Tumorigenic effects of tamoxifen on the female genital tract.

Tamoxifen is a nonsteroidal triphenylethyl compound that belongs to a class of selective estrogen receptor modulators (SERMs) ( Fig. 1 ), binds to estrogen receptors (ERs) and elicits estrogen agonist or antagonist responses, depending on the target tissue ( Fig. 2 ) ( MacGregor and Jordan, 1998 ; Pasqualini et al. 1998 ). Its estrogen antagonist properties have made tamoxifen an important treatment modality for patients with breast cancer, especially those whose tumors are positive for ERs. At present, tamoxifen is one of the most widely prescribed drugs in the world ( ACOG Committee Opinion, 2006 ). Tamoxifen was approved in 1977 by the US Food and Drug Administration for the treatment of metastatic breast cancer in postmenopausal patients. Tamoxifen was also found to suppress ER-positive breast cancer in postmenopausal women, to increase the disease-free interval, and to reduce the incidence of contralateral breast cancer in breast cancer patients ( Early Breast Cancer Trialists’ Collaborative Group, 1992 ). It is also effective in premenopausal breast cancer patients ( Osborne, 1998 ; Clarke, 2006 ). Early treatment for breast cancer metastasis has been found to delay disease progression ( Nicolini et al. 1997 ). Currently, tamoxifen is an option along with aromatase inhibitors as the endocrine treatment of choice in all stages of breast cancer in both pre- and postmenopausal women ( Osborne, 1998 ; Clarke, 2006 ). Fisher and colleagues, (1989) recommended that all low-risk patients with breast cancer receive adjuvant therapy, and a large number of premenopausal and postmenopausal women have been placed on 3- to 5-year regimens of tamoxifen. In addition, prophylactic use of tamoxifen resulted in a 45% reduction in the incidence of ER-positive breast cancer in healthy women, according to the Breast Cancer Prevention Trial of the National Surgical Adjuvant Breast and Bowel Project ( Fisher et al. 1998 ). Therefore, in 1998 the US Food and Drug Administration approved the use of tamoxifen as a chemopreventive option in both premenopausal and postmenopausal women. Efficacy of tamoxifen in breast cancer is due to its anti-estrogen properties, but it may also exert a weak estrogenic effect resulting in a variety of lesions in the female genital tract. Many ex vivo and in vivo studies have shown quite strong estrogenic-like activity in the endometrium, vagina, liver, and bones ( Patterson et al. 1982 ; Wolf and Jordan, 1992 ). Tamoxifen blocks ERs in the hypothalamus, leading to the inhibition of estrogen feedback, which leads to increases in the production of gonadotropin-releasing hormone (GnRH), follicle-stimulating hormone (FSH), and luteinizing hormone (LH). Consequently, the ovaries may become hyperstimulated, form cysts, and produce more estrogen. FSH and LH levels fluctuated during the treatment period. Estradiol (E 2 ) levels were significantly higher in the study group than in the control group ( Sherman et al. 1979 ). Conversely, tamoxifen manifests estrogen agonist activity in the skeleton and uterus, as well as on a number of intermediate markers of cardiovascular risk. Tamoxifen has long been considered a safe medication with few serious side effects. However, it has become clear in the past decade that prolonged use of this agent is associated with significant gynecological complications ( Table 1 ). The potential adverse effects of the drug include the development of endometrial cancer ( Killackey et al. 1985 ; Fisher et al. 1994 ; Clarke et al. 1998 ; Bernstein et al. 1999 ), endometrial polyps ( Corley et al. 1992 ; Lahti et al. 1993 ; Kedar et al. 1994 ), adenomyomatous polyp ( Nasu et al. 1997 ), adenomyosis ( Cohen et al. 1995 ), leiomyoma ( Dilts et al. 1992 ; Leo et al. 1994 ; Kang et al. 1996 ), cervical polyps, and ovarian cysts ( Sawka et al. 1986 ; Cohen et al. 1994a ; Barbieri et al. 1993 ; Shushan et al. 1996a ; Terada et al. 1993 ; Nasu et al. 1999 ). Of these diseases, tamoxifen-associated endometrial pathologies have been evaluated exclusively. There is sufficient clinical data to suggest that postmenopausal tamoxifen therapy may increase the risk of developing benign and malignant endometrial pathologies ( Cohen et al. 1998a ). It is interesting that some of these ovarian cysts are functional, such as follicular or luteinized cysts ( Dilts et al. 1992 ; Barbieri et al. 1993 ; Shushan et al. 1996a ; Terada et al. 1993 ). Tamoxifen was listed in 1996 as a human carcinogen by the International Agency of Research on Cancer ( International Agency of Research on Cancer, 1996 ). This review focuses on the effects of prolonged tamoxifen treatment on the adult human female genital tract and considers its tumorigenicity in the gynecologic organs based on an analysis of clinical data.

