Compositions and methods for demonstrating secretory immune system regulation of steroid hormone responsive cancer cell growth

Description of equivalent WO 0185210 A2

COMPOSITIONS AND METHODS FOR DEMONSTRATING

SECRETORY IMMUNE SYSTEM REGULATION OF

STEROTD HORMONE RESPONSIVE CANCER CELL GROWTH

CROSS-REFERENCE TO RELATED APPLICATIONS

0005This application claims the benefit under 35 U.S.C. § 119(e) of U.S. Provisional Patent Application Nos. 60/203,314 filed May 10, 2000; 60/208,348 filed May 31, 2000; 60/208,111 filed May 31, 2000; 60/229,071 filed August 30, 2000 and 60/231,273 filed September 8, 2000.

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

0007Research leading to the present invention was supported in part by the federal government under Grant Nos. DAMD17-94-J-4473, DAMD17-98-8337 and DAMD17-99-1-9405 awarded by the Defense Department through the US Army Medical Research and Materiel Command, Breast Cancer Research Program. The United States government may have certain rights in the invention.

0008BACKGROUND OF THE INVENTION Field of the Invention The present invention generally relates to the regulation of steroid hormone responsive cancer cell growth, and more particularly to compositions and in vitro methods and models for demonstrating secretory immune system immunoglobulin regulation of mucosal epithelial cancer cell growth. Description of Related Art STEROID HORMONE RESPONSIVE TUMOR CELL GROWTH

0009L 1896, a physician named Beatson reported in the medical journal Lancet (Beatson GT (1896) Lancet (Part 1, July 11), 104-107 and Lancet (Part 2, July 18), 162-165) that removal of the ovaries from breast cancer patients slowed or stopped the growth of their tumors. As medical science has moved forward, it is now understood that Dr. Beatson had found that the estrogens made by the ovaries promoted the growth of breast cancers. In the 1940s and 1950s, work by Professor Charles Huggins (Huggins CB and Hodges CV (1941) Cancer Res 1, 293-297; Huggins et al. (1941) Arch Surg 43, 209-223) proved that surgical or chemical castration very substantially reduced the growth of prostate cancers. These results indicated that testicular androgens were important promoters of the growth of tumors of this male accessory organ. In subsequent work, researchers have established that estrogens and androgens act on breast and prostate cancer cells via receptors within the cell nucleus (Tsai M-J and O'Malley BW (1994) Annu Rev Biochem 63, 451-486; Evans RE (1988) Science (Wash DC) 240, 889-895). In fact, estrogen receptors are now commonly measured in breast cancer specimens to assist in decisions regarding the most effective therapies for each patient, and chemical and surgical castration are common treatments for prostate cancer. The regulation of estrogen target tissue cell growth has been a topic of dynamic experimental interest for several years (Jensen EV and DeSombre ER (1973) Science (Wash DC) 182, 126-134; O'Malley BW and Means AR (1974) Science (Wash DC) 183, 610-620). Today, it is generally accepted that estrogen interaction with specific nuclear located DNA binding receptors is necessary to initiate critical cell cycle events (Dickson RB and Stancel GM (2000) J Natl Cancer Inst Monogr No 27, 135-145). It is also highly likely that other non-steroid factors are essential participants in this process (Sirbasku DA (1978) Proc Natl Acad Sci USA 75, 3786-3790; Sirbasku DA (1981) Banbury Report 8, 425-443; Dickson RB and Lippman ME (1987) Endocr Rev 8, 29-43; Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52). Many of these new regulators fall into the general class of positive acting substances called growth factors (Gospodarowitz D and Moran JS (1976) Annu Rev Biochem 45, 531-558; Goustin AS et al. (1986) Cancer Res 46, 1015-1029). Simply stated, these agents cause cells to undergo cell division and thereby lead to growth. Because the hallmark of cancer is uncontrolled cell division, understanding these molecules and how they act is of vital importance. Other members of this regulatory family include negative acting agents called growth inhibitors (Rnabbe et al. (1987) Cell 48, 417-428; de Jong JS et al. (1998) J Pathol 184, 44-52). They block cell division, and because of this, are important targets for new anticancer therapies. A great deal of study has focused on cellular site(s) of estrogen action, and various models have been proposed attempting to explain how estrogen participates with these additional factors to regulate growth.

0010The relative merits of positive versus negative regulation of cell growth have been debated (Dickson RB and Lippman ME (1987) Endocr Rev 8, 29-43; Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52). Although the positive direct and positive indirect models (as defined by Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52) have received the most attention, the concept of negative regulation has intrinsic appeal because its loss offers a ready explanation for the uncontrolled replication of cancer cells. Factors that negatively regulate cell proliferation are now classified as members of the "tumor suppressor" family (Sager R (1997) Proc Natl Acad Sci USA 94, 952-955). Defining and understanding this family of intracellular and extracellular growth regulators is a primary focus of current cancer research. A number of years ago, studies were reported which indicated that serum-borne inhibitors, later named "estrocolyones," had an important if not essential role in steroid responsive cell growth (Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52; Soto AM et al. (1992) J Steroid Biochem Mol Biol 43, 703-712; Soto AM et al. (1986) Cancer Res 46, 2271-2275; Soto AM and Sonnenschein C (1984) Biochem Biophys Res Comrnun 111, 1097-1103; Schatz RW et al. (1985) J Cell Physiol 124, 386-390; Soto AM and Sonnenschein C (1985) J Steroid Biochem 23,

