as described previously (Gunin et al. 2001). Mitoses were counted in sections stained with iron haematoxylin. BrdU was detected immunohistochemically. Anti-BrdU mouse monoclonal antibody conjugated with biotin (catalog number, MO 5215; Caltag Laboratories, Burlingame, CA, USA) diluted 1:50 in Tris-buffered saline (TBS) (pH 7.2–7.6) was used as the primary antibody. Streptavidin conjugated with alkaline phosphatase (catalog number, SA 1008; Caltag Laboratories) diluted 1:50 in TBS with 0.1% (v/v) Triton X-100 was then applied. Alkaline phosphatase activity was revealed through the use of naphtol AS-BI-phosphate and new fuchsin as chromogens. Control sections were stained in a similar manner, except the primary antibody was replaced with normal mouse serum. All results were expressed as the percentage of mitotic or labelled cells.
Estrogen receptor-α, progesterone receptors and β-catenin
Estrogen receptor-α, progesterone receptors and β-catenin were detected using routine indirect immunohistochemical staining. Rabbit anti-estrogen receptor-α polyclonal antibody (catalog number, sc-542; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) diluted 1:50, rabbit anti-progesterone receptors antiserum (catalog number, sc-538; Santa Cruz Biotechnology Inc.) diluted 1:50 and rabbit anti-β-catenin antiserum (catalog number, C2206; Sigma) diluted 1:50 were used as primary antibodies. For detection of estrogen receptors, goat anti-rabbit immunoglobulin G antibody conjugated with alkaline phosphatase (catalog number 111–055–045; Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA) was used as secondary antibody, and alkaline phosphatase activity was then revealed using naphtol AS-BI-phosphate and new fuchsin as chromogens. For progesterone receptors and β-catenin, goat anti-rabbit immunoglobulin G antibody conjugated with peroxidase (catalog number 111–035–045; Jackson ImmunoResearch Laboratories Inc.) was used as the secondary antibody, peroxidase activity was then developed by the use of hydrogen peroxide and diaminobenzidine (Sigma) techniques, slides were also preincubated in 0.1% hydrogen peroxide in distilled water for 10 min to block endogenous peroxidase activity. Control sections were stained in a similar manner, except the primary antibody was replaced with normal rabbit serum. To avoid possible differences in the intensity of staining, sections from all mice were processed simultaneously for each antigen, so that all sections were incubated in exactly the same TBS, the same mixtures of primary and secondary antibodies, the same mixture for development of enzyme activity, for the same times, at the same temperature.
Intensity of immunostaining was quantified by photo-metric measurement of optical density (D) for positive-stained components of a tissue (Gunin et al. 2002, 2004a). The photometric procedure was performed using Sigma Scan Pro software (SPSS Inc., Chicago, IL, USA). Initially, uterine sections immunohistochemically stained for estrogen or progesterone receptors were photographed using an Olympus light microscope (objective magnification 40), an Olympus C3040-ADU camera adapter and an Olympus Camedia 4040z 4 megapixel digital camera. At least three randomly selected sections were photographed for each mouse, and three to five randomly selected fields were photographed from one section. Photographs were then loaded in Sigma Scan Pro software and analyzed to find the optical density. It was performed by measuring the intensity of staining of positive-stained structures (F) and structures with no staining (F0). The intensify of staining was measured from equal areas of tissue image (19 pixels). Positive staining for estrogen and progesterone receptors was detected in the nuclei of all uterine tissues (luminal epithelium, glandular epithelium, stromal cells, myometrial cells). Therefore, the intensity of nuclear staining (F; positive staining) and the intensity of staining of the internuclear space in the endometrial stroma (F0; negative staining) were measured for estrogen and progesterone receptors.
β-Catenin was detected in luminal and glandular epithelia. Therefore, the intensity of staining of cytoplasm of these epithelial cells (F; positive staining) and intensity of staining of the internuclear space in the endometrial stroma (F0; negative staining) was measured for β-catenin. Optical density (light absorption) was calculated from the formula D=l g(F0/F). The level of expression was considered as the value of optical density (Gunin et al. 2002, 2004a). At least 100 nuclei were analyzed for each structure in each mouse.
Arithmetic means and standards errors were calculated for each data group. The significance of differences was determined by Student’s t-test (uterine mass, proliferation, estrogen and progesterone receptors, β-catenin) and by the use of the observed versus expected χ2 test (gland types, epithelium types, pathology). Values of P<0.05 were considered significant.