In breast cancer patients, ovarian cyst formation during prolonged tamoxifen treatment ( Cohen et al. 1994a ; Barbieri et al. 1993 ; Kedar et al. 1994 ; Shulman et al. 1994 ; Nasu et al. 1999 ) and in series of tamoxifen-treated breast cancer patients ( Cohen et al. 1996 ; Shushan et al. 1996a ) has been reported. Ovarian cysts also have been described in a breast cancer prevention study ( Powles et al. 1994 ). These reports described a heterogeneous group of ovarian pathologies with numerous histologic diagnoses, but they did not assess hormones or define menopausal status. In premenopausal patients, tamoxifen disrupts the menstrual cycle and causes functional ovarian cysts ( Cohen et al. 1994a ; Hochner-Celnikier et al. 1995 ). The ovarian pathology in these instances includes simple cysts, follicular cysts, luteinized follicular cysts, and corpus luteum cysts ( Cohen et al. 1994a ; Hochner-Celnikier et al. 1995 ; Shushan et al. 1996a ). Some studies suggest that benign ovarian pathologies may be expected in most premenopausal tamoxifen users ( Cohen et al. 1994a ). It is interesting that these cysts regress if tamoxifen is withdrawn ( Shushan et al. 1996a ) or if premenopausal patients are treated with GnRH agonists during tamoxifen treatment ( Cohen et al. 1994a ; Shushan et al. 1996b ). These lesions, although benign, may be complicated by torsion or cystic necrosis and may pose a diagnostic dilemma in patients at risk of ovarian metastases from breast cancer or of primary ovarian cancer ( Cohen et al. 1994a ). In premenopausal women, these cysts may be associated with hyperestrogenism ( Cohen et al. 1994a ; Hochner-Celnikier et al. 1995 ). In the ovaries of premenopausal patients, tamoxifen stimulates estrogen production by affecting the hypothalamic-pituitary-ovarian feedback mechanism ( Kedar et al. 1994 ). The hyperestrogenemia described during tamoxifen therapy may reflect a simultaneous maturation of multiple ovarian follicles or an enhanced gonadotropin stimulation of a single maturing follicle ( Sherman et al. 1979 ). Such a phenomenon may lead to an increased risk of fibroid ovaries and ovarian cysts ( Cohen et al. 1994a ). However, the mechanism by which tamoxifen stimulates the development of ovarian cysts has not yet been fully explored. It was suggested that the mechanism by which tamoxifen induces ovarian cysts in premenopausal women could be by a direct action on the ovaries to stimulate excessive growth of ovarian follicles, resulting in elevated estradiol levels (up to 3,700 pg/ml), throughout all phases of the menstrual cycle ( Terada et al. 1993 ). Mourits et al. (1999) performed a prospective study using transvaginal ultrasound with hormonal assessment and reported ovarian cysts in 40% of premenopausal women during tamoxifen treatment, whereas none of the postmenopausal patients developed cystic ovaries. In patients with regular menstrual cycles during tamoxifen treatment, 81% developed ovarian cysts. In these premenopausal women with cystic ovaries, the serum estrogen levels were markedly elevated, with gonadotropin concentrations either unchanged or slightly increased. Sawka et al. (1986) reported that 7 of 84 premenopausal women being treated with tamoxifen for breast cancer developed cystic enlargement of the ovaries. It is uncertain how many of these cases will require surgical intervention. Shushan et al. (1996a) reported that 5 of 79 tamoxifen treated premenopausal women with breast cancer had cystic enlargement of the ovaries. They also reported that, in 8 of 11 patients, the ovarian cystic enlargement disappeared after the cessation of tamoxifen treatment. Barbieri et al. (1993) reported a case of a 45-year-old woman with breast cancer treated with tamoxifen for 2 years. She had bilateral functional ovarian cysts and torsion of unilateral adnexa, and underwent surgical treatment. Terada et al. (1993) also reported a case of a large ovarian follicular cyst with torsion, whose serum estradiol level was significantly increased. Sadan et al. (2001) reported