001187-94). Estrocolyones appeared to act as estrogen reversible inhibitors of steroid hormone target tissue cell growth. Subsequently, the inhibitor has been variously identified as an unstable M_r70,000 to 80,000 protein (Soto AM et al. (1992) J Steroid Biochem Mol Biol 43, 703-712), the intact serum albumin molecule (Laursen I et al. (1990) Anticancer Res 10, 343-352; Sonnenschein C et al. (1996) J Steroid Biochem Mol Biol 59, 147-154), two domains of serum albumin (Sonnenschein C et al (1996) J Steroid Biochem Mol Biol 59, 147-154), and the plasma steroid carrier protein sex hormone binding globulin (SHBG) (Reese CC et al. (1988) Ann NY Acad Sci 538, 112-121; Fissore F et al. (1994) Steroids 59, 661-667; Fortunati N et al. (1993) J Steroid Biochem Mol Biol 45, 435-444). Other investigators also thought it possible that SHBG, as well as the other major plasma steroid hormone carrier protein' orticosteroid-binding globulin (CBG), were potential growth regulators independent of their steroid hormone binding capacity. This conclusion was based on the fact that specific cellular membrane receptors have been identified for steroid free CBG and SHBG (Hryb DJ et al. (1986) Proc Natl Acad Sci USA 83, 3253-3256; Hryb DJ et al. (1990) J Biol Chem 265, 6048-6054) and that binding of SHBG and CBG to cells caused changes to cell growth mediators such as cyclic AMP and protein kinase A (Rosner W (1990) Endocrine Rev 11, 80-91; Fortunati N et al. (1996) Endocrinology 137, 686-692; Rosner W et al. (1991) J Steroid Biochem Mol Biol 40, 813-820; Nakhla AM et al. 153, 1012-1018; Rosner W (1992) JAndrology 13, 101-106).

0012Nonetheless, the roles of both albumin and SHBG as estrogen reversible serum-borne growth regulators have been challenged by the present Inventor, and others (Soto AM et al. (1992) J Steroid Biochem Mol Biol 43, 703-712; Damassa DA et al. (1991) Endocrinology 129, 75-84). In fact, in one report, SHBG stimulated growth of the androgen responsive ALNA-41 human prostate cancer cell line (Plymate SR et al. (1991) J Steroid Biochem Mol Biol 40, 833-839). In 1997, Sirbasku et al. reported that nearly pure CBG and an approximately 85% homogeneous SHBG-like protein were obtained from horse serum (Sirbasku DA et al. "Serum factor regulation of estrogen responsive mammary tumor cell growth. "Proceedings of the 1997 Meeting of the "Department of Defense Breast Cancer Research Program: An Era of Hope", (Abstract) pp. 739- 740, Washington, D.C., Oct. 31- Nov. 4, 1997) by employing a procedure similar to that described for use with human cord serum (Fernlund P and Laurell C-B (1981) J Steroid Biochem 14, 545- 552). The Fernlund and Laurell procedure was stated to produce human CBG and SHBG in pure or very nearly pure states using cortisol-agarose affinity chromatography at pH 5.5 followed by Phenyl Sepharose™ chromatography at pH 7.4. Under serum-free defined cell culture conditions, the partially purified SHBG-like fraction obtained by Sirbasku et al. demonstrated progressive inhibition of cell growth in a rat mammary tumor cell line (MTW9/PL2) with increasing concentration of the SHBG-like fraction. Addition of 17β-estradiol (E₂) completely reversed even the maximum inhibition. Sirbasku et al. found that the active SHBG-like fraction contained little or no serum albumin as judged by immunological methods and by standard polyacylamide gel electrophoresis in the presence of reducing agents and sodium dodecyl sulfate (SDS-PAGE) (Laemmli UK (1976) Nature (Lond) 227, 680-685). Although the SHBG-like inhibitor displayed certain immunological similarities to SHBG, it was clearly distinguishable from SHBG based on physiologic, physical and biochemical analyses. Despite its first proposal more than fifteen years ago, the purified estrogen reversible serum-borne inhibitor has yet to be described. Sirbasku et al, as well as others (Soto AM et al. (1992) J Steroid Biochem Mol Biol 43, 703-712), has observed that the estrogen reversible inhibitory activity of serum was very labile during isolation by conventional protein purification methods. Other investigators have used a combination of cortisol affinity chromatorgraphy and an ammonium sulfate precipitation to isolate a cell growth inhibitor from human serum. These studies (Tanji M et al. (2000) Anticancer Res. 20, 2779-2783; Tanji M et al. (2000) Anticancer Res. 20, 2785-2789) describe estrogen inhibition of MCF-7 human breast carcinoma cells that had been maintained at least 3 months in serum-free medium, but no estrogenic effect was observed with normally cultured MCF-7 cells (i.e., cells not long term conditioned to serum-free medium). An isolated steroid-binding protein was stated to mediate an estrogen- dependent inhibition of cell growth. Other serum-borne inhibitors also have been separated from whole serum by diethylaminoethyl (DEAE) chromatography (Dell'Aquila ML and Gaffhey EN (1984) J Natl Cancer Inst 73, 397-403). The properties of these inhibitors have not been defined further nor have they been shown to act as estrogen-reversible inhibitors.

