- Email: [email protected]
Mechanisms for tamoxifen resistance in breast cancer: Possible role of tamoxifen metabolism
J. Steroid Biochem. Molec. Biol. Vol. 47, No. I-6, pp. 83-89, 1993
0960-0760/93 $6.00 + 0.00 Copyright © 1993 Pergamon Press Ltd
Printed in Great Britain. All fights reserved
MECHANISMS FOR TAMOXIFEN RESISTANCE IN BREAST CANCER: POSSIBLE ROLE OF TAMOXIFEN METABOLISM C. KENT OSBORNE Department of Medicine, Division of Medical Ontology, University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, TX 78284-7884, U.S.A.
Summary--Potential mechanisms for tamoxifen resistance include loss or alteration in estrogen receptor or other transcription factors and altered tamoxifen pharmacology. Using an experimental model, we have previously demonstrated that one form of tamoxifen resistance is related to the acquired ability of tamoxifen to stimulate tumor growth. These tamoxifen-stimulated tumors contain a reduced tamoxifen concentration and an altered metabolite profile suggesting that accumulation of more estrogenic metabolites could explain this phenomenon. However, /n vivo treatment of nude mice carrying tamoxifen-stimulated tumors with fixed ring non-isomerizable analogs, or other analogs resistant to conversion to metabolite E (a full estrogen), still resulted in tumor growth stimulation. Growth of these tamoxifen-stimulated tumors was inhibited by a pure steroidal antiestrogen, ICI 182,780, suggesting that this drug should be investigated in patients with tamoxifen resistance. These tamoxifen-stimulated tumors could be further stimulated by estrogen replenishment, and estrogen stimulation was blocked by tamoxifen, indicating that tamoxifen has both agonist and antagonist properties in these tumors. Our data suggest that although tamoxifen-stimulated tumors display a markedly altered metabolite profile, isomerization or metabolism of tamoxifen does not fully explain the development of tamoxifen-stimulated growth. The mechanisms by which tamoxifen acquires more potent in vivo agonist properties over time remains to be defined.
for preventing or reversing the emergence of resistant cells.
INTRODUCTION
Tamoxifen is a non-steroidal antiestrogen which is now the most frequently used drug in the treatment of breast cancer. The primary mode of action of tamoxifen and other antiestrogens is inhibition of estrogen-induced growth by competitive antagonism of the estrogen receptor (ER). The drug effectively prolongs disease-free and overall survival of women treated adjuvantly following primary surgery [1], and it can also induce tumor regression in more than half of patients with metastatic disease who have ER-positive tumors[2]. Although tamoxifen treatment is effective in many patients, still 40 to 50% of patients are resistant initially to tamoxifen therapy, and all patients, even those who initially respond, eventually develop acquired tamoxifen resistance leading to tumor progression. Understanding the mechanisms by which tumors develop resistance to tamoxifen might uncover new strategies
Clinical clues f o r the development o f acquired tamoxifen resistance
Clinical experience with tamoxifen has provided several clues for the mechanisms by which tumors become resistant. Patients who initially respond to tamoxifen but later develop resistance and tumor progression frequently respond to second or third-line endocrine therapies such as an aromatase inhibitor or megestrol acetate [2]. Thus, acquired tamoxifen resistance does not necessarily indicate global hormonal unresponsiveness, but rather selective resistance to tamoxifen itself. Some patients with acquired tamoxifen resistance, however, do develop resistance to all forms of endocrine therapy via selection of an ER-negative tumor cell clone. We have recently evaluated a series of patients with acquired tamoxifen resistance using immunohistochemical as well as ligand binding assays to measure ER and progesterone receptors (PgR)[3]. Immunohistochemical assays were used to circumvent the problem of false negative assays caused by receptor occupancy
Proceedings of the llth International Symposium of the Journal of Steroid Biochemistry and Molecular Biology, Recent Advances in Steroid Biochemistry and Molecular Biology, Seefeld, Austria, 30 May-2 June 1993. SBMB 4711-6--43
83
84
C. KENT OSBOR~
by the drug. Using these methods, more than two thirds of patients expressed ER and/or PgR in their tumors after the development of progression on tamoxifen. These data indicate that mechanisms of resistance other than receptor loss are common. Although it has not been systematically studied, anecdotal experience suggests that some patients with progression on tamoxifen will respond to a rechallenge with tamoxifen after an interval in which they received alternative treatments. Furthermore, patients who received tamoxifen adjuvant therapy and then later developed disease progression frequently respond to a second treatment with the drug. This suggests that tamoxifen resistance in some cases may not be a permanent phenotype but is reversible when the drug is stopped. Patients may also respond to an increase in the tamoxifen dose after developing progression with a lower dose schedule [4]. Finally, as with highdose estrogen therapy, some patients with metastatic disease who have responded to tamoxifen will have a withdrawal response when the drug is stopped at the time of tumor progression [5]. The long half-life of tamoxifen makes it difficult for clinicians to withhold alternative therapy while waiting for a withdrawal response to the drug. Nevertheless, these data strongly suggest that in some patients with acquired resistance, tamoxifen may actually be stimulating tumor growth. Results of two previously published clinical trials also suggest that tamoxifen can behave as an estrogen and stimulate tumor growth at the time of progression in some patients [6, 7]. In these studies of premenopausal women with advanced breast cancer, second line ovarian ablation was performed in patients who initially responded to tamoxifen. In one of these studies, secondary response to ovarian ablation was common in patients who had previously responded to tamoxifen suggesting that tamoxifen treatment served as an /n vivo tumor estrogen sensitivity assay[6]. However, in the other study, contradictory results were obtained and no patient responded to subsequent ovarian ablation[7]. In the latter study, tamoxifen therapy was continued after castration while in the first study, tamoxifen treatment was stopped. No response to ovarian ablation would be expected if tamoxifen itself was behaving as an estrogen agonist and stimulating tumor growth in patients on the second study. Tamoxifen-stimulated tumor growth as a mechanism
for acquired resistance is further supported by data from preclinical experimental models (see below). Altered ER as a mechanism for tamoxifen resistance
Although the ER protein can be detected in many patients with acquired tamoxifen resistance, mutations in the protein or in other transcription factors could conceivably contribute to tamoxifen resistance. Human ER contains at least five functional domains[8]. Transcription activating functions are located in the A/B region and in the E region, which also serves as the hormone binding domain. The C region is involved in DNA binding and the D region is involved with dimerization. Mutations in the C region would render the receptor non-functional by preventing it from binding to estrogen response elements on target genes. Mutations in the E region could alter the transcription activating function or alter binding of estrogen and other ligands. Mutations in the A/B region could also alter the transcription activating activity of the receptor conceivably leading to tamoxifen resistance. While several different mutations in human ER have been identified in cultured human breast cancer cells as well as in human breast cancer specimens, the clinical significance of these altered receptors and whether they are responsible for tamoxifen resistance in some patients remains to be demonstrated [9-15]. Experimental model of tamoxifen resistance