that 7 of 10 (70%) women with tamoxifen administration developed ultrasonographically benign ovarian cysts ranging from 1.5 to 6.0 cm in diameter. One woman underwent surgery to remove an enlarging cyst. In all of the other patients, ovarian cysts disappeared within three months after the cessation of therapy. We have also reported the torsion of an ovarian functional cyst in a premenopausal breast cancer patient who was treated surgically ( Nasu et al. 1999 ). Although torsion of a cystic ovary during tamoxifen treatment has been described ( Barbieri et al. 1993 ; Nasu et al. 1999 ), surgical intervention is rarely required, and functional asymptomatic monolocular cysts in these patients should be followed conservatively ( Mourits et al. 2001 ). The discontinuation of tamoxifen usually leads to the gradual reduction and disappearance of these lesions ( Cohen et al. 1994a ). Whereas tamoxifen in postmenopausal patients induces ovarian cystic tumors and endometriomas ( Kedar et al. 1994 ; Shushan et al. 1996a ; Varras et al. 2003 ), Cohen et al. (1996) reported that 10 of 16 tamoxifen-treated women who had undergone hysterectomy had ovarian neoplasms. Four of these women had serous cystadenomas, and two had serous cystadenofibromas. Other ovarian tumors in this group included an endometrioid adenocarcinoma ( Cohen et al. 1994b ), a Brenner tumor, a thecoma, and ovarian fibromas. It is controversial whether or not tamoxifen users risk developing ovarian cancer ( Cohen et al. 1996 ; Ismail, 1999 ; Lewis 2000 ).

Tamoxifen binds to the ER with a Kd of <2 nM, which is ∼20-fold lower than that of 17β-estradiol ( Capony and Rochefort, 1978 ). Administered as a single oral dose of 20 mg, tamoxifen is rapidly absorbed, with its concentration peaking in ∼5 hours. The terminal elimination half-life is ∼5–7 days. Steady-state concentrations in plasma are reached after ∼4 weeks of tamoxifen therapy in women. Tamoxifen is extensively metabolized after oral administration: ∼65% of the administered dose is excreted over 2 weeks, primarily through the feces. Tamoxifen is excreted mainly as polar conjugates, which account for ∼70% of the elimination products. Tamoxifen is hydrated by cytochrome P450 (CYP) 2D6 to the potent metabolites 4-hydroxytamoxifen and 4-hydroxy- N -desmethyl tamoxifen (endoxifen). The major metabolite, 4-hydroxy- N -desmethyl tamoxifen, is similar in biological activity to tamoxifen ( Zeneca Pharmaceuticals, 1998 ). CYP2D6 activity is considered as a determinant of tamoxifen efficacy and adverse effects ( Goetz et al. 2005 ). Breast cancer patients who were poor metabolizers of CYP2D6 had a worse clinical outcome and fewer adverse effects compared with those who were extensive metabolizers of CYP2D6. The cell-specific effects of tamoxifen in genital tissues and its divergent effects in premenopausal and postmenopausal women are complex, thus making it difficult to determine what defines the biologic effect (i.e. agonist or antagonist) on a specific gynecologic organ or tissue. Tamoxifen also interacts with cellular proteins other than the estrogen receptor, such as protein kinase C, calmodulin, transforming growth factor-β, insulin-like growth factor-I, phosphoinositide kinase, P-glycoprotein, and membrane-associated proteins through the ER-independent pathway ( Lam, 1984 ; O’Brian et al. 1985 ; Butta et al. 1992 ; Laatikainen et al. 1995 ; Cabot et al. 1997 ; Duk et al. 1997 ; Elkas et al. 1998 ; Friedman, 1998 ; Zhao et al. 1998 ). In 1996, a second isoforms, ERβ, was discovered ( Kuiper et al. 1996 ; Mosselman et al. 1996 ). One possible explanation for the tissue-selective activity of different ER ligands is that they interact with different receptors. In addition, splice variants of each of these receptors have been observed, allowing them to express various isoforms of ERα and ERβ in different tissues ( Ogawa et al. 1998 ). ERβ is expressed in tissues other than ERα, and both ERs have different ligand binding properties. These differences might contribute to the selective action