0013Carcinogen-induced rat mammary tumors have been studied extensively as models for the in vivo role of hormones in the induction and growth of breast cancer (Welsch CW (1985) Cancer Res 45, 3415-3443). Despite ample evidence of hormone dependence in vivo,- the carcinogen-induced tumors have not yet yielded permanent tissue culture cell lines that show the same responsiveness to steroid hormones in in vitro culture. Typically, cultures initiated from primary tumors very quickly lose hormone responsiveness. Because of this, the earliest endocrine studies were done with organ cultures (Welsch CW and Rivera EM (172) Proc Soc Exp Biol Med 139, 623-626; Lewis D and Hallowes RC (1974) JEndocrinol 62, 225-240; Chan P-C et al. (1976) Proc Soc Exp Biol Med 151, 362-365; Pasteels J-L et al. (1976) Cancer Res 36, 2162-2170) and short-term cultures of dissociated cells (Chan P-C et al. (1976) Proc Soc Exp Biol Med 151, 362-365). Now investigators recognized that those approaches were inadequate. More recently, cell lines have been developed from carcinogen-induced rat mammary tumors (Bennett DC et al. (1978) Cell 15, 283-298; Rudland PS (1987) Cancer Metast Rev 6, 55-83; Webster MK et al. (1990) J Biol Chem 265, 4831-4838; Lichmer RB et al. Cancer Res 51, 5943-5950; Lichtner RB et al. (1995) Oncogene 10, 1823-1832). Although these lines have been useful for investigations related to breast properties, investigators have found that in general they do not display steroid hormone responsiveness in cell culture. To compound the difficulties, most of these lines could not be evaluated for hormone responsiveness in vivo because they were derived from outbred rats. Simply stated, they lack the syngeneic inbred hosts absolutely required for in vivo transplantation. One of the basic tenets of endocrine physiology is that estrogens and androgens cause coordinate growth of several target tissues (Clark JH et al. (1992) In: Williams Textbook of Endocrinology, 8^th Edition, WB Saunders, Philadelphia, pp 35-90). A partial list of estrogen target tissues includes breast, uterus, cervix, vagina, ovary, pituitary, liver, leukocytes and kidney. A partial list of androgen target tissues includes the male reproductive tract (e. g. prostate, epididymus, and testis), kidney, bladder, liver and muscle. Whatever mechanism is proposed to explain sex steroid dependent growth, one would expect it to be applicable to cells from several of the major target tissues.

0014The history of attempts to demonstrate steroid hormone responsive tumor cell growth in culture has led to two important conclusions. First, demonstration of estrogen and androgen responsive cell growth in culture required the presence of hormone deficient/depleted serum. One of the first studies to demonstrate this requirement was done with human breast cancer cells (Page MJ et al. (1983) Cancer Res 43, 1244-1250). Some notable examples of demonstration by others of estrogen responsiveness in serum containing culture include studies with the MCF-7 human breast cancer cells (Lippman ME et al. (1977) Cancer Res 37, 1901-1907; Soto AM and Sonnenschein C (1985; J Steroid Biochem 23, 87-94; Wiese TE et al. (1992) In Vitro Cell Dev Biol 28A, 595-602), the T47D human breast cancer cells (Chalbos D et al. (1982) J Clin Endocrinol Metab 55, 276-283; Schatz RW et al. (1985) J Cell Physiol 124, 386-390; Soto AM et al. (1986) Cancer Res 46, 2271- 2275), the ZR-75-1 human breast cancer cells (Darbre P et al. (1983) Cancer Res 43, 349-355), the GHjCi rat pituitary tumor cells (Amara JF and Dannies PS (1983) Endocrinology 112, 1141-1143), and the H-301 Syrian hamster kidney tumor cell line (Soto AM et al. (1988) Cancer Res 48, 3676- 3680). Two reports have proposed that estrogen responsiveness can be observed in serum-free defined medium with ZR-75-1 cells (Allegra JC and Lippman ME (1978) Cancer Res 38, 3823- 3829; Darbre PD et al. (1984) Cancer Res 44, 2790-2793). However, in both of those studies, the cells were first incubated for several days in medium supplemented with serum before changing to serum-free defined medium conditions. M Ogasawara and DA Sirbasku previously demonstrated that this approach leaves a problematic serum factor "memory" with cells (Ogasawara M and Sirbasku DA (1988) In Vitro Cell Dev Biol 24, 911-920). When completely serum-free defined medium conditions were applied (Barnes D and Sato G (1980) Nature 281, 388-389; Danielpour D et al. (1988) In Vitro Cell Dev Biol 24, 42-52; Karey KP and Sirbasku DA (1988) Cancer Res 48, 4083-4092; Ogasawara M and Sirbasku DA (1988) In Vitro Cell Dev Biol 24, 911-920; Riss TL and Sirbasku DA (1989) In Vitro Cell Dev Biol 25, 136-142), no growth effects of estrogens were observed. Comparison of the observations in serum-free defined medium versus those in medium with serum led to the second important conclusion. Serum contains a mediator(s) that is required for steroid hormone responsiveness in culture. When the mediator is completely purified and defined chemically, its addition to serum-free defined medium will be expected to provide unequivocal confirmation of its role in hormone dependent cell growth. The purification of the serum-borne mediator has been a challenging undertaking. Sirbasku et al. originally proposed that estrogens per se were not mitogenic, but instead caused the production of endocrine, paracrine or autocrine "estromedins" that were themselves the promoters of target tissue cell growth (Sirbasku DA (1978) Proc Natl Acad Sci USA 75, 3786-3790; Sirbasku DA (1981) Banbury Report 8, 425-443; J- eda T et al. (1982) In Vitro 18, 961-979; Sirbasku DA and Leland FE (1982) Biochemical Action of Hormones 9, 115-140; Leland FE et al. In: Cold Spring Harbor Conferences on Cell Proliferation, Volume 9, Books A and B, Growth of Cells in Hormonally Defined Media, Cold Spring Harbor, New York, pp 741-750). From 1970 through 1984, estrogenic mitogenic effects were most often not seen in culture. Although some laboratories were reporting positive results in serum containing medium, as cited above, others were at the same time recording negative results using the same or related cell lines (Sirbasku DA (1978) Proc Natl Acad Sci USA 75, 3786-3790; Sirbasku DA and Kirkland WL (1976) Endocrinology 98, 1260-1272; Kirkland WL et al. (1976) JNatl Cancer Inst 56, 1159-1164; Dceda T et al. (1982) In Vitro 18, 961-979; Butler WB et al. (1983) Cancer Res 41, 82-88; Edwards DP et al. (1980) Biochem Biophys Res Commun 93, 804-812; Shafie SM (1980) Science (Wash DC) 209, 701-702). Part of the problem may have been due to culture conditions (Ruedl C et al. (1990) J Steroid Biochem Mol Biol 37, 195-200; Zugmaier G et al. (1991) J Cell Physiol 141, 353-361) or possibly caused by differences that arose because of variations in the properties of cell lines in different laboratories (Seibert K et al. (1983) Cancer Res 43, 2223-2239). In addition, there are other more technical issues that are well known in this field, have been described in the literature, and which are addressed in more detail elsewhere herein and in subsequent publications (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427; Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446; and Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 447-464.) Another vital matter has been how "growth" is defined. Sonnenschein and Soto (Sonnenschein C and Soto AM (1980) JNatl Cancer List 64, 211-215) have addressed this issue very effectively. To be accepted as valid, sex steroids must cause significant changes in cellular logarithmic growth rates. Elucidation of the nature and activity of the estrogen reversible serum inhibitor(s) continues to be an area of intense experimental interest.