The uterine mass of ovariectomized mice receiving olive oil instead of estradiol and saline for 30 days (group 4) was 115.2±22.4 mg per 100 g body mass (mean±s.e.m.; n=5); the addition of trichostatin A (group 5; 120.6±24.1 mg per 100 g body mass; n=5) or sodium butyrate (group 6; 116.8±19.7 mg per 100 g body mass; n=5) for 30 days had no effect on uterine mass.
Treatment with estradiol and trichostatin A for 30 days (group 2) resulted in a 34% increase in uterine mass as compared with data for control mice receiving estradiol alone (group 1). Administration of estradiol and sodium butyrate (group 3) also produced a 17% increase in uterine mass (Fig. 1).
All uteri of ovariectomized mice, which were not treated with estradiol and received trichostatin A (group 5) or sodium butyrate (group 6) or no additional treatments (group 4) for 30 days, were diagnosed with atrophic endometrium (Fig. 2). All endometrial glands in all these uteri had a narrow lumen, and had a round, oval, or elongated shape (a microscopical reflection of simple tubular glands), which were regarded as normal. All glands were lined with simple cuboidal epithelium.
In control mice receiving estradiol for 30 days (group 1), abnormal glands, especially glands with daughter glands and glands forming conglomerates, were observed (Figs 2 and 3). Glands lined with pseudostratified or atypical stratified epithelium were also found in these uteri (Figs 2 and 3). Atypical endometrial hyperplasia was diagnosed in 36.5% of control mice treated with estradiol for a month.
In animals treated with estradiol and trichostatin A for 30 days (group 2), glands with daughter glands, glands forming conglomerates and glands with atypical stratified columnar epithelium were observed more often (Figs 2 and 3). Atypical endometrial hyperplasia was found in 78.6% of cases, and there were no cases of simple endometrial hyperplasia. Normal proliferative endometrium was also not diagnosed in mice receiving estradiol and trichostatin A (Figs 2 and 3).
In the uteri of mice treated with estradiol and sodium butyrate for 30 days (group 3), glands with daughter glands, conglomerates of glands and glands with atypical epithelium were also found in a greater percentage of cases (Figs 2 and 3). Atypical endometrial hyperplasia was found in 64.3% of cases (Figs 2 and 3).
Proliferation in the uterus was estimated by two parameters: the number of mitotic cells and the number of BrdU-labelled cells. Treatment with estradiol and trichostatin A for 30 days (group 2) led to an increase in the percentage of mitotic and BrdU-labelled cells in all uterine tissues (Fig. 4). Treatment with estradiol and sodium butyrate for 30 days (group 3) also produced an increase in the number of mitotic and BrdU-labelled cells in all structures (Fig. 4). Mice that received the estradiol vehicle (olive oil) with no additional treatments (group 4) or with trichostatin A (group 5) or sodium butyrate (group 6) for 30 days had no changes in proliferative parameters in all uterine tissues (Fig. 4b and d).
Using immunohistochemistry, estrogen receptor-α was found in luminal and glandular epithelia, in stromal and myometrial cells of the uterus. Treatment with estradiol and trichostatin A for 30 days (group 2) reduced the level of estrogen receptor-α in all uterine compartments, as compared with that in control animals (group 1) treated with estradiol (Figs 2 and 5). Treatment with estradiol and sodium butyrate also led to a reduction in estrogen receptor levels in all uterine tissues.
Mice that received the estradiol vehicle (olive oil) and trichostatin A or sodium butyrate, or with no additional treatments for 30 days (groups 4 to 6) had no changes in estrogen receptor expression (Fig. 5).
Using immunohistochemistry, progesterone receptors were detected in luminal and glandular epithelia, stromal and myometrial cells of the uteri of mice in all treatments group. Treatment with estradiol and trichostatin A for 30 days (group 2) led to marked reduction in the level of progesterone receptors in all uterine compartments, as compared with that in control animals (group 1) treated with estradiol (Figs 2 and 6). Treatment with estradiol and sodium butyrate (group 3) also resulted in a marked decrease in progesterone receptor expression in all uterine tissues.
The data from mice treated with estradiol vehicle only or with trichostatin A, or with sodium butyrate for a month (groups 4 to 6) are shown in Fig. 6b. There are no differences in progesterone receptor levels among these groups.
Immunohistochemical staining for β-catenin showed that this protein was clearly detected in luminal and glandular epithelia of the uteri of mice in all treatment groups. Treatment with estradiol and trichostatin A (group 2) or with sodium butyrate (group 3) for 30 days led to a reduction in the level of β-catenin in both epithelia (Figs 2 and 7). The effect of trichostatin A was more expressed.
The data from mice treated with estradiol vehicle only or with trichostatin A, or with sodium butyrate for a month (groups 4 to 6) are shown in Fig. 7b. There are no differences in β-catenin level among these groups.
