- Email: [email protected]
Analogous tamoxifen and estrogen effects on transforming growth factor-betas 1 and 2 in the rat uterus
Reproductive
l
Toxicology, Vol. 9, No. 3, pp. 225-231, 1995 Copyright 0 1995 Elsevier Science Ltd Printed in the USA. All rights reserved 0@+623&9S $9.50 + .oO
Original Contributions
ANALOGOUS TAMOXIFEN AND ESTROGEN EFFECTS ON TRANSFORMING GROWTH FACTOR-BETAS 1 AND 2 IN THE RAT UTERUS
BELIND.AM.SARTOR,OLIVERSARTOR,~~~ KATHLEEN CFLANDERS Laboratory of Chemoprevention, National Cancer Institute, National Institutes of Health, Bethesda, Maryland Abstract - Estrogenic stimulation is a potent risk factor for the development of uterine cancer. More recently, analysis of patients in prospective breast cancer trials have established that tamoxifen also increases uterine cancer risk. In this report, uteri of oophorectomized rats were examined to ascertain the effects of estrogen and tamoxifen on lthe uterine induction of two isoforms of transforming growth factor-beta (TGF-8). In contrast to studies of cells derived from breast epithelium, our studies reveal that both estrogen and tamoxifen increase immunoreactive TGF-j3. These changes were particularly pronounced in the endometrial stroma, Effects of progesterone also were examined and found to be distinct and relatively restricted to the glandular epithelium. These studies indicate that, in the uteri of oophorectomized rats, tamoxifen exerts estrogen-like effects on a peptide previously implicated in the control of cellular growth and differentiation. We hypothesize that induction of TGF-p isoforms may be an important mediatior of both estrogen- and tamoxifen-induced proliferative disorders in the uterus. Key Words: estrogen;
transforming growth factor-beta
(TGF-/3); rat uterus;tamoxifen.
INTlXODUCTION
these trials (3). One of the most serious of these issues is that tamoxifen may be associated with uterine cancer. Epidemiologic studies have implicated tamoxifen use with a variety of uterine abnormalities including atypical hyperplasia, polyps, and invasive carcinomas (4-6). More recently, two separate prospective, multicenter, randomized breast cancer trials have clearly associated use of tamoxifen with an increased risk of atypical uterine hyperplasia (7) and adenocarcinoma (8). Prolonged estrogenic stimulation has long been correlated with atypical uterine hyperplasia and uterine cancer (9). Since tamoxifen was developed as an estrogen antagonist, the development of uterine cancers in women treated with tamoxifen was unexpected by many clinicians. In fact, some trials have reported the successful use of tamoxifen in patients with uterine cancer (10). The transduction of steroid hormone signals is an area of active investigation. Some effects of steroid hormone are hypothesized to be mediated by autocrine and/or paracrine actions of peptide growth factors (11). For example, several studies indicate
Tamoxifen has a clear role in the treatment of breast cancer patients in both the adjuvant and metastatic setting (1,2). Based upon these and other observations, clinical investigators have hypothesized that tamoxifen might also be useful for breast cancer prevention. Prospective randomized clinical trials are currently being conducted in several countries (United States, Italy, Australia, and the United Kingdom) to test this hypothesis. Given that the potential risk/benefit ratios of cancer prevention are considerably different from that of treating metastatic cancer, multiple issues and potential objections have been raised regarding Belinda M. Sartor is now at Section OfReproductive Endocrinology, Department of OB!GYN, LSU Medical Center, Shreveport LA; Oliver Sartor is now at Section of HematologylOncology, Department of Medicine, LSU Medical Center, Shreveport, LA. Address correspondence to Dr. Belinda M. Sartor, Department of OB/GYN, LSU Medical Center, 1501 Kings Highway, Shreveport, LA 71130. Received 29 September 1994; Accepted 16 November 1994. 225
226
Reproductive
Toxicology
that estrogens induce TGF-a and insulin-like growth factor-I (12,13) while decreasing expression of TGF/3 (14-16). Some reports also indicate that antiestrogens (such as tamoxifen) or estrogen withdrawal can induce TGF-p expression both in vitro (l-5-18) and in vivo (19). Furthermore, some investigators have used TGF-P measurements as a surrogate end-point for assays of antiestrogenic effects (18). In this study, we have compared the effects of estrogen, tamoxifen, and progesterone treatments on the expression of two TGF-fi isoforms in the rat uterus. In some model systems, these regulatory peptides are differentially regulated by estrogens and antiestrogens. However, the studies reported herein indicate that, in the uteri of oophorectomized rats, tamoxifen and estrogen exert essentially the same effects on TGF-/3 expression. In contrast, the effects of progesterone in this model system are clearly distinct. METHODS Adult Sprague-Dawley rats were oophorectomized at least 10 d prior to experimentation and randomly assigned to one of five different treatment groups. Each group contained at least three rats, and results were essentially the same for each of the animals within each group. One group received no treatment (Group C). Other groups received daily SC steroid injections mixed in corn oil. These groups included: Group I, 17-p-estradiol (10 pg/kg/d); Group II, tamoxifen 500 pglkgld; and Group III, progesterone (10 mg/kg/d). Another group (Group IV) received daily SC injections of corn oil alone. Rats from groups I to IV were sacrificed 1 or 7 d after initiating injections. After sacrifice, uteri were dissected from the rats and washed in cold buffered saline. Tissues were then fixed in 10% buffered formalin, paraffin embedded, and sectioned (5 pm thick) for immunohistochemical staining. TGF-Pl and TGF-/32 were identified in tissue sections using specific affinity purified rabbit-derived polyclonal antibodies (see below). Visualization of antibody binding was accomplished with avidin-biotin linked peroxidase (Vector Laboratories, Burlingame, CA). In brief, sections were deparaffinized and endogenous peroxidases were blocked by incubations with hydrogen peroxidase/methanol. Tissue permeabilization of antibodies was then increased by the addition of hyaluronidase (Calbiothem, La Jolla, CA) at a concentration of 1 mg/mL. Sections were flooded with 5% normal goat serum, 1% bovine serum albumin (BSA), and 1% ovalbumin and incubated for 1 h at room temperature. After washing, incubations with the primary antibodies
Volume 9, Number 3, 1995