We have developed an in vivo experimental model of tamoxifen resistance using the MCF-7 human breast cancer cells growing as solid tumors in athymic nude mice [16, 17]. Tamoxifen suppresses tumor growth in this model for several months, and then tumor growth resumes despite continued treatment with the drug. Resistance to tamoxifen is not due to the emergence of an ER-negative clone or to alterations in tamoxifen serum levels. Interestingly, we have shown that acquired tamoxifen "resistance" in this model is related to the acquired ability of tamoxifen after several months of treatment to stimulate rather than to inhibit tumor growth. Similar data have been reported by Gottardis and Jordan [18]. These data support the clinical observations described above suggesting that one form of tamoxifen resistance may be due to growth stimulation of the tumor by tamoxifen. Since ER is normal in these tumors, we next
Tamoxifen resistance in breast cancer
investigated tamoxifen pharmacology and metabolism as a possible mechanism for tamoxifen-stimulated growth. Tamoxifen pharmacology and metabolism Tamoxifen exists in both a trans and a cis configuration [19]. The trans isomer has potent antiestrogenic properties but the cis isomer is a full estrogen agonist. Pharmaceutically manufactured tamoxifen is the trans isomer. In humans the major metabolites of tamoxifen include N-desmethyltamoxifen and 4-hydroxytamoxifen. Although trans-4-hydroxytamoxifen quantitatively is a minor metabolite, it binds to ER with much higher affinity than the parent drug. Thus, it is a more potent metabolite and it may contribute significantly to the antiestrogenic effects of tamoxifen. Trans-4-hydroxytamoxifen can also spontaneously isomerize to cis-4-hydroxytamoxifen, a much less potent antiestrogen [20]. Two other metabolites in which the dimethylaminoethoxy side chain of tamoxifen has been cleaved (metabolite E and bisphenol) are potent estrogen agonists [21]. Metabolism either in the host or in the tumor to less antiestrogenic or more estrogenic metabolites could lead to tamoxifenstimulated growth and tamoxifen "resistance" in the patient. To further investigate the mechanism for tamoxifen-stimulated tumor growth in our experimental model, we compared levels of tamoxifen and several of its metabolites in both the tamoxifen-inhibited and -stimulated tumors[17]. In our initial studies, we found that tumor extracts from tamoxifen-stimulated tumors were characterized by a 10-fold lower tamoxifen concentration and by isomerization of trans-4-hydroxytamoxifen to the less potent cis isomer. A similar pattern of reduced tamoxifen concentration and a relative increase in cis-4-hydroxytamoxifen was found in both cytosol as well as nuclear extracts. The mechanism for this reduced tamoxifen concentration in tamoxifen-resistant tumors is unknown but, as described below, it has now been reproduced in two preliminary studies from patients with tamoxifen resistance [22, 23]. The fact that serum levels from mice with tamoxifen-resistant tumors remain normal while tumor levels are markedly reduced suggests the possibility of an efflux mechanism that reduces net uptake of tamoxifen by the tumor. Since tamoxifen is known to interact with P-glycoprotein, we evaluated its ex-
85
pression at both the RNA and protein level in tamoxifen-stimulated tumors. However, P-glycoprotein was not overexpressed in any of these tumors suggesting that it is not involved in the resistance phenomenon (unpublished data). To further investigate the hypothesis that metabolic tolerance as manifested by decreasing accumulation of tamoxifen in the tumor, by isomerization of the potent antiestrogenic metabolite trans-4-hydroxytamoxifen to the less potent cis isomer, and by accumulation of estrogenic metabolites is a cause of tamoxifenstimulated growth and clinical tamoxifen resistance, we measured tumor concentrations of tamoxifen and several of its metabolites in a small series of patients who were on tamoxifen at the time of the tissue biopsy [22]. Tumor tamoxifen levels varied over a wide range. Low concentrations were observed in tumors from 8 patients all demonstrating progressive disease at the time of biopsy after a minimum duration of treatment of 6 months. Six tumors had moderate to high tamoxifen levels--2 from patients responding to tamoxifen, 1 from a patient with stable disease and 3 from patients with disease progression. Both the cis and trans isomers of the potent antiestrogenic metabolite 4-hydroxytamoxifen were detected in 11 tumors. Six tumors had high ratios of the eis to trans isomer similar to our observation in the mouse model--all from patients not responding to tamoxifen. The 5 tumors with low cis to trans ratios included 2 tumors from responding patients and 3 from patients with progression. All but 1 of the 11 non-responding patients in this study had either a low tumor tamoxifen level, a high eis/trans ratio, or both. Reduced tamoxifen concentrations in patients with tamoxifen resistance have recently been confirmed in another larger study [23]. A reduction in the concentration of tamoxifen in the tumor or a relative increase in the eis isomer of 4-hydroxytamoxifen alone could not explain tamoxifen-stimulated growth. Using both an HPLC assay as well as mass spectrometry, we investigated the possibility that accumulation of the estrogenic metabolites, metabolite E and bisphenol, might be responsible for tamoxifen resistance in this model. Both metabolites were identified in tumors from our nude mouse model as well as in tumors from patients on tamoxifen lending further support to the hypothesis [24].
86
C. KENT OSBORNE
Use of tamoxifen analogs to investigate mechanisms for tamoxifen resistance To further investigate whether tamoxifen metabolism is responsible for tamoxifen-stimulated growth in our preclinical model, we utilized several tamoxifen analogs and a pure steroidal antiestrogen. Two non-isomerizable 7-membered ring analogs fixed in the trans position were used to determine the importance of isomerization to the development of resistance [25]. The potential importance of conversion to metabolite E or bisphenol was evaluated by using a deoxytamoxifen analog that is resistant to cleavage of the dimethylaminoethoxyside chain. The pure steroidal antiestrogen, ICI 182,780, was also employed since it has a different mechanism of action than tamoxifen and does not undergo a similar metabolic fate [26].
Quantitative levels of tamoxifen and major metabolites in tamoxifen-inhibited versus tamoxifen-stimulated MCF-7 tumors We first quantified levels of tamoxifen and other major metabolites in the tamoxifen-inhibited versus tamoxifen-stimulated MCF-7 tumors. Similar to our previous report, levels of the parent drug tamoxifen were significantly reduced in tamoxifen-stimulated tumors [27]. In general, tumor concentrations varied widely, but tamoxifen-stimulated tumors had a mean concentration of 998 ng/g tissue compared to 15,780ng/g tissue in tamoxifen-inhibited tumors. The levels of N-desmethyltamoxifen and the cis and trans isomers of 4-hydroxytamoxifen were also reduced in tamoxifen-stimulated tumors. Confirming our previous report, the ratio of cis to trans 4-hydroxytamoxifen was higher in stimulated tumors than in inhibited tumors. Metabolite E, an estrogenic metabolite lacking the dimethylaminoethoxy side chain, was also detected in both groups including tumors still being inhibited by tamoxifen. However, the relative concentration of metabolite E compared to tamoxifen itself or compared to the potent antiestrogenic metabolite trans-4-hydroxytamoxifen, was distinctly different in tumors that were being stimulated by tamoxifen treatment. Although the absolute level of metabolite E was also reduced slightly in tamoxifen-stimulated tumors, it was relatively more abundant compared to the parent drug or the other metabolites. In fact, the tamoxifen-metabolite E ratio was 20:1 in these tumors compared to 132:1 in tamoxifen-inhibited tu-
mors. Very similar results were observed when the median of the individual tumor ratios was compared (33:1 compared to 136:1). These differences were highly statistically significant and they lend additional support to the hypothesis that tamoxifen metabolism may contribute to tamoxifen-stimulated growth.