A group of parameters was employed to estimate estradiol action in the uterus. Uterine mass is a well-known indicator of estrogen action (Emons et al. 2000, Bigsby 2002). Proliferation, which was determined by the numbers of mitotic and BrdU-labelled cells, also depends on estrogen action (Martin et al. 1973, Gunin et al. 2000). In addition, morphogenetic alterations – such as shape of glands, type of glandular epithelium and pathology findings – appear in the uterus under long-term estrogen action (Martin et al. 1973, Gunin et al. 2002). Moreover, hyperplastic changes in the endometrium are often found. Some hyperplastic changes have nonfavorable prognosis, especially complex and atypical hyperplasies. Complex hyperplasia is characterized by architectural disarray in gland shape and glandular epithelium. Glands may have multiple lumens which interconnect. Glandular epithelium may be tall and columnar or pseudostratified. However, epithelial cells in general do retain their orientation to the lumen. Atypical hyperplasia has approximately the same characteristics, but hyperplastic endometrium shows cytologic atypia including large nuclei of variable size and shape that have lost polarity. Of all endometrial hyperplasies, atypical hyperplasia has the most increased risk for progression to endometrial cancer (Scully et al. 1994, Deligdisch 2000).
Results clearly showed that both histone deacetylase blockers had unilateral effects and led to an increase in uterine mass and proliferation and to more expressed morphogenetic alterations than in control mice. The effect of trichostatin A is more expressed than that of butyrate. This situation is probably associated with the fact that trichostatin A has a more specific and a stronger action in blocking histone deacetylase activity (Riester et al. 2004). For all parameters tested, the effects of histone deacetylase inhibitors were found only in estrogen-treated mice and were not documented in control animals receiving olive oil instead of estradiol. Hence, these results support the supposition that histone deacetylase inhibitors affect some steps in the mechanism of estrogen action.
In our previous experiments, treatments sometimes had an effect on proliferation but had no action on morpho-genetic changes in the uterus, and vice versa (Gunin et al. 2001, 2004a,Gunin et al.b). The present results show that histone deacetylase blockers affect both proliferation and morphogenetic alterations. It is known that proliferation and control of cell shape, differentiation, adhesion and apoptosis are regulated by different mechanisms and depend on the work of different genes (Bigsby 2002, Klotz et al. 2002). Hence, histone deacetylases are involved in the control of all these estrogen-dependent processes, and inhibition of these deacetylases intensifies proliferative and morphogenetic estrogen actions. One more well-known estrogen effect in the uterus is the induction of progesterone receptors (Couse & Korach 1999). Our current results showed that both histone deacetylase blockers led to a decrease in progesterone receptor expression in all uterine tissues. Hence, it can be concluded that histone deacetylases also manage this part of estrogen action, but their blockade attenuates the estrogen effect on progesterone receptor expression.
To define some possible mechanisms involved in the action of histone deacetylases on estrogen-induced effects, the expression of estrogen receptor-α and progesterone receptors in uterine tissues was examined. Results showed that both blockers reduced the level of estrogen receptor-α in all uterine compartments, as compared with control. Other data support our observation and also show that histone deacetylase blockers decrease estrogen receptor-α levels in ovarian, endometrial and mammary gland cancer cell lines (Alao et al. 2004, Margueron et al. 2004).
In general, the level of receptors in a tissue depends on a balance between their synthesis and degradation (Ing & Ott 1999, Nephew et al. 2000). Histone deacetylase blockers probably attenuate estrogen receptor synthesis or activate their degradation. The effect of histone deacetylase blockers on progesterone receptor levels may also be caused by a decrease in the speed of their synthesis or by acceleration of their degradation. Other researchers also reported that trichostatin A decreased the levels of progesterone receptor coactivators and impaired progesterone receptor function (Wilson et al. 2002, Condon et al. 2003).
It is interesting to note that in mice treated with histone deacetylase blockers and estradiol, the more intensive estrogen-dependent processes (increase in mass, proliferation, morphogenesis) in the uterus proceed with lower levels of estrogen and progesterone receptors. Other data also showed that more malignant and less-differentiated endometrial tumors had low levels of estrogen and progesterone receptors (Sivridis et al. 2001, Ali et al. 2004). It is possible that a diminished level of estrogen and progesterone receptors does not allow estrogens to adequately control the processes managing the morphogenesis that leads to atypical hyperplasia formation.