were performed in 1% bovine serum albumin in Trisbuffered saline at 4 “C for 12 to 16 h. The affinity purified primary antibodies were used at concentrations of 3.5 to 10 pg/mL. In some control slides, the primary antibodies were replaced with normal rabbit IgG at similar concentrations. The sections were then copiously washed with Tris-buffered saline in 1% BSA and incubated with secondary biotinylated antibodies (goat anti-rabbit). After incubation with avidin-peroxidase complex, sections were exposed to 3,3’diaminobenzidine (50 mg/mL) in 0.1% hydrogen peroxide for 3 to 5 min to produce the visible peroxidase reaction product. The sections were then counterstained with Mayer’s hematoxylin. The primary antibodies used in this study were affinity-purified rabbit polyclonals developed by immunization of rabbits with synthetic peptides representing sequences of specific TGF-/3 isoforms. Descriptions of this process (20,21), the affinity purification procedures (22), and the specificity of these antibodies have been published in detail elsewhere (20,21,23,24). The antibodies to TGF-@I were generated after immunization with a peptide corresponding to the amino-terminal region of mature TGF-/31 (amino acids 1 to 30). TGF-P2 antibodies were elicited by immunization with peptides corresponding to amino acids 50 to 75 of the mature peptide. These antibodies do not cross-react with other TGF-P isoforms in tissue sections (unpublished data). RESULTS In oophorectomized rats, the affinity purified anti-TGF-Pl antibodies demonstrated weak but definitively positive immunoreactivity within all regions of the uterus (see Figure 1A). Staining was most apparent within the endometrial glands and the luminal epithelium. Detectable, but less pronounced, reactivity was apparent in the endometrial stroma and myometrium. Interestingly, immunostaining was more pronounced within the outer (as compared to inner) layer of the myometrium. No differences were observed between the uninjected controls as compared to the corn oil injected controls (data not shown). Anti-TGF-/31 immunoreactivity was compared to normal rabbit sera using additional control uteri. As shown in Figure lB, no staining was detectable using the normal rabbit serum. In addition, experiments using an antigenic peptide to neutralize antibody binding were also conducted. These controls also yielded no evidence of nonspecific staining (data not shown).
Analogous tamoxifen and estrogen effects
l
B. M. SARTOR ET AL.
227
(W
0
(F)
W)
Fig. 1. Effects of daily estrogen, tamoxifen, or progesterone injections on TGF-PI immunoreactivity in the uterus. (A) TGF-/31 immunostaining in rats receiving no injections. (B) Immunostaining with normal rabbit antibodies in rats receiving no injections, (C and D) TGF+l immunostaining in rats 1 or 7 d after the initiation of 17p estradiol injections, respectively. (E and F) TGF- pl immunostaining in rats 1 or 7 d after the initiation of tamoxifen injections, respectively. (G and H) TGF-/31 immunostaining in rats 1 or 7 d after the initiation of progesterone injections, respectively.
228
Reproductive
Toxicology
Uterine tissues were examined 1 and 7 d after initiating the daily SC 17-p estradiol injections (Figures IC and lD, respectively). Twenty-four hours postinjection, an increase in anti-TGF-pl immunoreactivity was readily apparent within all uterine layers. In particular, the endometrial stroma was prominently stained in comparison to the corn-oiltreated animals. The estrogen-induced changes in TGF-/31 persisted during the 7-d treatment course, however, the initially prominent changes within the endometrial glands and epithelium appeared slightly less prominent by the seventh day. To compare estrogen and antiestrogen treatments, uteri from tamoxifen-treated rats were next examined (see Figures 1E and 1F). Twenty-four hours after the first tamoxifen injection, anti-TGF/31 immunostaining was increased in all uterine layers as compared to the corn-oil-treated animals. At 24 h, the pattern of immunoreactivity was very similar to that of the estrogen-treated animals, albeit slightly less intense. This pattern of TGF-Pl staining persisted during the 7 d of tamoxifen treatment. To examine the effects of progestogens on our model system, uteri from rats treated with progesterone were also assayed for TGF-P1 immunoreactivity. As shown in Figures 1G and IH, anti-TGF-pl staining was intense within the luminal epithelium, endometrial glands, and myometrium within 24 h postinjection. Endometrial stromal staining was also slightly increased. After 7 d of progesterone injections, the immunoreactive staining remained intense within the endometrial epithelium and glands; however, staining within the endometrial stromal elements was indistinguishable from that of controls. This lack of stromal staining in combination with a robust pattern of immunoreactive epithelium was in striking contrast to that observed for both the estrogen- and tamoxifen-treated rats. To determine the effects of estrogen, tamoxifen, and progesterone treatments on TGF-/32 immunoreactivity, assays were conducted using specific antiTGF-fi2 antibodies. As shown in Figure 2A, uteri from oophorectomized (otherwise untreated) rats contained a small but definite amount of reaction product as compared to normal rabbit sera controls (see Figure 1B). All uterine layers were stained, with the outer layer of the myometrium and the endometrial glands being most affected. Twenty-four hours after estrogen injection, uterine TGF+2 immunostaining was clearly increased (Figure 2B). As in the previous experiments, changes in stromal reactivity were the most prominent. Minor changes in the luminal epithelium, glands, and myometrium were variably apparent in
Volume 9, Number 3, 1995
some uteri. This pattern of immunoreactivity was persistent for 7 d (Figure 2C). In comparison to the vehicle-injected controls, changes in the endometrial stroma were clear while changes in the other uterine layers were not. Tamoxifen and estradiol injections elicited a qualitatively similar pattern of TGF-02 immunoreactivity. As shown in Figure 2D, 24 h after tamoxifen administration, TGF$32 immunoreactivity was clearly increased in the endometrial stroma. Staining in the endometrial epithelium and glands was variable as was staining in the myometrium. After 7 d of tamoxifen administration, stromal TGF-P2 staining was even more intense. Changes in other layers were equivocal. Twenty-four hours after initiating progesterone treatment, uteri showed clear increases in TGF-P2 immunoreactivity in the luminal epithelium and glands. Other layers were not affected. After 7 d, these changes were even more dramatic. Intense staining was noted in the endometrial epithelial structures with very little change in the endometrial stroma or myometrium. These alterations are in stark contrast to the changes noted after either estradiol or tamoxifen.