Effect of non-isomerizable tamoxifen analogs This cis isomer of the parent drug tamoxifen and its metabolites display quite different biological activities compared to the trans isomers. Trans-tamoxifen is not known to isomerize under physiologic conditions, but the cis isomer is a full estrogen with little or no antiestrogenic activity[19]. Trans-4-hydroxytamoxifen, a potent antiestrogen, readily and spontaneously isomerizes to the cis isomer, a weak antiestrogen [20]. To investigate further whether isomerization of tamoxifen or its metabolites is important for the development of tamoxifenstimulated growth observed in our experimental model, we employed two fixed-ring non-isomerizable analogs of trans-tamoxifen. Fragments of a tamoxifen-stimulated tumor were transplanted into recipient mice [27]. Fragments failed to form tumors when introduced into ovariectomized mice, unless the mice received estrogen supplementation. As we have previously reported, tamoxifen treatment, like estrogen, stimulated tumor growth, although it was less effective than estrogen [14, 27]. Toremifene, a triphenylethylene antiestrogen similar to tamoxifen, also stimulated tumor growth. Interestingly, both the dimethyl and diethyl fixed-ring analogs had a similar stimulatory effect on tumor growth. Under the same conditions, nafoxidine, another non-steroidal antiestrogen that is structurally distinct from tamoxifen, was just as potent as the other analogs in promoting growth of these tumors. To ensure that the fixed-ring analogs remained intact in vivo, tumor extracts were analyzed by HPLC. A large peak corresponding to the parent drug was identified. In no case were doublet or two closely separated peaks identified that might suggest isomerization of the 4-hydroxy metabolite. Thus, the fixed-ring structure appeared to remain intact in vivo, preventing the formation of cis metabolites.
Effect of deoxytamoxifen on growth of tamoxifen-stimulated tumors Due to the elimination of the oxygen atom and the presence of a carbon-carbon bond,
Tamoxifen resistancein breast cancer deoxytamoxifen is relatively resistant to cleavage of the dimethylaminoethoxy side chain necessary for antiestogenic activity. When mice transplanted with tamoxifen-stimulated tumors were treated with the deoxy analog, tumor growth stimulation nearly identical to tamoxifen itself was observed [27]. When extracts of several of these tumors were analyzed by HPLC, peaks comigrating in the region of the cis and trans-4-hydroxy metabolites were observed, but no peaks comigrating with the metabolite E or bisphenol standards could be detected. Thus, the deoxy tamoxifen analog, that is resistant to side chain cleavage, was still able to stimulate growth of these tumors.
Effects of the pure steroidal antiestrogen ICI t 82, 780 Gottardis et al. [28] reported in their model that the pure steroidal antiestrogen, ICI 164,384, blocked growth of a serially-transplanted tamoxifen-stimulated tumor. We have confirmed this observation using a new steroidal antiestrogen (ICI 182,780). Treatment of ovariectomized mice transplanted with a fragment from a tamoxifen-stimulated tumor with ICI 182,780 alone or with the oil vehicle alone failed to stimulate tumor growth [27]. ICI 182,780, in fact, inhibited tamoxifen and toremifene-stimulated tumor growth. This drug also inhibited estrogen-induced tumor growth and growth stimulated by deoxytamoxifen. These data suggest that stimulation of tumor growth by tamoxifen and related analogs is mediated through ER and that by virtue of a different mechanism of action, pure steroidal antiestrogens may be useful in reversing this form of tamoxifen resistance.
Agonist and antagonist properties of tamoxifen If tamoxifen-stimulated growth is due to the conversion of tamoxifen, by whatever mechanism, to a pure estrogen agonist, then combinations of tamoxifen with a suboptimal concentration of estrogen might be expected to have an additive effect on stimulating tumor growth. To test this hypothesis, tamoxifenstimulated tumor fragments were transplanted into fresh recipient castrated mice treated with vehicle alone or with optimal or suboptimal doses of estrogen without or with tamoxifen [27]. Again, tumors failed to grow without hormone supplementation. Tamoxifen also stimulated tumor growth though less effectively than either dose of estrogen. Interestingly, when
87
tamoxifen was combined with estrogen, antagonistic, not additive properties were observed. Thus, tamoxifen had a dual effect, capable of stimulating tumor growth when used alone, but still able to antagonize estrogen when the drugs are used together. SUMMARYAND CONCLUSIONS The data using analogs of tamoxifen that are resistant to isomerization and/or cleavage of the side chain necessary for its antagonist properties do not support the hypothesis that metabolism of tamoxifen is responsible for tamoxifen-stimulated growth. The two seven-membered ring analogs fixed in the trans configuration were just as potent as tamoxifen itself in stimulating MCF-7 tumor growth. Thus, isomerization of tamoxifen to its estrogenic cis configuration, the production of the potent estrogen, cis-metabolite E, or conversion of the potent antagonist trans-4-hydroxytamoxifen to the weak cis isomer, seem unlikely to contribute to this form of tamoxifen resistance. We cannot exclude the possibility from our data, however, that the fixed-ring compounds are still flexible enough to have altered receptor binding profiles or biologic activities under certain in vivo conditions. The deoxytamoxifen analog studies also reject the idea that metabolism to metabolite E is responsible for the estrogenic effects of tamoxifen in our model. We found no evidence of metabolite E in tumors from mice treated with the deoxy analog. Yet, this compound was as effective as tamoxifen in promoting growth of a tamoxifen-stimulated tumor transplant. Taken together, our studies implicate mechanisms other than metabolism for tamoxifen-stimulated tumor growth in this model. The altered metabolite profile may simply be a marker for the development of this tumor phenotype. Another reproducible observation with this model of tamoxifen resistance is that tumors, harvested during the phase of tamoxifen-inhibited growth contain 10-15 times as much of the parent drug as do tumors harvested at the time of progression. The mechanism responsible for this reduction in tumor tamoxifen concentration in tamoxifen-stimulated tumors remains obscure. These tumors do not overexpress P-glycoprotein, but an active efflux mechanism is an interesting possibility in view of the persistently high serum concentrations of tamoxifen in these mice. Whether the reduced tumor tamoxifen level is related somehow to the development of
88
C. KENT OSBORNE
tamoxifen-stimulated growth is not clear. Supportive evidence, however, is provided by recent reports demonstrating that a significant proportion of patients with acquired tamoxifen resistance have a marked reduction in tamoxifen concentration in their tumors compared to those with de novo resistance [22, 23]. The data presented here confirm and extend a recent publication using a fixed-ring tamoxifen analog in a serially transplanted model of tamoxifen-stimulated tumor growth[29]. This model differs from ours in that it uses a serially transplanted MCF-7 tumor and a different method of tamoxifen administration which resuits in significantly lower serum and tumor tamoxifen concentrations. Similar to our data, a non-isomerizable tamoxifen analog was as effective as tamoxifen in stimulating tumor growth. Taken together, the observations that non-isomerizable antiestrogens such as nafoxidine and the fixed-ring tamoxifen analogs, as well as the deoxytamoxifen analog resistant to side-chain cleavage are all capable of enhancing in vivo growth of tamoxifen-stimulated tumors, suggests that isomerization and/or metabolism of tamoxifen are not key elements for the development of resistance in this model. The data presented do support the idea that tamoxifen-stimulated growth in our model is mediated by the ER. These tumors continue to express estrogen and progesterone receptor [17], and preliminary sequencing studies demonstrate that the A/B and E regions of the receptor, those domains involved in ligand binding and transcriptional activation, are normal in tamoxifenstimulated tumors (Fuqua S. and Osborne K., unpublished observation). Finally, pure steroidal antiestrogens such as ICI 182,780 are effective inhibitors of tamoxifen-stimulated growth. Clinical investigation of these antiestrogens in patients with tamoxifen resistance will be of interest. The mechanisms by which tamoxifen acquires the ability to stimulate tumor growth after a period of inhibition remains a mystery. It has long been known that tamoxifen has dominant estrogen agonist properties in some tissues or in some species, while antagonistic properties dominate in others such as the breast [30]. In our MCF-7 tumor model estrogen antagonistic properties dominate initially and tumor growth is inhibited. Thereafter, tumor progression occurs as the agonist activity of the drug increases, although antagonistic properties of the drug are still preserved. It is interesting to speculate that
changes in accessory proteins that affect transcriptional activation through ER contribute to tamoxifen-stimulated growth, a possibility that will require further study. Acknowledgements--This work was supported in part by NIH Grants RO1 CA 30251 and P50 CA58183.