0015As cited above, AM Soto and C Sonnenschein have proposed that the serum mediator is an estrogen reversible inhibitor they have named estrocolyone. They have alternately described the inhibitor as a pituitary factor (Sonnenschein C and Soto AM (1978) J Steroid Biochem 6, 533- 537), α-fetoprotein (Sonnenschein C et al. (1980) JNatl Cancer Inst 64, 1141-1146; Sonnenschein C et al. (1980) JNatl Cancer Inst 64, 1147-1152; Soto AM and Sonnenschein C (1980) Proc Natl Acad Sci USA 11, 2084-2087), a serum protein different than human serum albumin (Soto AM et al. (1992) J Steroid Biochem Mol Biol 43, 703-712), and in a later reversal of this view, stated that estrocolyone 1 (i.e. the serum-borne estrogen reversible inhibitor) was human serum albumin or a combination of two domains of albumin (Sonnenschein C (1996) J Steroid Biochem Mol Biol 59, 147-154). They have also sought the inhibitor as an estrogen-binding glycoprotein different than SHBG using Concanavalin-A chromatography (Reny J-L and Soto AM (1989) J Gin Endocrinol Metab 68, 938-945). The outcome of this effort did not identify the inhibitor. The exact chemical nature of the inhibitor was even further complicated by U.S. Patent No. 4,859,585 (Sonnenschein) and U.S. Patent No. 5,135,849 (Soto) describing an inhibitor that was derived from heat inactivated serum depleted of its endogenous estrogens and androgens by a 37.5°C charcoal- dextran procedure. Alternatively, the inhibitor was obtained from serum by ammonium sulfate precipitation. This inhibitor is said to be useful for in vitro testing of substances of interest for activity as an estrogen or androgen agonist or antagonist using the MCF-7 cell line grown in Dulbecco's modified Eagle minimal essential medium supplemented with 5% (v/v) fetal bovine serum. However, the two above-mentioned U.S. Patents do not address the issues of (i) whether there are one or more inhibitors, (ii) what is/are the exact chemical composition of the inhibitor(s), and (iii) what conditions were required to yield the long term stable product(s) necessary for the commercial application of the testing methodology described. STEROID HORMONE RECEPTORS