DISCUSSION
These studies clearly demonstrate that immunoreactivity of at least two TGF-P isoforms are strikingly altered by estrogen, progesterone, and tamoxifen in the uteri of oophorectomized rats. For both estrogen and tamoxifen, these changes are most pronounced in the endometrial stroma. For progesterone, these changes are most conspicuous in the epithelial cells lining the endometrial surface and glands. In vitro studies of breast-cancer-derived cells have demonstrated that estrogens decrease TGF/3 levels (14-16). Our findings clearly indicate that estrogens can induce the opposite effect in the uteri of oophorectomized rats. In vitro examination of normal human endometrium explants also demonstrate that TGF-/3 protein expression can be induced by estrogens (25). In addition, estrogens increase TGF-0 expression in cultured osteosarcoma cells (26). Taken together these data suggest that estrogens regulate TGF-/3 expression in a tissue-specific manner. Das et al. (27) have previously reported the effects of in vivo estrogens on uterine TGF-P2 and TGF-/33 messenger RNA (mRNA). In that report, estrogen increased TGF-fl2 mRNA levels (as as-
Analogous tamoxifen and estrogen effects
l
B. M. SARTOR ET AL.
229
(W
(F)
Fig. 2. Effects of daily estrogen, tamoxifen, or progesterone injections on TGF-P2 immunoreactivity in the uterus. (A) TGF-/32 immunostaining in rats receiving no injections. (B and C) TGF-P2 immunostaining in rats 1 or 7 d after the initiation of 17p estradbl injections, respectively. (D and E) TGF-/32 immunostaining in rats 1 or 7 d after the initiation of tamoxifen injections, respectively. (F and G) TGF-P2 immunostaining in rats 1 or 7 d after the initiation of progesterone injections, respectively.
230
Reproductive
Toxicology
sessed by northern blots) without altering levels of TGF-/33 mRNA. TGF-PI was not examined in that report. The TGF-j32 data reported herein are compatible with the report by Das et al. and extend their observation to the TGF-/32 protein level. Tamoxifen and other antiestrogens increase TGF-/3 expression in vitro (15-17) and in vivo (19) in breast-derived epithelium. Our data, in uterine tissues, also demonstrate that tamoxifen increases at least two isoforms of TGF-#3. In contrast to breast epithelium, however, tamoxifen effects are estrogen-like in the uterus. It is also important to note that uterine IGF-I expression is stimulated by tamoxifen (28) Because estrogens increase uterine IGF-I (13), these data are also consistent with tamoxifen exerting estrogen-like effects on the uterus. When taken together, these data suggest that tamoxifen exerts estrogen-like effects in the uterus by modulating the expression of at least two peptide growth factors. Based upon these data, we hypothesize that tamoxifen increases TGF-j? in the uterus as a consequence of its estrogen-like actions. In these studies we demonstrate that estrogens and tamoxifen have analogous effects on TGF-P, a peptide implicated in cell growth, wound healing, angiogenesis, differentiation, and regulation of the extracellular matrix. Whether the TGF-@s mediate estrogen-induced effects on the uterus awaits further experimentation; however, we suggest that TGF-/3 may be an important peptide intermediary in both estrogen- and tamoxifen-linked uterine proliferative disorders. Several pieces of data, in addition to our own, support this hypothesis. TGF-P can stimulate the growth of certain endometrial carcinoma cell lines (29). In addition, a recent study examining pathologic proliferative disorders of human endometrium concluded that TGF-/3 immunostaining is markedly increased in uteri containing endometriai hyperplasia and adenocarcinoma (30). Thus, despite inhibiting epithelial growth in a number of other tissues, we hypothesize that the TGF+s act as positive mediators of estrogen-induced uterine growth. REFERENCES Bonadonna G, Valagussa P, Brambilla C, Molitemi A, Zambitti M, Ferrari LS. Adjuvant and neoadjuvant treatment of breast cancers with chemotherapy and endocrine therapy. Semin Oncol. 1991;18:515-24. Early Breast Cancer Trials Collaborative Group. Systemic treatments of early breast cancer by hormonal, cytotoxic, and immune therapy: 133 randomized trials involving 33,000 recoveries and 24,000 deaths among 75,ooO women. Lancet. 1992;399:1-15. Fugh-Berman A, Epstein S. Tamoxifen: disease prevention or disease substitution? Lancet. 1992;340:1143-5.