REFERENCES 1. Early Breast Cancer Trialists Collaborative Group: Systemic treatment of early breast cancer by hormonal, systemic or immune therapy: 133 randomized trials involving 31,000 recurrences and 24,000 deaths among 75,000 women. Lancet 339 (1992) 1-5, 71-85. 2. Saez R. A. and Osborne C. K.: Hormonal treatment of advanced breast cancer. In Current Clinical Oncology (Edited by B. J. Kennedy). Alan R. Liss, Inc., New York (1989) pp. 163-172. 3. Encarnacion C. A., Ciocca D. R., McGuire W. L., Clark G. M., Fuqua S. A. W. and Osborne C. K.: Measurement of steroid hormone receptors in breast cancer patients on tamoxifen. Breast Cancer Res. Treat. 26 (1993) 237-246. 4. Manni A. and Arafah B. M.: Tamoxifen-induced remission in breast cancer by escalating the dose to 40 mg daily after progression on 20 mg daily: A case report and review of the literature. Cancer 48 (1981) 873-875. 5. Taylor S. G. IV, Gelman R. S., Falkson G. and Cummings F. J.: Combination chemotherapy compared to tamoxifen as initial therapy for stage IV breast cancer in elderly women. Ann. Int. Med. 104 (1986) 455-461. 6. Pritchard K. I., Thomson D. B., Myers R. E., Sutherland J. A., Mobbs B. G. and Meaken J. W.: Tamoxifen therapy in premenopausal patients with metastatic breast cancer. Cancer Treat. Rep. 64 (1980) 787--796. 7. Hoogstraten B., Gad-EI-Mawla N., Maloney T. R., Fletcher W. S., Vaughn C. B., Tranum B. L., Athens J. W., Costanzi J. J. and Foulkes M.: Combined modality therapy for first recurrence of breast cancer. A Southwest Oncology Group Study. Cancer 54 (1984) 224-256. 8. White T. J., Metzger D. and Chambon P.: Expression and function of the human estrogen receptor in yeast. In Cold Spring Harbor Symposia on Quantitative Biology. Cold Spring Harbor Laboratory, Cold Spring Harbor, Vol. LIII (1988) p. 819. 9. Graham M. L. II, Krett N. L., Miller L. A., Leslie K. A., Gordon D. F., Wood W. M., Wei L. L. and Horwitz K. B.: T47Dco cells, genetically unstable and containing estrogen receptor mutations, are a model for the progression of breast cancer to hormone resistance. Cancer Res. 50 (1990) 6208-6217. 10. Raam S., Robert N., Pappas C. H., and Tamura H.: Defective estrogen receptors in human mammary cancers: their significance in defining hormone dependence. J. Natn. Cancer Inst. 811 (1988) 756 761. 11. Scott G. K., Kushner P., Vigne J. L. and Benz C. C.: Truncated forms of DNA-binding estrogen receptors in human breast cancer. J. Clin. Invest. 88 (1991) 700-706. 12. Foster B. D., Cavener D. R. and Parl F. F.: Binding analysis of the estrogen receptor to its specific DNA target site in human breast cancer. Cancer Res. 51 (1991) 3405-3410. 13. Murphy L. C. and Dozlaw H.: Regulation of transforming growth factor 0t and transforming growth factor fl messenger ribonucleic acid abundance in T-47D human breast cancer cells. Molec. Endocr. 3 (1989) 611 617. 14. Fnqua S. A. W., Fitzgerald S. D., Chamness G. C., Tandon A. K., McDonnell D. O., Nawaz Z., O'Malley B. W. and McGuire W. L.: A variant human breast
Tamoxifen resistance in breast cancer
15.
16.
17.
18.
19.
20.
21. 22.
tumor estrogen receptor with constitutive transcriptional activity. Cancer Res. 51 (1991) 105-109. Fuqua S. A. W., Fitzgerald S. D., AUred D. C., Elledge R. M., Nawaz Z., McDonnell D. P., O'Malley B. W., Greene G. L. and McGuire W. L.: Inhibition of estrogen receptor action by a naturally occurring variant in human breast cancer. Cancer Res. 52 (1992) 483-486. Osborne C. K., Coronado E. B. and Robinson J. P.: Human breast cancer in the athymic nude mouse: Cytostatic effects of long-term antiestrogen therapy. Eur. J. Cancer Clin. Oncol. 23 (1987) 118-196. Osborne C. K., Coronado E., Allred D. C. and Wiebe V., DeGregorio M.: Acquired tamoxifen resistance: correlation with reduced tumor levels of tamoxifen and isomerization of trans-4-hydroxytamoxifen. J. Nam. Cancer Inst. 83 (1991) 147%1482. Gottardis M. M. and Jordan V. C.: Development of tamoxifen-stimulated growth of MCF-7 tumors in athymic mice after long-term antiestrogen administration. Cancer Res. 48 (1988) 5183-5187. Buckley M. M.-T. and Goa K. L.: Tamoxifen: A reappraisal of its pharmacodynamic and pharmacokinetic properties and therapeutic use. Drugs 37 (1989) 451-490. Katzenellenbogen J. A., Carlson K. E. and Katzcnellenbogen B. S.: Facile geometric isomerization of pehnolic non-steroidal estrogens and antiestrogens: Limitations to the interpretation of experiments characterizing the activity of individual isomers. J. Steroid Biochem. 22 (1985) 589-596. Lyman S. D. and Jordan V. C.: Metabolism of tamoxifen and its uterotrophic activity. Biochem. Pharmac. 34 (1985) 2787-2794. Osborne C. K., Wiebe V. J., McGuire W. L., Ciocca D. R. and DeGregorio M. W.: Tamoxifen and the isomers of 4-hydroxytamoxifen in tamoxifen-resistant
23.
24.
25.
26. 27.
28.
29.
30.
89
tumors from breast cancer patients. J. Clin. Oncol. 10 (1992) 304-310. Dowser M., Johnston S. R. D., Haynes B. P , Sacks N. P. M., Baum M. B., Ebbs S. R. and Smith I. E.: Impaired intra-tumoral uptake as a mechanism for resistance in human breast cancer. Proc. Am. Assoc. Cancer Res. 34 (1993) 1463 (Abs.). Wiebe V. J., Osborne C. K., McGuire W. L. and DeGregorio M. W.: Identification of estrogenic tamoxifen metabolite(s) in tamoxiren-resistant human breast tumors. J. Clin. OncoL 10 (1992) 990-994. McCague R., Leclercq G. and Jordan V. C.: Nonisomerizable analogues of (Z)- and (E)- 4-hydroxytamoxifen. Synthesis and endocrinological properties of substituted diphenylbenzocycloheptenes. J. Med. Chem. 31 (1988) 1284-1290. Wakeling A. E.: The future of new pure antiestrogens in clinical breast cancer. Breast Cancer Res. Treat. 25 (1993) 1-9. Osborne C. K., Jarman M., McCague R., Coronado E. B., DeGregorio M. W., Hilsenbeck S. G. and Wakeling A. E.: The importance of tamoxifen metabolism in tamoxifen-stimulated breast tumor growth. Cancer Chemother. Pharmac. In press. Gottardis M. M., Jiang S.-Y., Jeng M.-H. and Jordan V. C.: Inhibition of tamoxifen-stimulated growth of an MCF-7 tumor variant in athymic mice by novel steroidal antiestrogens. Cancer Res. 49 (1989) 4090-4093. Wolf D. M., Langan-Fahey S. M., Parker C. J., McCague R. and Jordan V. C.: Investigation of the mechanism of tamoxifen-stimulated breast tumor growth with nonisomerizable analogues of tamoxifen and metabolites. J. Natn. Cancer Inst. 85 (1993) 806-812. Buckley M. M. and Goa K. L.: Tamoxifen: A reappraisal of its pharmacodynamic and pharmacokinetic properties and therapeutic use. Drugs 37 (1989) 451-490.