0016As the matter stands today, it has not been established beyond doubt which of the many estrogen receptors and/or variants is the one that regulates the estrogen induced mitogenic effect. It is generally assumed that the ERcc is the most likely positive growth mediator. Estrogens, androgens, progestins, corticosteroids, mineral steroids, vitamin D, retinoic acid and thyroid hormone receptors all belong to a family of DNA binding intracellular receptors that are activated by binding of the appropriate hormone/ligand (Evans RM (1988) Science (Wash DC) 240, 889- 895; Giguere V (1990) Genetic Eng (NY) 12, 183-200; Williams GR and Franklyn JA (1994) Baillieres Gin Endocrinol Metab 8, 241-266; Kumar R and Thompson EB (1999) Steroids 64, 310-319; Pemrick SM et al. (1994) Leukemia 8, 1797-806; Carson-Jurica MA et al. (1990), Endocr Rev 11, 201-220; Tsai MJ and O'Malley BW (1994) Annu Rev Biochem 63, 451-486; Alberts B et al. (1994) Molecular Biology of The Cell, 3rd edition, Garland Publishing, New York, pp 729-731). The estrogen receptor described in the citations above is now designated the classical estrogen receptor alpha (ERα). Its role in steroid regulated gene expression has been studied extensively and often reviewed (Yamamoto KR (1985) Annu Rev Genet 19, 209-252; Green S and Chambon P (1991) In: Nuclear Hormone Receptors, Academic Press, New York, pp 15-38; Tsai M-J and O'Malley BW (1994) Annu Rev Biochem 63, 451-486; McDonnell DP et al. (1992) Proc Natl Acad Sci USA 89, 10563-10567; Landel CC et al. (1994) Mol Endocrinol 8, 1407-1419; Landers JP and Spelsberg TC (1992) Crit Rev Eukary Gene Exp 2, 19-63; Cavailles V et al. (1994) Proc Natl Acad (Sci USA 91, 10009-10013; Halachmi S et al. (1994) Science (Wash DC) 264, 1455-1458; Brasch K and Ochs RL (1995) Int rev Cyto 159, 161-194; Hard T and Gustafsson J-A (1993) Ace Chem Res 26, 644-650). It is noteworthy that estrogen resistance in man is caused by a mutation in the ERα (Smith EP et al. NEng J Med 331, 1056-1061). The most startling fact is that this point mutation (i.e. cytosine → thymidine) generated a premature stop codon, but was not lethal. Although many metabolic abnormalities were noted, development into adulthood was observed without expression of a functional ERα. This fact is further strengthened by the experiments with ERα gene knockout mice (Couse JF and Korach KS (1999) Endocr Rev 20, 358-417). The authors state "the list of unpredictable phenotypes in the α ERKO (estrogen receptor knockout) must begin with the observation that generation of an animal lacking a functional ER α gene was successful and produced animals of both sexes that exhibit a life span comparable to wild-type". Furthermore, in the review of the ERKO results, it was not possible to conclude that the ERα regulated estrogen responsive cell growth. Indeed, functions normally ascribed to the ERα seemed unaffected. In fact, only relationships to development in tissues such as breast seemed best correlated (Boccchinfuso WP and Korach KS (1997) J Mammary Gland Biol Neoplasia 2, 323-334). The situation with ERKO mice and ERβ is similar (Couse JF and Korach KS (1999) Endocr Rev 20, 358-417). The results from ERβ knockout suggest an indirect role of this receptor via stromal tissue (Gustafsson J-A and Warner M (2000) J Steroid Biochem Mol Biol 74, 254-248). Certainly a direct growth role for ERβ in breast epithelial cells was not established. The results available from ERKO do not yet provide confidence that either the ERα or the ERβ mediate estrogen responsive cell growth. There are other pertinent lines of evidence that relate to the role of the ERα and growth.

0017The first is from a study of transfection of estrogen receptor negative cells with the full length functional ERα (Zajchowski DA et al. (1993) Cancer Res 53, 5004-5011). Those investigators arrived at a remarkable result. They had expected to regain estrogen responsive growth in the transfected hormone independent cells. This was definitely not the case. Instead, addition of E₂ caused cell growth inhibition. Their results indicated that ERα was not a positive mediator, but instead a negative regulator. However, similarly transfected estrogen responsive cell lines such as MCF-7 and T47D were not E₂ inhibited in those studies.

0018More recently, another estrogen receptor has been cloned and cDNA sequenced from rat prostate and ovary (Kuiper GG et al. (1996) Proc Natl Acad Sci USA 93, 5925-5930). It has now also been cloned from mouse (Tremblay GB et al. (1997) Mol Endocinol 11, 353-365) and human (Mossehnan S et al. (1996) FEBS Lett 392, 49-53). This new receptor has been named estrogen receptor beta (ERβ). Evidence that ERβ is separate from ERα comes from the fact that the genes are located on different chromosomes (Enmark E et al. (1997) 82, 4258-4265). Therefore, ERβ is not simply an alternate splicing product of the ERα gene. Furthermore, ERβ is distinguishable from ERα based on critical differences in the amino acid sequences of functional domains (Kuiper GG et al. (1996) Proc Natl Acad Sci USA 93, 5925-5930; Enmark E et al. (1997) 82, 4258-4265; Dickson RB and Stancel GM (2000) JNatl Cancer Inst Monogr No. 27, 135-145). For example, the sequence homology between the two receptors is 97% in the DNA binding domain, but 59% in the C-terminal ligand-binding (i.e. steroid hormone-binding) domain, and only 17% in the N- terminal domain. The ERβ N-terminal domain is much abbreviated compared to the ERα (Enmark E et al. (1997) 82, 4258-4265). Rat ERβ contains an 18 amino acid insert in the domain binding the ligand. Despite the significant differences in structure, ERα and ERβ bind E₂ with the same affinity (Kuiper GG et al. (1996) Proc Natl Acad Sci USA 93, 5925-5930; Dickson RB and Stancel GM (2000) J Natl Cancer Inst Monogr No. 27, 135-145). In fact, others (Tremblay GB et al. (1997) Mol Endocrinol 11, 353-365) have stated that ERβ has a slightly lower affinity for E₂ than ERα (Tremblay GB et al. (1997) Mol Endocrinol 11, 353-365). Therefore, if either of these receptors mediates estrogen-induced growth, the steroid hormone concentrations required for one- half maximum growth (i.e. ED₅₀), or for optimum growth (i.e. ED₁₀o), are expected to be about the same. It is thought that ERα and ERβ are functionally interrelated (Kuiper GG et al. (1998)