Volume 9, Number 3, 1995 4. Neven P, De Muylder X, Van Belle Y, et al. Hysteroscopic follow- up during tamoxifen treatment. Eur J Obstet Gynaeco1 Rep Biol. 1990;35:235-8. 5. Femanader T, Rutqvist IE, Cedermark B, et al. Adjuvant tamoxifen in early breast cancer: occurrence of new primary cancers. Lancet. 1989;i: 117-20. 6. Lahti E, Guillermo B, Kauppila A, et al. Endometrial changes in postmenopausal women receiving tamoxifen. Obstet Gynecol. 1993;81:660-4. 7. Kedar RP, Boume TH, Powles TJ, et al. Effects of tamoxifen on uterus and ovaries of postmenopausal women in a randombreast ized cancer trial. prevention Lancet. 1994;343:1318-21. 8. Fisher B, Constantino JP, Redmond CK, et al. Endometrial cancer in tamoxifen-treated breast cancer patients: findings from the National Adjuvant Breast and Bowel Project (NSABP-14). J Nat1 Cancer Inst. 1994;86:527-37. 9. Ziel HK. Finkle WD. Increased risk of endometrial carcinoma among users of conjugated estrogens. N Engl J Med. 1975;293: 1167-70. 10. Senerton KD. Treatment of advanced endometrial adenocarcinema with tamoxifen. Cancer Treat Rep. 1980$4:805-l 1. 11. Lippman ME, Dickson RB, Bates S, et al. Autocrine and paracrine growth regulation of human breast cancer. Breast Cancer Res Treat. 1986;7:59-70. 12. Bates SE, Davidson NE, Valvarius EM, et al. Expression of transforming growth factor-a and its messenger ribonucleic acid in human breast cancer: its regulation by estrogen and its possible functional significance. Mol Endocrinol. 1988;2:543-55. 13. Murphy LJ, Murphy LC, Friesen HG. Estrogen induces insulin-like growth factor-1 expression in the rat uterus. Mol Endocrinol. 1987; 1445-50. 14. Arrick BA, Korc M, Derynck R. Differential expression of three transforming growthfactor beta species in human breast cancer cell lines bv estradiol. Cancer Res. 1990:50:299-303. 15. Knabbe C, Lippman ME, Wakefield LM, et al. Evidence that transforming growth factor-p is a hormonally regulated negative growth factor in human breast cancer cells. Cell. 1987;48:417-28. 16. Jeng MH, ten Dijke P, Iwata KK, Jordan VC. Regulation of the levels of three transforming growth factor beta mRNAs by estrogen and their effects on the proliferation of human breast cancer cells. Mol Cell Endocrinol. 1993;97: 115-23. 17. Murphy CS, Pink JJ, Jordan VC. Characterization of a receptor- negative, hormone nonresponsive clone derived from a T47D human breast cancer cell line kept under estrogen-free conditions. Cancer Res. 1990;50:7285-92. 18. Knabbe C, Zugmaier G, Schmahl M, Dietel M, Lippman ME, Dickson RB. Induction of transforming growth factor beta by the antiestrogens droloxifene, tamoxifen, and toremifene in MCF-7 cells. Am J Clin Oncol. 1991;14(suppl 2):S15-20. 19. Butta A, MacLennan K, Flanders KC, et al. Induction of transforming growth factor beta 1 in human breast cancer in vivo following tamoxifen treatment. Cancer Res. 1992;52:4261-4. 20. Flanders KC, Thompson NL, Cissel DS, et al. Transforming growth factor-pl: histochemical localization with antibodies to different epitopes. J Cell Biol. 1989;108:653-60. 21. Flanders KC, Cissel DS, Mullen LT, et al. Antibodies to transforming growth factor-p2 peptides: specific detection of TGF- p2 in immunoassays. Growth Factors. 1990;3:45-52. 22. Flanders KC, Roberts AB, Ling N, et al. Antibodies to peptide determinants of transforming growth factor-b and their applications. Biochemistry. 1988;27:739-46. 23. McCune BK, Mullin BR, Flanders KC, Jaffurs WJ, Mullen LT, Spom MB. Localization of transforming growth factor/3 isotypes in lesions of the human breast. Hum Pathol. 1992;23:13-20. 24. Horikoshi S, McCune BK, Ray PE, et al. Water deprivation stimulates transforming growth factor-p2 accumulation in the
Analogous tamoxifen and estrogen effects juxtaglomerular apparatus of mouse kidney. J Clin Invest. 1991;88:2117-22. 25. Ohsteen KG, Bruner IKL, Gold LI, Hargrove JT. Steroidal regulation of transforming growth factor-p (TGF-/3) expression in normal endometrium and endometriosis. F&i1 Steril. 1993(S199):P-260. 26. Komm BS, Terpening CM, Benz DJ, et al. Estrogen binding, receptor mRNA, and biologic responses in osteoblast-like osteosarcoma cells. Science. 1988:241:81-4. 27. Das SK, Flanders KC, Andrews GK, Dey SK. Expression of transforming growth factor-p isoforms @2 and p3) in the mouse uterus: analysis of the perimplantation period and effects of ovarian steroids. Endocrinology. 1992;130: 3459-66.
l
B. M. SARTOR ET AL.
231
28. Huynh HT, Pollak M. Insulin-like growth factor I gene expression in the uterus is stimulated by tamoxifen and inhibited by the pure antiestrogen ICI 182780. Cancer Res. 1993;53:5585-8. 29. Anzai Y, Gong Y, Holinka CF, et al. Effects of transforming growth factors and regulation of their mRNA levels in two human endometrial adenocarcinoma cell lines. J Steroid Biothem Mol Biol. 1992;42:449-55. 30. Gold LI, Saxena B, Mittal KR, et al. Increased expression of transforming growth factor fi isoforms and basic fibroblast growth factor in complex hyperplasia and adenocarcinoma of the endometrium: evidence for paracrine and autocrine action. Cancer Res. 1994;54:2347-58.