β-Catenin is implicated in cell adhesion and is a component of the Wnt-pathway (Cong et al. 2003). β-Catenin provides intercellular adhesion and it is possible that if its concentration is high, cell–cell connection is more stable and that this protects from the formation of precancerous changes. In the case of low β-catenin concentration, cell–cell interactions are less solid, which provides a foundation for cancer development. It has been reported that the level of β-catenin in the uterus was decreased following estrogen action and cancer formation (Fujimoto et al. 1998, Nei et al. 1999, Miyamoto et al. 2000, Gunin et al. 2004b). There is a decrease in β-catenin expression in uterine epithelia in mice treated with estradiol and histone deacetylase blockers compared with that of mice treated with estradiol alone. Our results showed that more expressed morphogenetic shifts, which were found in estradiol and trichostatin A or sodium butyrate treated mice, are accompanied by decreased levels of β-catenin. Hence, β-catenin is involved in changes in estrogen-dependent uterine morphology which are affected by histone deacetylase blockers.
β-Catenin content in a tissue is also a result of the balance between its synthesis and degradation. It has previously been shown that changes in β-catenin expression can be caused by the work of glycogen-synthase kinase-3β, an enzyme which takes part in β-catenin degradation (Gunin et al. 2003). Other data showed that estrogen hormones can attenuate β-catenin biosynthesis in the uterus (Fujimoto et al. 1996). Direct interactions between β-catenin and histone deacetylases were documented (Billin et al. 2000). It was also shown that β-catenin led to a loss of activity of histone deacetylases (Henderson et al. 2002, Baek et al. 2003). A function of β-catenin is compromised by fusion to a transcriptional repressor domain from histone deacetylase (Cong et al. 2003). However, further studies are needed to elucidate the roles of histone deacetylases in the regulation of β-catenin content in the uterus, as well as to define the role of β-catenin in uterine morphogenesis.
There are reports showing that histone deacetylase blockers – trichostatin A, sodium butyrate and others – led to a decrease in proliferation in endometrial and mammary gland cancer cell lines (Takai et al. 2004a,b). Our results showed that trichostatin A or sodium butyrate enhanced proliferation and hyperplasia formation in the murine uterus. What are the sources of these contradictions? First, most of the studies examining the effects of changes in the level of histone acetylation on proliferation and differentiation in uterine and mammary gland cell lines were performed in the absence of estrogen administration (Adhikari et al. 2000, Terao et al. 2001). Secondly, most works that showed the antiproliferative and anticancer potency of histone deacetylase blockers were performed in vitro. However, in vitro and in vivo conditions are principally different. Concentrations of reagents are constant in vitro, but in vivo, concentrations of acting substances are subjected to fluctuations depending on the activity of metabolic processes. There is a report showing a rapid increase in proliferation of human endometrial adenocarcinoma cells after removal of sodium butyrate from a medium (Saito et al. 1991). Other works demonstrated that in large intestine sodium butyrate has an antineoplastic effect when used in vitro (Deschner et al. 1990, Lupton 2004) and accelerates cancer formation in vivo (Freeman 1986, Lupton 2004). There are, however, observations documenting an increase in colon cancer formation under sodium butyrate treatment in vitro and in vivo (Deschner et al. 1990). Finally, all published data were obtained on already formed cancer cells, but not on normal cells that are transforming to cancerous cells. Our experiments were done on normal, nonmalignant, uterine tissues which were then exposed to estrogen followed by hyperplasia formation. One more interesting remark, valproic acid, which is used for treatment of epilepsy and has a mechanism of action which involves the blockade of histone deacetylases, revealed expressed teratogenic effects in humans (Phiel et al. 2001, Kultima et al. 2004). Trichostatin A also has teratogenic effects in vertebrate embryos (Phiel et al. 2001). However, further studies are needed to explain some contradictory points and to elucidate the exact mechanisms involved in the interactions between estrogen signalling and histone deacetylases.
Thus, this research provides evidence that histone deacetylase blockers, trichostatin A and sodium butyrate, enhance proliferative and morphogenetic estrogen action and support the development of estrogen-dependent endometrial hyperplasia. However, the exact mechanisms of the actions of histone deacetylase blockers on estrogen-dependent changes in uterine tissues remain unclear. It is possible that histone deacetylase inhibitors affect the activity of estrogen-regulated genes in the uterus and this is an important avenue for further research. We hope that this research will lead to a better understanding of the origin and progression of estrogen-dependent cancer of the female reproductive system.
This work was supported by grants from the Russian Foundation for Basic Research (03–04–48000) and the Ministry of Education and Science of Russia (E02–6.0–136; yp. 11.01.026). The authors declare that there is no conflict of interest that would prejudice the impartiality of this scientific work.
ScullyRE1994 Histological typing of female genital tract tumours. In International Histological Classification of Tumours