0019Endocrinology 139, 4252-4263) and that one role of ERβ is to modulate the transcriptional activity of ERα (Hall JM and McDonnell DP (1999) Endocrinology 140, 5566-5578). Clearly however, there are significant functional differences between ERα and ERβ, which have been discussed (Gustafsson J-A (1999) J Endocrinol 163, 379-383). Also, there are functional differences expected because of the different pattern of steroid hormone binding shown by ERβ (Kuiper GG et al. (1996) Proc Natl Acad Sci USA 93, 5925-5930). For example, ERβ binds androgens whereas ERα does not. This fact, plus the location of ERβ in prostate indicates a new function that may be androgen related.

0020It should also be noted that there have been "estrogen related receptors" (ERR 1 and 2) or "orphan" receptors identified that share properties with ERα but do not have a known function and do not have a known ligand (Giguere V et al. (1988) Nature (Lond) 331, 91-94; Gustafsson J- A (1999) J Endocrinol 163, 379-383). In fact, today, there are more than 70 "orphan" receptors seeking ligands and functions (Gustafsson J-A ( 1999) Science (Wash DC) 284, 1285-1286). THE SECRETORY ΓMMUNE SYSTEM Turning now to discussion of a separate body of work from that described above, as further background for understanding the present invention, it should be noted that the immunological function and physiological properties of the body's secretory immune system have been recognized for many years (Tomasi TB et al. (1965) J Exp Med 121, 101-124; Brandtzaeg P and Baklien K (1977) Ciba Foundation Symposium 46, 77-113; Tomasi TB (1970) Ann Rev Med 21, 281-298; Spiegelberg HL (1974) Adv Immunol 19, 259-294; Tomasi TB (1976) 77 e Immune System of Secretions, Prentice-Hall, Englewood Clifts, New Jersey; Mestecky J and McGhee JR (1987) Adv Immunol 40, 153-245). The major immunoglobulins secreted as mucosal immune protectors include IgA, IgM and IgG. In human serum, the percent content of IgG, IgA and IgM are 80, 6 and 13%, respectively. In humans, the major subclasses of IgG are IgGl, IgG2, IgG3 and IgG4. These are 66, 23, 7 and 4 % of the total IgG, respectively. The relative content of human immunoglobulin classes/subclasses in adult serum follow the order IgGl > IgG2 > IgAl > IgM > IgG3 > IgA2 > IgD > IgE (Spiegelberg HL (1974) Adv Immunol 19, 259-294). When the serum concentrations of immunoglobulins are compared to those in exocrine secretion fluids, the relative contents change dramatically (Brandtzaeg P (1983) Ann NY Acad Sci 409, 353-382; Brandtzaeg P (1985) Scand J Immunol 22, 111-146). For example in colostrum (a breast fluid secretion), IgA is ≥ 80% of the total immunoglobulins. IgM is < 10 % of the total. IgG represents a few percent. In human colostrum and milk, IgGl and IgG2 are the major subclasses of IgG (Kim K et al. (1992) Ada Paediatr 81, 113-118). Clearly, comparison of serum and mucosal fluid concentrations indicate selective immunoglobulin secretion.

0021Immunoglobulin Function. All human mucus membranes are protected by the secretory immune system (Hanson L A and Brandtzaeg P (1989) In: Immunological Disorders in Infants and Children, 3^rd edition, Stiehm ER, ed, Saunders, Philadelphia, pp 169-172). The primary protector is slgA that is produced as dimers and larger polymers. A single joining "J"chain connects IgA monomers to form the dimers and polymers (Garcia-Pardo A et al. (1981) J Biol Chem 256, 11734-11738), and connects monomers of IgM to give pentamers (Niles MJ et al. (1995) Proc Natl Acad Sci USA 92, 2884-2888). This critical joining endows these structures with a very important immunological property. IgA and IgM are known to bind to bacterial, parasite and viral surface antigens. These complexes bind to receptors on inflammatory cells leading to destruction of the pathogen by antibody-dependent cell-mediated cytotoxicity (Hamilton RG (1997) "Human Immunoglobulins" In: Handbook of Human Immunology, Leffell MS et al, eds, CRC Press, Boca Raton, Chapter 3). Dimeric and polymeric slgA have a high antigen binding valence that effectively agglutinates/neutralizes bacteria and virus (Janeway CA Jr et al. (1999) Immunobiology, The Immune System in Health and Disease, 4^th edition, Garland Publishing, New York, pp 326-327). Also, slgA shows little or no complement activation. This means that it does not cause inflammatory responses (Johansen FE et al. (2000) Scand J Immunol 52, 240-248). In addition, the fact that IgA exists as two separate forms is significant (Loomes LM et al (1991) J Immunol Methods 141, 209-218). The IgAl predominates in the general circulation. In contrast, IgA2 is often higher in mucosal secretions such as those from breast, gut, and respiratory epithelium, salivary and tear glands, the male and female reproductive tracts, and the urinary tracts of both males and females. This difference in proportions is important to immune protection of mucosal surfaces. Although the secretory form of IgAl is by and large resistant to proteolysis