l
Toxicology, Vol. 9, No. 3, pp. 225-231, 1995 Copyright 0 1995 Elsevier Science Ltd Printed in the USA. All rights reserved 0@+623&9S $9.50 + .oO
Original Contributions
ANALOGOUS TAMOXIFEN AND ESTROGEN EFFECTS ON TRANSFORMING GROWTH FACTOR-BETAS 1 AND 2 IN THE RAT UTERUS
BELIND.AM.SARTOR,OLIVERSARTOR,~~~ KATHLEEN CFLANDERS Laboratory of Chemoprevention, National Cancer Institute, National Institutes of Health, Bethesda, Maryland Abstract - Estrogenic stimulation is a potent risk factor for the development of uterine cancer. More recently, analysis of patients in prospective breast cancer trials have established that tamoxifen also increases uterine cancer risk. In this report, uteri of oophorectomized rats were examined to ascertain the effects of estrogen and tamoxifen on lthe uterine induction of two isoforms of transforming growth factor-beta (TGF-8). In contrast to studies of cells derived from breast epithelium, our studies reveal that both estrogen and tamoxifen increase immunoreactive TGF-j3. These changes were particularly pronounced in the endometrial stroma, Effects of progesterone also were examined and found to be distinct and relatively restricted to the glandular epithelium. These studies indicate that, in the uteri of oophorectomized rats, tamoxifen exerts estrogen-like effects on a peptide previously implicated in the control of cellular growth and differentiation. We hypothesize that induction of TGF-p isoforms may be an important mediatior of both estrogen- and tamoxifen-induced proliferative disorders in the uterus. Key Words: estrogen;
transforming growth factor-beta
(TGF-/3); rat uterus;tamoxifen.
INTlXODUCTION
these trials (3). One of the most serious of these issues is that tamoxifen may be associated with uterine cancer. Epidemiologic studies have implicated tamoxifen use with a variety of uterine abnormalities including atypical hyperplasia, polyps, and invasive carcinomas (4-6). More recently, two separate prospective, multicenter, randomized breast cancer trials have clearly associated use of tamoxifen with an increased risk of atypical uterine hyperplasia (7) and adenocarcinoma (8). Prolonged estrogenic stimulation has long been correlated with atypical uterine hyperplasia and uterine cancer (9). Since tamoxifen was developed as an estrogen antagonist, the development of uterine cancers in women treated with tamoxifen was unexpected by many clinicians. In fact, some trials have reported the successful use of tamoxifen in patients with uterine cancer (10). The transduction of steroid hormone signals is an area of active investigation. Some effects of steroid hormone are hypothesized to be mediated by autocrine and/or paracrine actions of peptide growth factors (11). For example, several studies indicate
Tamoxifen has a clear role in the treatment of breast cancer patients in both the adjuvant and metastatic setting (1,2). Based upon these and other observations, clinical investigators have hypothesized that tamoxifen might also be useful for breast cancer prevention. Prospective randomized clinical trials are currently being conducted in several countries (United States, Italy, Australia, and the United Kingdom) to test this hypothesis. Given that the potential risk/benefit ratios of cancer prevention are considerably different from that of treating metastatic cancer, multiple issues and potential objections have been raised regarding Belinda M. Sartor is now at Section OfReproductive Endocrinology, Department of OB!GYN, LSU Medical Center, Shreveport LA; Oliver Sartor is now at Section of HematologylOncology, Department of Medicine, LSU Medical Center, Shreveport, LA. Address correspondence to Dr. Belinda M. Sartor, Department of OB/GYN, LSU Medical Center, 1501 Kings Highway, Shreveport, LA 71130. Received 29 September 1994; Accepted 16 November 1994. 225
226
Reproductive
Toxicology
that estrogens induce TGF-a and insulin-like growth factor-I (12,13) while decreasing expression of TGF/3 (14-16). Some reports also indicate that antiestrogens (such as tamoxifen) or estrogen withdrawal can induce TGF-p expression both in vitro (l-5-18) and in vivo (19). Furthermore, some investigators have used TGF-P measurements as a surrogate end-point for assays of antiestrogenic effects (18). In this study, we have compared the effects of estrogen, tamoxifen, and progesterone treatments on the expression of two TGF-fi isoforms in the rat uterus. In some model systems, these regulatory peptides are differentially regulated by estrogens and antiestrogens. However, the studies reported herein indicate that, in the uteri of oophorectomized rats, tamoxifen and estrogen exert essentially the same effects on TGF-/3 expression. In contrast, the effects of progesterone in this model system are clearly distinct. METHODS Adult Sprague-Dawley rats were oophorectomized at least 10 d prior to experimentation and randomly assigned to one of five different treatment groups. Each group contained at least three rats, and results were essentially the same for each of the animals within each group. One group received no treatment (Group C). Other groups received daily SC steroid injections mixed in corn oil. These groups included: Group I, 17-p-estradiol (10 pg/kg/d); Group II, tamoxifen 500 pglkgld; and Group III, progesterone (10 mg/kg/d). Another group (Group IV) received daily SC injections of corn oil alone. Rats from groups I to IV were sacrificed 1 or 7 d after initiating injections. After sacrifice, uteri were dissected from the rats and washed in cold buffered saline. Tissues were then fixed in 10% buffered formalin, paraffin embedded, and sectioned (5 pm thick) for immunohistochemical staining. TGF-Pl and TGF-/32 were identified in tissue sections using specific affinity purified rabbit-derived polyclonal antibodies (see below). Visualization of antibody binding was accomplished with avidin-biotin linked peroxidase (Vector Laboratories, Burlingame, CA). In brief, sections were deparaffinized and endogenous peroxidases were blocked by incubations with hydrogen peroxidase/methanol. Tissue permeabilization of antibodies was then increased by the addition of hyaluronidase (Calbiothem, La Jolla, CA) at a concentration of 1 mg/mL. Sections were flooded with 5% normal goat serum, 1% bovine serum albumin (BSA), and 1% ovalbumin and incubated for 1 h at room temperature. After washing, incubations with the primary antibodies
Volume 9, Number 3, 1995