0960-0760/93 $6.00 + 0.00 Copyright © 1993 Pergamon Press Ltd
Printed in Great Britain. All fights reserved
MECHANISMS FOR TAMOXIFEN RESISTANCE IN BREAST CANCER: POSSIBLE ROLE OF TAMOXIFEN METABOLISM C. KENT OSBORNE Department of Medicine, Division of Medical Ontology, University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, TX 78284-7884, U.S.A.
Summary--Potential mechanisms for tamoxifen resistance include loss or alteration in estrogen receptor or other transcription factors and altered tamoxifen pharmacology. Using an experimental model, we have previously demonstrated that one form of tamoxifen resistance is related to the acquired ability of tamoxifen to stimulate tumor growth. These tamoxifen-stimulated tumors contain a reduced tamoxifen concentration and an altered metabolite profile suggesting that accumulation of more estrogenic metabolites could explain this phenomenon. However, /n vivo treatment of nude mice carrying tamoxifen-stimulated tumors with fixed ring non-isomerizable analogs, or other analogs resistant to conversion to metabolite E (a full estrogen), still resulted in tumor growth stimulation. Growth of these tamoxifen-stimulated tumors was inhibited by a pure steroidal antiestrogen, ICI 182,780, suggesting that this drug should be investigated in patients with tamoxifen resistance. These tamoxifen-stimulated tumors could be further stimulated by estrogen replenishment, and estrogen stimulation was blocked by tamoxifen, indicating that tamoxifen has both agonist and antagonist properties in these tumors. Our data suggest that although tamoxifen-stimulated tumors display a markedly altered metabolite profile, isomerization or metabolism of tamoxifen does not fully explain the development of tamoxifen-stimulated growth. The mechanisms by which tamoxifen acquires more potent in vivo agonist properties over time remains to be defined.
for preventing or reversing the emergence of resistant cells.
INTRODUCTION
Tamoxifen is a non-steroidal antiestrogen which is now the most frequently used drug in the treatment of breast cancer. The primary mode of action of tamoxifen and other antiestrogens is inhibition of estrogen-induced growth by competitive antagonism of the estrogen receptor (ER). The drug effectively prolongs disease-free and overall survival of women treated adjuvantly following primary surgery [1], and it can also induce tumor regression in more than half of patients with metastatic disease who have ER-positive tumors[2]. Although tamoxifen treatment is effective in many patients, still 40 to 50% of patients are resistant initially to tamoxifen therapy, and all patients, even those who initially respond, eventually develop acquired tamoxifen resistance leading to tumor progression. Understanding the mechanisms by which tumors develop resistance to tamoxifen might uncover new strategies
Clinical clues f o r the development o f acquired tamoxifen resistance
Clinical experience with tamoxifen has provided several clues for the mechanisms by which tumors become resistant. Patients who initially respond to tamoxifen but later develop resistance and tumor progression frequently respond to second or third-line endocrine therapies such as an aromatase inhibitor or megestrol acetate [2]. Thus, acquired tamoxifen resistance does not necessarily indicate global hormonal unresponsiveness, but rather selective resistance to tamoxifen itself. Some patients with acquired tamoxifen resistance, however, do develop resistance to all forms of endocrine therapy via selection of an ER-negative tumor cell clone. We have recently evaluated a series of patients with acquired tamoxifen resistance using immunohistochemical as well as ligand binding assays to measure ER and progesterone receptors (PgR)[3]. Immunohistochemical assays were used to circumvent the problem of false negative assays caused by receptor occupancy
Proceedings of the llth International Symposium of the Journal of Steroid Biochemistry and Molecular Biology, Recent Advances in Steroid Biochemistry and Molecular Biology, Seefeld, Austria, 30 May-2 June 1993. SBMB 4711-6--43
83
84
C. KENT OSBOR~
by the drug. Using these methods, more than two thirds of patients expressed ER and/or PgR in their tumors after the development of progression on tamoxifen. These data indicate that mechanisms of resistance other than receptor loss are common. Although it has not been systematically studied, anecdotal experience suggests that some patients with progression on tamoxifen will respond to a rechallenge with tamoxifen after an interval in which they received alternative treatments. Furthermore, patients who received tamoxifen adjuvant therapy and then later developed disease progression frequently respond to a second treatment with the drug. This suggests that tamoxifen resistance in some cases may not be a permanent phenotype but is reversible when the drug is stopped. Patients may also respond to an increase in the tamoxifen dose after developing progression with a lower dose schedule [4]. Finally, as with highdose estrogen therapy, some patients with metastatic disease who have responded to tamoxifen will have a withdrawal response when the drug is stopped at the time of tumor progression [5]. The long half-life of tamoxifen makes it difficult for clinicians to withhold alternative therapy while waiting for a withdrawal response to the drug. Nevertheless, these data strongly suggest that in some patients with acquired resistance, tamoxifen may actually be stimulating tumor growth. Results of two previously published clinical trials also suggest that tamoxifen can behave as an estrogen and stimulate tumor growth at the time of progression in some patients [6, 7]. In these studies of premenopausal women with advanced breast cancer, second line ovarian ablation was performed in patients who initially responded to tamoxifen. In one of these studies, secondary response to ovarian ablation was common in patients who had previously responded to tamoxifen suggesting that tamoxifen treatment served as an /n vivo tumor estrogen sensitivity assay[6]. However, in the other study, contradictory results were obtained and no patient responded to subsequent ovarian ablation[7]. In the latter study, tamoxifen therapy was continued after castration while in the first study, tamoxifen treatment was stopped. No response to ovarian ablation would be expected if tamoxifen itself was behaving as an estrogen agonist and stimulating tumor growth in patients on the second study. Tamoxifen-stimulated tumor growth as a mechanism
for acquired resistance is further supported by data from preclinical experimental models (see below). Altered ER as a mechanism for tamoxifen resistance
Although the ER protein can be detected in many patients with acquired tamoxifen resistance, mutations in the protein or in other transcription factors could conceivably contribute to tamoxifen resistance. Human ER contains at least five functional domains[8]. Transcription activating functions are located in the A/B region and in the E region, which also serves as the hormone binding domain. The C region is involved in DNA binding and the D region is involved with dimerization. Mutations in the C region would render the receptor non-functional by preventing it from binding to estrogen response elements on target genes. Mutations in the E region could alter the transcription activating function or alter binding of estrogen and other ligands. Mutations in the A/B region could also alter the transcription activating activity of the receptor conceivably leading to tamoxifen resistance. While several different mutations in human ER have been identified in cultured human breast cancer cells as well as in human breast cancer specimens, the clinical significance of these altered receptors and whether they are responsible for tamoxifen resistance in some patients remains to be demonstrated [9-15]. Experimental model of tamoxifen resistance
We have developed an in vivo experimental model of tamoxifen resistance using the MCF-7 human breast cancer cells growing as solid tumors in athymic nude mice [16, 17]. Tamoxifen suppresses tumor growth in this model for several months, and then tumor growth resumes despite continued treatment with the drug. Resistance to tamoxifen is not due to the emergence of an ER-negative clone or to alterations in tamoxifen serum levels. Interestingly, we have shown that acquired tamoxifen "resistance" in this model is related to the acquired ability of tamoxifen after several months of treatment to stimulate rather than to inhibit tumor growth. Similar data have been reported by Gottardis and Jordan [18]. These data support the clinical observations described above suggesting that one form of tamoxifen resistance may be due to growth stimulation of the tumor by tamoxifen. Since ER is normal in these tumors, we next
Tamoxifen resistance in breast cancer