0022(Lindh E (1975) J Immunol 114, 284-286), a number of different bacteria secrete proteolytic enzymes that cleave it into Fab and Fc fragments (Warm JH et al. (1996) Infect Immun 64, 3967- 3974; Poulsen K et al. (1989) Infect Immun 57, 3097-3105; Gilbert JN et al. (1988) Infect Immun 56, 1961-1966; Reinholdt J et al. (1993) Infect Immun 61, 3998-4000; Blake MS and Eastby C (1991) J Immunol Methods 144, 215-221; Burton J et al. (1988) J Med Chem 31, 1647-1651; Mortensen SB and Kilian M (1984) Infect Immun 45, 550-557; Simpson DA et al. (1988) J Bacteriol 170, 1866-1873; Blake MS and Swanson J et al. (1978) Infect Immun 22, 350-358; Labib RS et al. (1978) Biochim Biophys Ada 526, 547-559). In effect, the bacterial proteinases negate the neutralizing effects of multivalent slgAl. In contrast, because of structural differences (Chintalacharuvu KR and Morrison SL (1996) J Immunol 157, 3443-3449), IgA2 lacks sites required for proteolysis. This makes IgA2 more resistant to bacterial digest than IgAl (Hamilton RG (1997) "Human immunoglobulins" In: Handbook of Human Immunology, Leffell MS et al, eds, CRC Press, Boca Raton, Chapter 3).

0023With regard to IgM, its function is somewhat different. IgM antibodies serve primarily as efficient agglutinating and cytolytic agents. They appear early in the response to infection and are largely confined to the bloodstream. Whether secreted or plasma-borne, IgM is a highly effective activator of the classical complement cascade. It is less effective as a neutralizing agent or an effector of opsinization (i.e. facilitation of phagocytosis of microorganisms). Nonetheless, IgM complement activation causes lysis of some bacteria. The effects of the IgG class are more encompassing. All four subclasses cause neutralization, opsinization and complement activation to defend against mucosal microorganisms. IgGl is an active subclass in this regard (Janeway CA Jr et al. (1999) Immunobiology, The Immune System in Health and Disease, 4^th edition, Garland Publishing, New York, pp 326-327).

0024Immunoglobulin Structure. It was established that immunoglobulin A (IgA) represents 5 to 15% of the total plasma immunoglobulins in humans (Spiegelberg HL (1974) Adv Immunol 19, 259-294). IgA has a typical immunoglobulin four-chain structure (M_r 160,000) made up of two heavy chains (M_r 55,000) and two light chains (M_r 23,000) (Fallgreen-Gebauer E et al (1993) Biol Chem Hoppe-Seyler 374, 1023-1028; Kratzin H et al. (1978) Hoppe-Seylers Z Physiol Chem 359, 1717-1745; Yang C et al. (1979) Hoppe-Seylers Z Physiol Chem 360, 1919-1940; Eiffert H et al. (1984) Hoppe-Seylers Z Physiol Chem 365, 1489-1495). In humans, there are two subclasses of IgA. These are IgAl and IgA2 that have 1 and 2 heavy chains, respectively. The IgA2 subclass has been further subdivided into A₂m(l) and A₂m(2) allotypes (Mestecky J and Russell MW (1986) Monogr Allergy 19, 277-301; Morel A et al. (1973) Clin Exp Immunol 13, 521-528). IgA can occur as monomers, dimers, trimers or multimers (Lϋllau E et al. (1996) J Biol Chem 271, 16300-16309). In plasma, 10% of the total IgA is polymeric while the remaining 90% is monomeric. Formation of dimeric or multimeric IgA requires the participation of an elongated glycoprotein of approximately M_r 15,000, designated the "J" chain (Mestecky J et al. (1990) Am J Med 88, 411-416; Mestecky J and McGhee JR (1987) Adv Immunol 40, 153-245; Cann GM et al. (1982) Proc Natl Acad Sci USA 19, 6656-6660). Structurally, the J chain is disulfide linked to the penultimate cysteine residue of heavy chains of two IgA monomers to form a dimeric complex of approximately M_r 420,000. The general structure of the dimer has been well described in the literature (Fallgreen-Gebauer E et al (1993) Biol Chem Hoppe-Seyler 374, 1023-1028). Multimeric forms of IgA and IgM require only a single J chain to form (Mestecky J and McGhee JR (1987) Adv Immunol 40, 153-245; Chapus RM and Koshland ME (1974) Proc Natl Acad Sci USA 71, 657-661; Brewer JW et al. (1994) J Biol Chem 269, 17338-17348). The structures and chemical properties of IgA and IgM have been described in detail (Janeway CA Jr et al. (1996) Immunobiology, The Immune System in Health and Disease, Second edition, Garland Publishing, New York, pp 3-32 and pp 8-19). Immunoglobulin Production. Dimeric and multimeric IgA and IgM are secreted by a number of exocrine tissues. IgA is the predominant secretory immunoglobulin present in colostrum, saliva, tears, bronchial secretions, nasal mucosa, prostatic fluid, vaginal secretions, and mucous secretions from the small intestine (Mestecky J et al. (1987) Adv Immunol 40, 153-245; Goldblum RM, et al. (1996) In: Stiehm ER, ed, Immunological Disorders in Infants and Children, 4^th edition, Saunders, Philadelphia, pp 159-199; Heremans JF (1970) In: Immunoglobulins, Biological Aspects and Clinical Uses, Merler E, ed, National Academy of Sciences, Wash DC pp 52-73; Tomasi TB Jr (1971) In: Immunology, Current Knowledge of Basic Concepts in Immunology and their Clinical Applications, Good RA and Fisher DW, eds, Sinauer Associates, Stanford, CT, p 76; Brandtzaeg P (1971) Ada Path Microbiol Scand 79, 189-203). IgA output exceeds that of all other immunoglobulins, making it the major antibody produced by the body daily (Heremans JF (1974) In: The Antigens, Vol 2, Sela M, ed, Academic Press, New York, pp 365-522; Conley ME et al. (1987) Ann Intern Med 106, 892-899. IgA is the major immunoglobulin found in human milk/whey/colostrum (Ammann AJ et al. (1966) Soc Exp Biol Med 122, 1098-1113; Peitersen B et al. (1975) A a Paediatr Scand 64, 709-717); Woodhouse L et al. (1988) NutrRes 8, 853-864). IgM secretion is less abundant but can increase to compensate for deficiencies in IgA secretion.