were performed in 1% bovine serum albumin in Trisbuffered saline at 4 “C for 12 to 16 h. The affinity purified primary antibodies were used at concentrations of 3.5 to 10 pg/mL. In some control slides, the primary antibodies were replaced with normal rabbit IgG at similar concentrations. The sections were then copiously washed with Tris-buffered saline in 1% BSA and incubated with secondary biotinylated antibodies (goat anti-rabbit). After incubation with avidin-peroxidase complex, sections were exposed to 3,3’diaminobenzidine (50 mg/mL) in 0.1% hydrogen peroxide for 3 to 5 min to produce the visible peroxidase reaction product. The sections were then counterstained with Mayer’s hematoxylin. The primary antibodies used in this study were affinity-purified rabbit polyclonals developed by immunization of rabbits with synthetic peptides representing sequences of specific TGF-/3 isoforms. Descriptions of this process (20,21), the affinity purification procedures (22), and the specificity of these antibodies have been published in detail elsewhere (20,21,23,24). The antibodies to TGF-@I were generated after immunization with a peptide corresponding to the amino-terminal region of mature TGF-/31 (amino acids 1 to 30). TGF-P2 antibodies were elicited by immunization with peptides corresponding to amino acids 50 to 75 of the mature peptide. These antibodies do not cross-react with other TGF-P isoforms in tissue sections (unpublished data). RESULTS In oophorectomized rats, the affinity purified anti-TGF-Pl antibodies demonstrated weak but definitively positive immunoreactivity within all regions of the uterus (see Figure 1A). Staining was most apparent within the endometrial glands and the luminal epithelium. Detectable, but less pronounced, reactivity was apparent in the endometrial stroma and myometrium. Interestingly, immunostaining was more pronounced within the outer (as compared to inner) layer of the myometrium. No differences were observed between the uninjected controls as compared to the corn oil injected controls (data not shown). Anti-TGF-/31 immunoreactivity was compared to normal rabbit sera using additional control uteri. As shown in Figure lB, no staining was detectable using the normal rabbit serum. In addition, experiments using an antigenic peptide to neutralize antibody binding were also conducted. These controls also yielded no evidence of nonspecific staining (data not shown).
Analogous tamoxifen and estrogen effects
l
B. M. SARTOR ET AL.
227
(W
0
(F)
W)
Fig. 1. Effects of daily estrogen, tamoxifen, or progesterone injections on TGF-PI immunoreactivity in the uterus. (A) TGF-/31 immunostaining in rats receiving no injections. (B) Immunostaining with normal rabbit antibodies in rats receiving no injections, (C and D) TGF+l immunostaining in rats 1 or 7 d after the initiation of 17p estradiol injections, respectively. (E and F) TGF- pl immunostaining in rats 1 or 7 d after the initiation of tamoxifen injections, respectively. (G and H) TGF-/31 immunostaining in rats 1 or 7 d after the initiation of progesterone injections, respectively.
228
Reproductive
Toxicology
Uterine tissues were examined 1 and 7 d after initiating the daily SC 17-p estradiol injections (Figures IC and lD, respectively). Twenty-four hours postinjection, an increase in anti-TGF-pl immunoreactivity was readily apparent within all uterine layers. In particular, the endometrial stroma was prominently stained in comparison to the corn-oiltreated animals. The estrogen-induced changes in TGF-/31 persisted during the 7-d treatment course, however, the initially prominent changes within the endometrial glands and epithelium appeared slightly less prominent by the seventh day. To compare estrogen and antiestrogen treatments, uteri from tamoxifen-treated rats were next examined (see Figures 1E and 1F). Twenty-four hours after the first tamoxifen injection, anti-TGF/31 immunostaining was increased in all uterine layers as compared to the corn-oil-treated animals. At 24 h, the pattern of immunoreactivity was very similar to that of the estrogen-treated animals, albeit slightly less intense. This pattern of TGF-Pl staining persisted during the 7 d of tamoxifen treatment. To examine the effects of progestogens on our model system, uteri from rats treated with progesterone were also assayed for TGF-P1 immunoreactivity. As shown in Figures 1G and IH, anti-TGF-pl staining was intense within the luminal epithelium, endometrial glands, and myometrium within 24 h postinjection. Endometrial stromal staining was also slightly increased. After 7 d of progesterone injections, the immunoreactive staining remained intense within the endometrial epithelium and glands; however, staining within the endometrial stromal elements was indistinguishable from that of controls. This lack of stromal staining in combination with a robust pattern of immunoreactive epithelium was in striking contrast to that observed for both the estrogen- and tamoxifen-treated rats. To determine the effects of estrogen, tamoxifen, and progesterone treatments on TGF-/32 immunoreactivity, assays were conducted using specific antiTGF-fi2 antibodies. As shown in Figure 2A, uteri from oophorectomized (otherwise untreated) rats contained a small but definite amount of reaction product as compared to normal rabbit sera controls (see Figure 1B). All uterine layers were stained, with the outer layer of the myometrium and the endometrial glands being most affected. Twenty-four hours after estrogen injection, uterine TGF+2 immunostaining was clearly increased (Figure 2B). As in the previous experiments, changes in stromal reactivity were the most prominent. Minor changes in the luminal epithelium, glands, and myometrium were variably apparent in
Volume 9, Number 3, 1995
some uteri. This pattern of immunoreactivity was persistent for 7 d (Figure 2C). In comparison to the vehicle-injected controls, changes in the endometrial stroma were clear while changes in the other uterine layers were not. Tamoxifen and estradiol injections elicited a qualitatively similar pattern of TGF-02 immunoreactivity. As shown in Figure 2D, 24 h after tamoxifen administration, TGF$32 immunoreactivity was clearly increased in the endometrial stroma. Staining in the endometrial epithelium and glands was variable as was staining in the myometrium. After 7 d of tamoxifen administration, stromal TGF-P2 staining was even more intense. Changes in other layers were equivocal. Twenty-four hours after initiating progesterone treatment, uteri showed clear increases in TGF-P2 immunoreactivity in the luminal epithelium and glands. Other layers were not affected. After 7 d, these changes were even more dramatic. Intense staining was noted in the endometrial epithelial structures with very little change in the endometrial stroma or myometrium. These alterations are in stark contrast to the changes noted after either estradiol or tamoxifen.