investigated tamoxifen pharmacology and metabolism as a possible mechanism for tamoxifen-stimulated growth. Tamoxifen pharmacology and metabolism Tamoxifen exists in both a trans and a cis configuration [19]. The trans isomer has potent antiestrogenic properties but the cis isomer is a full estrogen agonist. Pharmaceutically manufactured tamoxifen is the trans isomer. In humans the major metabolites of tamoxifen include N-desmethyltamoxifen and 4-hydroxytamoxifen. Although trans-4-hydroxytamoxifen quantitatively is a minor metabolite, it binds to ER with much higher affinity than the parent drug. Thus, it is a more potent metabolite and it may contribute significantly to the antiestrogenic effects of tamoxifen. Trans-4-hydroxytamoxifen can also spontaneously isomerize to cis-4-hydroxytamoxifen, a much less potent antiestrogen [20]. Two other metabolites in which the dimethylaminoethoxy side chain of tamoxifen has been cleaved (metabolite E and bisphenol) are potent estrogen agonists [21]. Metabolism either in the host or in the tumor to less antiestrogenic or more estrogenic metabolites could lead to tamoxifenstimulated growth and tamoxifen "resistance" in the patient. To further investigate the mechanism for tamoxifen-stimulated tumor growth in our experimental model, we compared levels of tamoxifen and several of its metabolites in both the tamoxifen-inhibited and -stimulated tumors[17]. In our initial studies, we found that tumor extracts from tamoxifen-stimulated tumors were characterized by a 10-fold lower tamoxifen concentration and by isomerization of trans-4-hydroxytamoxifen to the less potent cis isomer. A similar pattern of reduced tamoxifen concentration and a relative increase in cis-4-hydroxytamoxifen was found in both cytosol as well as nuclear extracts. The mechanism for this reduced tamoxifen concentration in tamoxifen-resistant tumors is unknown but, as described below, it has now been reproduced in two preliminary studies from patients with tamoxifen resistance [22, 23]. The fact that serum levels from mice with tamoxifen-resistant tumors remain normal while tumor levels are markedly reduced suggests the possibility of an efflux mechanism that reduces net uptake of tamoxifen by the tumor. Since tamoxifen is known to interact with P-glycoprotein, we evaluated its ex-
85
pression at both the RNA and protein level in tamoxifen-stimulated tumors. However, P-glycoprotein was not overexpressed in any of these tumors suggesting that it is not involved in the resistance phenomenon (unpublished data). To further investigate the hypothesis that metabolic tolerance as manifested by decreasing accumulation of tamoxifen in the tumor, by isomerization of the potent antiestrogenic metabolite trans-4-hydroxytamoxifen to the less potent cis isomer, and by accumulation of estrogenic metabolites is a cause of tamoxifenstimulated growth and clinical tamoxifen resistance, we measured tumor concentrations of tamoxifen and several of its metabolites in a small series of patients who were on tamoxifen at the time of the tissue biopsy [22]. Tumor tamoxifen levels varied over a wide range. Low concentrations were observed in tumors from 8 patients all demonstrating progressive disease at the time of biopsy after a minimum duration of treatment of 6 months. Six tumors had moderate to high tamoxifen levels--2 from patients responding to tamoxifen, 1 from a patient with stable disease and 3 from patients with disease progression. Both the cis and trans isomers of the potent antiestrogenic metabolite 4-hydroxytamoxifen were detected in 11 tumors. Six tumors had high ratios of the eis to trans isomer similar to our observation in the mouse model--all from patients not responding to tamoxifen. The 5 tumors with low cis to trans ratios included 2 tumors from responding patients and 3 from patients with progression. All but 1 of the 11 non-responding patients in this study had either a low tumor tamoxifen level, a high eis/trans ratio, or both. Reduced tamoxifen concentrations in patients with tamoxifen resistance have recently been confirmed in another larger study [23]. A reduction in the concentration of tamoxifen in the tumor or a relative increase in the eis isomer of 4-hydroxytamoxifen alone could not explain tamoxifen-stimulated growth. Using both an HPLC assay as well as mass spectrometry, we investigated the possibility that accumulation of the estrogenic metabolites, metabolite E and bisphenol, might be responsible for tamoxifen resistance in this model. Both metabolites were identified in tumors from our nude mouse model as well as in tumors from patients on tamoxifen lending further support to the hypothesis [24].
86
C. KENT OSBORNE
Use of tamoxifen analogs to investigate mechanisms for tamoxifen resistance To further investigate whether tamoxifen metabolism is responsible for tamoxifen-stimulated growth in our preclinical model, we utilized several tamoxifen analogs and a pure steroidal antiestrogen. Two non-isomerizable 7-membered ring analogs fixed in the trans position were used to determine the importance of isomerization to the development of resistance [25]. The potential importance of conversion to metabolite E or bisphenol was evaluated by using a deoxytamoxifen analog that is resistant to cleavage of the dimethylaminoethoxyside chain. The pure steroidal antiestrogen, ICI 182,780, was also employed since it has a different mechanism of action than tamoxifen and does not undergo a similar metabolic fate [26].
Quantitative levels of tamoxifen and major metabolites in tamoxifen-inhibited versus tamoxifen-stimulated MCF-7 tumors We first quantified levels of tamoxifen and other major metabolites in the tamoxifen-inhibited versus tamoxifen-stimulated MCF-7 tumors. Similar to our previous report, levels of the parent drug tamoxifen were significantly reduced in tamoxifen-stimulated tumors [27]. In general, tumor concentrations varied widely, but tamoxifen-stimulated tumors had a mean concentration of 998 ng/g tissue compared to 15,780ng/g tissue in tamoxifen-inhibited tumors. The levels of N-desmethyltamoxifen and the cis and trans isomers of 4-hydroxytamoxifen were also reduced in tamoxifen-stimulated tumors. Confirming our previous report, the ratio of cis to trans 4-hydroxytamoxifen was higher in stimulated tumors than in inhibited tumors. Metabolite E, an estrogenic metabolite lacking the dimethylaminoethoxy side chain, was also detected in both groups including tumors still being inhibited by tamoxifen. However, the relative concentration of metabolite E compared to tamoxifen itself or compared to the potent antiestrogenic metabolite trans-4-hydroxytamoxifen, was distinctly different in tumors that were being stimulated by tamoxifen treatment. Although the absolute level of metabolite E was also reduced slightly in tamoxifen-stimulated tumors, it was relatively more abundant compared to the parent drug or the other metabolites. In fact, the tamoxifen-metabolite E ratio was 20:1 in these tumors compared to 132:1 in tamoxifen-inhibited tu-
mors. Very similar results were observed when the median of the individual tumor ratios was compared (33:1 compared to 136:1). These differences were highly statistically significant and they lend additional support to the hypothesis that tamoxifen metabolism may contribute to tamoxifen-stimulated growth.
Effect of non-isomerizable tamoxifen analogs This cis isomer of the parent drug tamoxifen and its metabolites display quite different biological activities compared to the trans isomers. Trans-tamoxifen is not known to isomerize under physiologic conditions, but the cis isomer is a full estrogen with little or no antiestrogenic activity[19]. Trans-4-hydroxytamoxifen, a potent antiestrogen, readily and spontaneously isomerizes to the cis isomer, a weak antiestrogen [20]. To investigate further whether isomerization of tamoxifen or its metabolites is important for the development of tamoxifenstimulated growth observed in our experimental model, we employed two fixed-ring non-isomerizable analogs of trans-tamoxifen. Fragments of a tamoxifen-stimulated tumor were transplanted into recipient mice [27]. Fragments failed to form tumors when introduced into ovariectomized mice, unless the mice received estrogen supplementation. As we have previously reported, tamoxifen treatment, like estrogen, stimulated tumor growth, although it was less effective than estrogen [14, 27]. Toremifene, a triphenylethylene antiestrogen similar to tamoxifen, also stimulated tumor growth. Interestingly, both the dimethyl and diethyl fixed-ring analogs had a similar stimulatory effect on tumor growth. Under the same conditions, nafoxidine, another non-steroidal antiestrogen that is structurally distinct from tamoxifen, was just as potent as the other analogs in promoting growth of these tumors. To ensure that the fixed-ring analogs remained intact in vivo, tumor extracts were analyzed by HPLC. A large peak corresponding to the parent drug was identified. In no case were doublet or two closely separated peaks identified that might suggest isomerization of the 4-hydroxy metabolite. Thus, the fixed-ring structure appeared to remain intact in vivo, preventing the formation of cis metabolites.