of tamoxifen and other SERMs in different tissues. ERβ transcripts have been detected in tissues such as prostate, ovaries, and lungs, as well as in various parts of the central and peripheral nervous systems. In contrast, ERα is predominantly detected in the pituitary gland, ovaries, uterus, kidneys, adrenals, and mammary glands ( Kuiper et al. 1996 ; Kuiper and Gustafsson, 1997 ). These differences in the distribution of ERα and ERβ may explain the selectivity of the compounds. Tamoxifen has a similar affinity to both receptors ( Kuiper and Gustafsson, 1997 ). Some scientists believe that uterotrophic activity of tamoxifen is caused by ERβ. Both receptors have identical DNA-binding domains, but within the ligand-binding domain, the amino acid sequence diverges considerably ( Kuiper et al. 1996 ; Kuiper and Gustafsson, 1997 ) ( Fig. 3 ). Thus, although both receptors have an equivalent affinity to 17β-estradiol with regard to other substances, there are important differences. Tamoxifen acts as an antagonist when both the transactivating domains, transcription activating function (TAF)-1 and TAF-2, are suppressed, but as an agonist when the TAF-1 activation overcomes TAF-2 inhibition. Cell-specific effect of tamoxifen in the same tissue can be explained by the tripartite theory of Katzenellenbogen et al. (1996) , who studied the pharmacologic basis for the cell-and promoter specific action of steroid hormones. They referred to the cell and tissue selectivity that steroid hormones display as a tripartite system comprising ligand, receptor, and effector. Their results showed that molecular elements within the cell nucleus interact with the ligand-receptor complex and influence ER transcriptional response to the ligand. The biocharacter of the ligand (i.e. agonist-antagonist balance) is determined principally through this receptor-effector coupling ( O’Brian et al. 1985 ). This molecular explanation of the agonistic and antagonistic effects of the ligand-binded ER is further supported by others ( Brzozowski et al. 1997 ; Parker, 1998 ). Hormone binding to the ligand-binding domain of the ER initiated a series of molecular events culminating in the activation or repression of target genes. Each ligand induces a distinct conformation in the transactivation domain, creating an interacting surface to which coactivators are likely to bind. Several candidate coactivating proteins have been identified, including receptor-interacting proteins RIP-140 and RIP-160, L7SPA, and steroid receptor coactivator-1 (SRC-1) ( Onate et al. 1995 ; Horwitz et al. 1996 ; Jackson et al. 1997 ; Shah and Rowan, 2005 ). These proteins interact with receptors only in the presence of their respective ligands, providing structural evidence for the mechanism of repression or activation of target genes ( Parker, 1998 ).

The vagina is lined by stratified squamous, non-keratinizing epithelium containing ER in pre- and postmenopausal women ( Wiegerink et al. 1980 ). The epithelium is multilayered and the cells in the middle and superficial zones contain glycogen only when stimulated by estrogen. Under estrogenic stimulation, the vaginal epithelium undergoes proliferation and maturation. Several studies have reported that tamoxifen exerts a weak estrogenic effect on the vaginal epithelium in postmenopausal patients during long-term tamoxifen use ( Lahti et al. 1994 ; Mourits et al. 2001 ; Varras et al. 2003 ). The relationship between tamoxifen use and the occurrence of vaginal neoplasms has not been reported.

During long-term use, tamoxifen use has estrogenic effects in squamous epithelial cells within the cervix in postmenopausal patients ( Mourits et al. 2001 ; Varras et al. 2003 ). The association of tamoxifen with cervical polyps has been demonstrated by several studies ( Varras et al. 2003 ). Lahti et al. (1993) found that endocervical polyps were twice as common in a tamoxifen-treated patients than in a control group. The relationship between tamoxifen use and the occurrence of cervical cancer has not been reported.