DISCUSSION
These studies clearly demonstrate that immunoreactivity of at least two TGF-P isoforms are strikingly altered by estrogen, progesterone, and tamoxifen in the uteri of oophorectomized rats. For both estrogen and tamoxifen, these changes are most pronounced in the endometrial stroma. For progesterone, these changes are most conspicuous in the epithelial cells lining the endometrial surface and glands. In vitro studies of breast-cancer-derived cells have demonstrated that estrogens decrease TGF/3 levels (14-16). Our findings clearly indicate that estrogens can induce the opposite effect in the uteri of oophorectomized rats. In vitro examination of normal human endometrium explants also demonstrate that TGF-/3 protein expression can be induced by estrogens (25). In addition, estrogens increase TGF-0 expression in cultured osteosarcoma cells (26). Taken together these data suggest that estrogens regulate TGF-/3 expression in a tissue-specific manner. Das et al. (27) have previously reported the effects of in vivo estrogens on uterine TGF-P2 and TGF-/33 messenger RNA (mRNA). In that report, estrogen increased TGF-fl2 mRNA levels (as as-
Analogous tamoxifen and estrogen effects
l
B. M. SARTOR ET AL.
229
(W
(F)
Fig. 2. Effects of daily estrogen, tamoxifen, or progesterone injections on TGF-P2 immunoreactivity in the uterus. (A) TGF-/32 immunostaining in rats receiving no injections. (B and C) TGF-P2 immunostaining in rats 1 or 7 d after the initiation of 17p estradbl injections, respectively. (D and E) TGF-/32 immunostaining in rats 1 or 7 d after the initiation of tamoxifen injections, respectively. (F and G) TGF-P2 immunostaining in rats 1 or 7 d after the initiation of progesterone injections, respectively.
230
Reproductive
Toxicology
sessed by northern blots) without altering levels of TGF-/33 mRNA. TGF-PI was not examined in that report. The TGF-j32 data reported herein are compatible with the report by Das et al. and extend their observation to the TGF-/32 protein level. Tamoxifen and other antiestrogens increase TGF-/3 expression in vitro (15-17) and in vivo (19) in breast-derived epithelium. Our data, in uterine tissues, also demonstrate that tamoxifen increases at least two isoforms of TGF-#3. In contrast to breast epithelium, however, tamoxifen effects are estrogen-like in the uterus. It is also important to note that uterine IGF-I expression is stimulated by tamoxifen (28) Because estrogens increase uterine IGF-I (13), these data are also consistent with tamoxifen exerting estrogen-like effects on the uterus. When taken together, these data suggest that tamoxifen exerts estrogen-like effects in the uterus by modulating the expression of at least two peptide growth factors. Based upon these data, we hypothesize that tamoxifen increases TGF-j? in the uterus as a consequence of its estrogen-like actions. In these studies we demonstrate that estrogens and tamoxifen have analogous effects on TGF-P, a peptide implicated in cell growth, wound healing, angiogenesis, differentiation, and regulation of the extracellular matrix. Whether the TGF-@s mediate estrogen-induced effects on the uterus awaits further experimentation; however, we suggest that TGF-/3 may be an important peptide intermediary in both estrogen- and tamoxifen-linked uterine proliferative disorders. Several pieces of data, in addition to our own, support this hypothesis. TGF-P can stimulate the growth of certain endometrial carcinoma cell lines (29). In addition, a recent study examining pathologic proliferative disorders of human endometrium concluded that TGF-/3 immunostaining is markedly increased in uteri containing endometriai hyperplasia and adenocarcinoma (30). Thus, despite inhibiting epithelial growth in a number of other tissues, we hypothesize that the TGF+s act as positive mediators of estrogen-induced uterine growth. REFERENCES Bonadonna G, Valagussa P, Brambilla C, Molitemi A, Zambitti M, Ferrari LS. Adjuvant and neoadjuvant treatment of breast cancers with chemotherapy and endocrine therapy. Semin Oncol. 1991;18:515-24. Early Breast Cancer Trials Collaborative Group. Systemic treatments of early breast cancer by hormonal, cytotoxic, and immune therapy: 133 randomized trials involving 33,000 recoveries and 24,000 deaths among 75,ooO women. Lancet. 1992;399:1-15. Fugh-Berman A, Epstein S. Tamoxifen: disease prevention or disease substitution? Lancet. 1992;340:1143-5.