Effect of deoxytamoxifen on growth of tamoxifen-stimulated tumors Due to the elimination of the oxygen atom and the presence of a carbon-carbon bond,
Tamoxifen resistancein breast cancer deoxytamoxifen is relatively resistant to cleavage of the dimethylaminoethoxy side chain necessary for antiestogenic activity. When mice transplanted with tamoxifen-stimulated tumors were treated with the deoxy analog, tumor growth stimulation nearly identical to tamoxifen itself was observed [27]. When extracts of several of these tumors were analyzed by HPLC, peaks comigrating in the region of the cis and trans-4-hydroxy metabolites were observed, but no peaks comigrating with the metabolite E or bisphenol standards could be detected. Thus, the deoxy tamoxifen analog, that is resistant to side chain cleavage, was still able to stimulate growth of these tumors.
Effects of the pure steroidal antiestrogen ICI t 82, 780 Gottardis et al. [28] reported in their model that the pure steroidal antiestrogen, ICI 164,384, blocked growth of a serially-transplanted tamoxifen-stimulated tumor. We have confirmed this observation using a new steroidal antiestrogen (ICI 182,780). Treatment of ovariectomized mice transplanted with a fragment from a tamoxifen-stimulated tumor with ICI 182,780 alone or with the oil vehicle alone failed to stimulate tumor growth [27]. ICI 182,780, in fact, inhibited tamoxifen and toremifene-stimulated tumor growth. This drug also inhibited estrogen-induced tumor growth and growth stimulated by deoxytamoxifen. These data suggest that stimulation of tumor growth by tamoxifen and related analogs is mediated through ER and that by virtue of a different mechanism of action, pure steroidal antiestrogens may be useful in reversing this form of tamoxifen resistance.
Agonist and antagonist properties of tamoxifen If tamoxifen-stimulated growth is due to the conversion of tamoxifen, by whatever mechanism, to a pure estrogen agonist, then combinations of tamoxifen with a suboptimal concentration of estrogen might be expected to have an additive effect on stimulating tumor growth. To test this hypothesis, tamoxifenstimulated tumor fragments were transplanted into fresh recipient castrated mice treated with vehicle alone or with optimal or suboptimal doses of estrogen without or with tamoxifen [27]. Again, tumors failed to grow without hormone supplementation. Tamoxifen also stimulated tumor growth though less effectively than either dose of estrogen. Interestingly, when
87
tamoxifen was combined with estrogen, antagonistic, not additive properties were observed. Thus, tamoxifen had a dual effect, capable of stimulating tumor growth when used alone, but still able to antagonize estrogen when the drugs are used together. SUMMARYAND CONCLUSIONS The data using analogs of tamoxifen that are resistant to isomerization and/or cleavage of the side chain necessary for its antagonist properties do not support the hypothesis that metabolism of tamoxifen is responsible for tamoxifen-stimulated growth. The two seven-membered ring analogs fixed in the trans configuration were just as potent as tamoxifen itself in stimulating MCF-7 tumor growth. Thus, isomerization of tamoxifen to its estrogenic cis configuration, the production of the potent estrogen, cis-metabolite E, or conversion of the potent antagonist trans-4-hydroxytamoxifen to the weak cis isomer, seem unlikely to contribute to this form of tamoxifen resistance. We cannot exclude the possibility from our data, however, that the fixed-ring compounds are still flexible enough to have altered receptor binding profiles or biologic activities under certain in vivo conditions. The deoxytamoxifen analog studies also reject the idea that metabolism to metabolite E is responsible for the estrogenic effects of tamoxifen in our model. We found no evidence of metabolite E in tumors from mice treated with the deoxy analog. Yet, this compound was as effective as tamoxifen in promoting growth of a tamoxifen-stimulated tumor transplant. Taken together, our studies implicate mechanisms other than metabolism for tamoxifen-stimulated tumor growth in this model. The altered metabolite profile may simply be a marker for the development of this tumor phenotype. Another reproducible observation with this model of tamoxifen resistance is that tumors, harvested during the phase of tamoxifen-inhibited growth contain 10-15 times as much of the parent drug as do tumors harvested at the time of progression. The mechanism responsible for this reduction in tumor tamoxifen concentration in tamoxifen-stimulated tumors remains obscure. These tumors do not overexpress P-glycoprotein, but an active efflux mechanism is an interesting possibility in view of the persistently high serum concentrations of tamoxifen in these mice. Whether the reduced tumor tamoxifen level is related somehow to the development of
88
C. KENT OSBORNE
tamoxifen-stimulated growth is not clear. Supportive evidence, however, is provided by recent reports demonstrating that a significant proportion of patients with acquired tamoxifen resistance have a marked reduction in tamoxifen concentration in their tumors compared to those with de novo resistance [22, 23]. The data presented here confirm and extend a recent publication using a fixed-ring tamoxifen analog in a serially transplanted model of tamoxifen-stimulated tumor growth[29]. This model differs from ours in that it uses a serially transplanted MCF-7 tumor and a different method of tamoxifen administration which resuits in significantly lower serum and tumor tamoxifen concentrations. Similar to our data, a non-isomerizable tamoxifen analog was as effective as tamoxifen in stimulating tumor growth. Taken together, the observations that non-isomerizable antiestrogens such as nafoxidine and the fixed-ring tamoxifen analogs, as well as the deoxytamoxifen analog resistant to side-chain cleavage are all capable of enhancing in vivo growth of tamoxifen-stimulated tumors, suggests that isomerization and/or metabolism of tamoxifen are not key elements for the development of resistance in this model. The data presented do support the idea that tamoxifen-stimulated growth in our model is mediated by the ER. These tumors continue to express estrogen and progesterone receptor [17], and preliminary sequencing studies demonstrate that the A/B and E regions of the receptor, those domains involved in ligand binding and transcriptional activation, are normal in tamoxifenstimulated tumors (Fuqua S. and Osborne K., unpublished observation). Finally, pure steroidal antiestrogens such as ICI 182,780 are effective inhibitors of tamoxifen-stimulated growth. Clinical investigation of these antiestrogens in patients with tamoxifen resistance will be of interest. The mechanisms by which tamoxifen acquires the ability to stimulate tumor growth after a period of inhibition remains a mystery. It has long been known that tamoxifen has dominant estrogen agonist properties in some tissues or in some species, while antagonistic properties dominate in others such as the breast [30]. In our MCF-7 tumor model estrogen antagonistic properties dominate initially and tumor growth is inhibited. Thereafter, tumor progression occurs as the agonist activity of the drug increases, although antagonistic properties of the drug are still preserved. It is interesting to speculate that
changes in accessory proteins that affect transcriptional activation through ER contribute to tamoxifen-stimulated growth, a possibility that will require further study. Acknowledgements--This work was supported in part by NIH Grants RO1 CA 30251 and P50 CA58183.