The development of endometriosis has been reported in women receiving tamoxifen for the treatment of breast cancer ( Ford et al. 1988 ; Cano et al. 1989 ; Hajjar et al. 1993 ; Cohen et al. 1994b ; Morgan et al. 1994 ). A histopathologic analysis of endometriosis in a tamoxifen-treated, postmenopausal breast cancer patient showed similarity to tamoxifen-induced epithelial differentiation of the endometrium ( Ismail and Maulik, 1997 ). The development of endometriosis has been frequently reported in postmenopausal patients taking tamoxifen ( Hajjar et al. 1993 ; Cohen et al. 1994b ; Ismail and Maulik, 1997 ). Several cases of endometriosis in the ovary have also been described in premenopausal women during tamoxifen administration ( Ford et al. 1988 ; Cano et al. 1989 ; Morgan et al. 1994 ). Endometriosis is common in premenopausal women, and its occurrence in tamoxifen-treated women in this age group may therefore be coincidental. In contrast, endometriosis is rare in postmenopausal women and the finding of endometriosis in tamoxifen-treated postmenopausal women raises the possibility of a link between tamoxifen use and endometriosis. This possibility is further supported by the unusual morphological features and behavior of endometriosis in tamoxifen treated postmenopausal women ( Hajjar et al. 1993 ; Cohen et al. 1994a ; Ismail and Maulik, 1997 ). One patient had pelvic endometriosis infiltrating the cervix, vagina, rectum, and sigmoid colon ( Hajjar et al. 1993 ). Another patient had a cystic ovarian endometriotic cyst with apparent infiltration of adjacent structures ( Ismail and Maulik, 1997 ). Another patient developed an ovarian endometrioid adenocarcinoma in an endometriotic cyst ( Cohen et al. 1994b ), suggesting that prolonged tamoxifen treatment may be associated with an increased risk of carcinoma arising in endometriotic foci. As endometriosis is an estrogen-dependent disease, the mechanism by which tamoxifen acts probably that its estrogen agonistic activity stimulates the ectopic endometriotic tissue even in postmenopausal women. However, it is not clear whether tamoxifen causes de novo endometriosis or only exacerbates preexisting endometriosis. Several cases of adenomyosis have been reported as having developed in postmenopausal patients during tamoxifen administration ( Cohen et al. 1995 ; Cohen et al. 1997 ). Cohen et al. (1997) found a higher incidence of adenomyosis (53.6%) in postmenopausal breast cancer patients treated with tamoxifen than in those not receiving the drug. The morphological features present within adenomyosis more often in those taking tamoxifen were cystic dilatation of glands, fibrosis of the stroma, and various epithelial metaplasias ( McCluggage et al. 2000 ). The proliferative activity within the adenomyosis was higher in the tamoxifen group. In postmenopausal tamoxifen-treated patients, ERs and PRs have been described in adenomyosis in similar concentrations as in premenopausal non-tamoxifen users ( Cohen et al. 1998b ).

As reviewed in this paper, tamoxifen has been demonstrated to induce tumorigenesis in the female genital tract through estrogen agonism. Of the tamoxifen-associated pathologies discussed above, practitioners should be aware of the occurrence of endometrial diseases in postmenopausal women and ovarian functional cysts in premenopausal women. Tamoxifen is associated with an increased incidence of benign endometrial lesions, such as polyps and hyperplasia, as well as with a two- to three-fold increased risk of endometrial cancer in postmenopausal patients. Practitioners should be aware of the occurrence of endometrial diseases, and diagnostic procedures should be performed at the discretion of the individual gynecologist. However, there is no general consensus regarding endometrial surveillance in postmenopausal tamoxifen users. The most important recommendation by the ACOG is to thoroughly evaluate any discharge or bleeding by means of endometrial biopsy in women treated with tamoxifen ( ACOG Committee Opinion, 2006 ) ( Table 2 ). The risk of functional ovarian cysts is significantly high in premenopausal tamoxifen users. Cohen et al. ( 1994 ) suggested that all premenopausal breast cancer patients being treated with tamoxifen should be under close gynecological and ultrasonographic surveillance. We recommend that gynecological, cytological, and ultrasonographic examination should be performed every 4–6 months for the women receiving tamoxifen. In case of abnormal observations, further examination is necessary for the denial of malignancies. Despite its gynecologic side effects, benefits of tamoxifen in pre- and postmenopausal breast cancer patients in controlling breast cancer or preventing its relapse are without debate. Therefore, there is a clear need to elucidate the mechanism underlying action of tamoxifen in the reproductive tract. Genital side effects of tamoxifen are an example of the complexity of its mechanism of action, with agonistic and antagonistic effects on various sites and tissues, dependent on the ambient E 2 concentration. Frequent gynecological examination including transvaginal ultrasonography should be performed for both premenopausal and postmenopausal tamoxifen-treated patients. The multidisciplinary team including the surgeon, oncologist, and the patient’s primary care physician should be familiar with these gynecologic complications of tamoxifen therapy. Further research will enable the prediction of which groups of patients are more susceptible to develop pathologies of the genital tract.