Volume 9, Number 3, 1995 4. Neven P, De Muylder X, Van Belle Y, et al. Hysteroscopic follow- up during tamoxifen treatment. Eur J Obstet Gynaeco1 Rep Biol. 1990;35:235-8. 5. Femanader T, Rutqvist IE, Cedermark B, et al. Adjuvant tamoxifen in early breast cancer: occurrence of new primary cancers. Lancet. 1989;i: 117-20. 6. Lahti E, Guillermo B, Kauppila A, et al. Endometrial changes in postmenopausal women receiving tamoxifen. Obstet Gynecol. 1993;81:660-4. 7. Kedar RP, Boume TH, Powles TJ, et al. Effects of tamoxifen on uterus and ovaries of postmenopausal women in a randombreast ized cancer trial. prevention Lancet. 1994;343:1318-21. 8. Fisher B, Constantino JP, Redmond CK, et al. Endometrial cancer in tamoxifen-treated breast cancer patients: findings from the National Adjuvant Breast and Bowel Project (NSABP-14). J Nat1 Cancer Inst. 1994;86:527-37. 9. Ziel HK. Finkle WD. Increased risk of endometrial carcinoma among users of conjugated estrogens. N Engl J Med. 1975;293: 1167-70. 10. Senerton KD. Treatment of advanced endometrial adenocarcinema with tamoxifen. Cancer Treat Rep. 1980$4:805-l 1. 11. Lippman ME, Dickson RB, Bates S, et al. Autocrine and paracrine growth regulation of human breast cancer. Breast Cancer Res Treat. 1986;7:59-70. 12. Bates SE, Davidson NE, Valvarius EM, et al. Expression of transforming growth factor-a and its messenger ribonucleic acid in human breast cancer: its regulation by estrogen and its possible functional significance. Mol Endocrinol. 1988;2:543-55. 13. Murphy LJ, Murphy LC, Friesen HG. Estrogen induces insulin-like growth factor-1 expression in the rat uterus. Mol Endocrinol. 1987; 1445-50. 14. Arrick BA, Korc M, Derynck R. Differential expression of three transforming growthfactor beta species in human breast cancer cell lines bv estradiol. Cancer Res. 1990:50:299-303. 15. Knabbe C, Lippman ME, Wakefield LM, et al. Evidence that transforming growth factor-p is a hormonally regulated negative growth factor in human breast cancer cells. Cell. 1987;48:417-28. 16. Jeng MH, ten Dijke P, Iwata KK, Jordan VC. Regulation of the levels of three transforming growth factor beta mRNAs by estrogen and their effects on the proliferation of human breast cancer cells. Mol Cell Endocrinol. 1993;97: 115-23. 17. Murphy CS, Pink JJ, Jordan VC. Characterization of a receptor- negative, hormone nonresponsive clone derived from a T47D human breast cancer cell line kept under estrogen-free conditions. Cancer Res. 1990;50:7285-92. 18. Knabbe C, Zugmaier G, Schmahl M, Dietel M, Lippman ME, Dickson RB. Induction of transforming growth factor beta by the antiestrogens droloxifene, tamoxifen, and toremifene in MCF-7 cells. Am J Clin Oncol. 1991;14(suppl 2):S15-20. 19. Butta A, MacLennan K, Flanders KC, et al. Induction of transforming growth factor beta 1 in human breast cancer in vivo following tamoxifen treatment. Cancer Res. 1992;52:4261-4. 20. Flanders KC, Thompson NL, Cissel DS, et al. Transforming growth factor-pl: histochemical localization with antibodies to different epitopes. J Cell Biol. 1989;108:653-60. 21. Flanders KC, Cissel DS, Mullen LT, et al. Antibodies to transforming growth factor-p2 peptides: specific detection of TGF- p2 in immunoassays. Growth Factors. 1990;3:45-52. 22. Flanders KC, Roberts AB, Ling N, et al. Antibodies to peptide determinants of transforming growth factor-b and their applications. Biochemistry. 1988;27:739-46. 23. McCune BK, Mullin BR, Flanders KC, Jaffurs WJ, Mullen LT, Spom MB. Localization of transforming growth factor/3 isotypes in lesions of the human breast. Hum Pathol. 1992;23:13-20. 24. Horikoshi S, McCune BK, Ray PE, et al. Water deprivation stimulates transforming growth factor-p2 accumulation in the
Analogous tamoxifen and estrogen effects juxtaglomerular apparatus of mouse kidney. J Clin Invest. 1991;88:2117-22. 25. Ohsteen KG, Bruner IKL, Gold LI, Hargrove JT. Steroidal regulation of transforming growth factor-p (TGF-/3) expression in normal endometrium and endometriosis. F&i1 Steril. 1993(S199):P-260. 26. Komm BS, Terpening CM, Benz DJ, et al. Estrogen binding, receptor mRNA, and biologic responses in osteoblast-like osteosarcoma cells. Science. 1988:241:81-4. 27. Das SK, Flanders KC, Andrews GK, Dey SK. Expression of transforming growth factor-p isoforms @2 and p3) in the mouse uterus: analysis of the perimplantation period and effects of ovarian steroids. Endocrinology. 1992;130: 3459-66.
l
B. M. SARTOR ET AL.
231
28. Huynh HT, Pollak M. Insulin-like growth factor I gene expression in the uterus is stimulated by tamoxifen and inhibited by the pure antiestrogen ICI 182780. Cancer Res. 1993;53:5585-8. 29. Anzai Y, Gong Y, Holinka CF, et al. Effects of transforming growth factors and regulation of their mRNA levels in two human endometrial adenocarcinoma cell lines. J Steroid Biothem Mol Biol. 1992;42:449-55. 30. Gold LI, Saxena B, Mittal KR, et al. Increased expression of transforming growth factor fi isoforms and basic fibroblast growth factor in complex hyperplasia and adenocarcinoma of the endometrium: evidence for paracrine and autocrine action. Cancer Res. 1994;54:2347-58.