REFERENCES 1. Early Breast Cancer Trialists Collaborative Group: Systemic treatment of early breast cancer by hormonal, systemic or immune therapy: 133 randomized trials involving 31,000 recurrences and 24,000 deaths among 75,000 women. Lancet 339 (1992) 1-5, 71-85. 2. Saez R. A. and Osborne C. K.: Hormonal treatment of advanced breast cancer. In Current Clinical Oncology (Edited by B. J. Kennedy). Alan R. Liss, Inc., New York (1989) pp. 163-172. 3. Encarnacion C. A., Ciocca D. R., McGuire W. L., Clark G. M., Fuqua S. A. W. and Osborne C. K.: Measurement of steroid hormone receptors in breast cancer patients on tamoxifen. Breast Cancer Res. Treat. 26 (1993) 237-246. 4. Manni A. and Arafah B. M.: Tamoxifen-induced remission in breast cancer by escalating the dose to 40 mg daily after progression on 20 mg daily: A case report and review of the literature. Cancer 48 (1981) 873-875. 5. Taylor S. G. IV, Gelman R. S., Falkson G. and Cummings F. J.: Combination chemotherapy compared to tamoxifen as initial therapy for stage IV breast cancer in elderly women. Ann. Int. Med. 104 (1986) 455-461. 6. Pritchard K. I., Thomson D. B., Myers R. E., Sutherland J. A., Mobbs B. G. and Meaken J. W.: Tamoxifen therapy in premenopausal patients with metastatic breast cancer. Cancer Treat. Rep. 64 (1980) 787--796. 7. Hoogstraten B., Gad-EI-Mawla N., Maloney T. R., Fletcher W. S., Vaughn C. B., Tranum B. L., Athens J. W., Costanzi J. J. and Foulkes M.: Combined modality therapy for first recurrence of breast cancer. A Southwest Oncology Group Study. Cancer 54 (1984) 224-256. 8. White T. J., Metzger D. and Chambon P.: Expression and function of the human estrogen receptor in yeast. In Cold Spring Harbor Symposia on Quantitative Biology. Cold Spring Harbor Laboratory, Cold Spring Harbor, Vol. LIII (1988) p. 819. 9. Graham M. L. II, Krett N. L., Miller L. A., Leslie K. A., Gordon D. F., Wood W. M., Wei L. L. and Horwitz K. B.: T47Dco cells, genetically unstable and containing estrogen receptor mutations, are a model for the progression of breast cancer to hormone resistance. Cancer Res. 50 (1990) 6208-6217. 10. Raam S., Robert N., Pappas C. H., and Tamura H.: Defective estrogen receptors in human mammary cancers: their significance in defining hormone dependence. J. Natn. Cancer Inst. 811 (1988) 756 761. 11. Scott G. K., Kushner P., Vigne J. L. and Benz C. C.: Truncated forms of DNA-binding estrogen receptors in human breast cancer. J. Clin. Invest. 88 (1991) 700-706. 12. Foster B. D., Cavener D. R. and Parl F. F.: Binding analysis of the estrogen receptor to its specific DNA target site in human breast cancer. Cancer Res. 51 (1991) 3405-3410. 13. Murphy L. C. and Dozlaw H.: Regulation of transforming growth factor 0t and transforming growth factor fl messenger ribonucleic acid abundance in T-47D human breast cancer cells. Molec. Endocr. 3 (1989) 611 617. 14. Fnqua S. A. W., Fitzgerald S. D., Chamness G. C., Tandon A. K., McDonnell D. O., Nawaz Z., O'Malley B. W. and McGuire W. L.: A variant human breast
Tamoxifen resistance in breast cancer
15.
16.
17.
18.
19.
20.
21. 22.
tumor estrogen receptor with constitutive transcriptional activity. Cancer Res. 51 (1991) 105-109. Fuqua S. A. W., Fitzgerald S. D., AUred D. C., Elledge R. M., Nawaz Z., McDonnell D. P., O'Malley B. W., Greene G. L. and McGuire W. L.: Inhibition of estrogen receptor action by a naturally occurring variant in human breast cancer. Cancer Res. 52 (1992) 483-486. Osborne C. K., Coronado E. B. and Robinson J. P.: Human breast cancer in the athymic nude mouse: Cytostatic effects of long-term antiestrogen therapy. Eur. J. Cancer Clin. Oncol. 23 (1987) 118-196. Osborne C. K., Coronado E., Allred D. C. and Wiebe V., DeGregorio M.: Acquired tamoxifen resistance: correlation with reduced tumor levels of tamoxifen and isomerization of trans-4-hydroxytamoxifen. J. Nam. Cancer Inst. 83 (1991) 147%1482. Gottardis M. M. and Jordan V. C.: Development of tamoxifen-stimulated growth of MCF-7 tumors in athymic mice after long-term antiestrogen administration. Cancer Res. 48 (1988) 5183-5187. Buckley M. M.-T. and Goa K. L.: Tamoxifen: A reappraisal of its pharmacodynamic and pharmacokinetic properties and therapeutic use. Drugs 37 (1989) 451-490. Katzenellenbogen J. A., Carlson K. E. and Katzcnellenbogen B. S.: Facile geometric isomerization of pehnolic non-steroidal estrogens and antiestrogens: Limitations to the interpretation of experiments characterizing the activity of individual isomers. J. Steroid Biochem. 22 (1985) 589-596. Lyman S. D. and Jordan V. C.: Metabolism of tamoxifen and its uterotrophic activity. Biochem. Pharmac. 34 (1985) 2787-2794. Osborne C. K., Wiebe V. J., McGuire W. L., Ciocca D. R. and DeGregorio M. W.: Tamoxifen and the isomers of 4-hydroxytamoxifen in tamoxifen-resistant
23.
24.
25.
26. 27.
28.
29.
30.
89
tumors from breast cancer patients. J. Clin. Oncol. 10 (1992) 304-310. Dowser M., Johnston S. R. D., Haynes B. P , Sacks N. P. M., Baum M. B., Ebbs S. R. and Smith I. E.: Impaired intra-tumoral uptake as a mechanism for resistance in human breast cancer. Proc. Am. Assoc. Cancer Res. 34 (1993) 1463 (Abs.). Wiebe V. J., Osborne C. K., McGuire W. L. and DeGregorio M. W.: Identification of estrogenic tamoxifen metabolite(s) in tamoxiren-resistant human breast tumors. J. Clin. OncoL 10 (1992) 990-994. McCague R., Leclercq G. and Jordan V. C.: Nonisomerizable analogues of (Z)- and (E)- 4-hydroxytamoxifen. Synthesis and endocrinological properties of substituted diphenylbenzocycloheptenes. J. Med. Chem. 31 (1988) 1284-1290. Wakeling A. E.: The future of new pure antiestrogens in clinical breast cancer. Breast Cancer Res. Treat. 25 (1993) 1-9. Osborne C. K., Jarman M., McCague R., Coronado E. B., DeGregorio M. W., Hilsenbeck S. G. and Wakeling A. E.: The importance of tamoxifen metabolism in tamoxifen-stimulated breast tumor growth. Cancer Chemother. Pharmac. In press. Gottardis M. M., Jiang S.-Y., Jeng M.-H. and Jordan V. C.: Inhibition of tamoxifen-stimulated growth of an MCF-7 tumor variant in athymic mice by novel steroidal antiestrogens. Cancer Res. 49 (1989) 4090-4093. Wolf D. M., Langan-Fahey S. M., Parker C. J., McCague R. and Jordan V. C.: Investigation of the mechanism of tamoxifen-stimulated breast tumor growth with nonisomerizable analogues of tamoxifen and metabolites. J. Natn. Cancer Inst. 85 (1993) 806-812. Buckley M. M. and Goa K. L.: Tamoxifen: A reappraisal of its pharmacodynamic and pharmacokinetic properties and therapeutic use. Drugs 37 (1989) 451-490.