0025During passage of IgA through the cell, its structure is modified. A M_r 80,000 fragment of the receptor containing all five of the extracellular domains becomes covalently attached to dimeric IgA to form secretory IgA (slgA) (Fallgreen-Gebauer E et al (1993) Biol Chem Hoppe- Seyler 374, 1023-1028). The receptor that mediates the translocation has been interchangeably called the "poly-Ig receptor" (poly-Ig receptor) or the "secretory component" (Krajci P et al. (1992) Eur J Immunol 22, 2309-2315). For the purposes of the present disclosure, however, the term "poly-Ig receptor" refers to the full length M_r 100,000 transmembrane protein and the term "secretory component" denotes only the M_r 80,000 extracellular five domains of the receptor that become covalently attached to IgA in forming the slgA structure (Fallgreen-Gebauer E et al (1993) Biol Chem Hoppe-Seyler 374, 1023-1028; Krajci P et al. (1992) Eur J Immunol 22, 2309-2315). Because of the unique structure of slgA, it is highly resistant to acid and proteolysis (Lindh E (1975) J Immunol 114, 284-286) and therefore remains intact in secretions to perform extracellular immunological functions. IgM also binds secretory component, but not covalently (Lindh E and Bjork I (1976) Eur J Biochem 62, 271-278). However, IgM is less stabilized because of its different association with the secretory component, and therefore has a shorter functional survival time in acidic secretions (Haneberg B (1974) Scand J Immunol 3, 71-76; Haneberg B (1974) Scand J Immunol , 191-197).

0026The secretion mechanism for IgA and IgM are well described. Conversely, there is a fundamental question surrounding IgG secretion. There is no "J" chain present in IgG land IgG2. From the known facts of transcytosis/secretion of immunoglobulins (Johansen FE et al. (2000) Scand J Immunol 52, 240-248), it is unlikely that IgG secretion is mediated by the poly-Ig receptor. An epithelial receptor specific for IgGl has been reported in bovine mammary gland (Kemler R et al. (1975) Eur J Immunol 5, 603-608). Apparently, it preferentially transports this class of immunoglobulins from serum into colostrum. Despite this 1975 report however, the receptor has not been chemically or structurally identified nor has the mechanism of transport of IgG monomers been satisfactorily defined. It is possible that this receptor is a member of a large group now designated as Fc receptors (Fridman WH (1991) FASEB J 5, 2684-2690), but there is one study with IgG showing that of 31 different long-term human carcinoma cell lines including breast "all lines were found to be consistently Fc receptor negative" (Kerbel RS et al. (1997) IntJ Cancer 20, 673-679). One possible candidate for the epithelial transport of IgGl is the neonatal Fc receptor (Raghavan M and Bjorkman PJ (1996) Annu Rev Cell Dev Biol 12, 181-220). However, there is no indication yet of the presence of this receptor in adult mucosal tissues.

0027Transcytosis Mediating Receptors. J chain-containing IgA is produced and secreted by plasma B immunocytes located in the lamina propria just beneath the basement membrane of exocrine cells (Brandtzaeg P (1985) Scan J Immunol 22, 111-146). The secreted IgA binds to a M_r 100,000 poly-Ig receptor positioned in the basolateral surface of most mucosal cells (Heremans JF (1970) In: Immunoglobulins, Biological Aspects and Clinical Uses, Merler E, ed, National Academy of Sciences, Wash DC, pp 52-73; Brandtzaeg P (1985) Gin Exp Immunol 44, 221-232; Goodman JW (1987) Li: Basic and Clinical Immunology, Stites DP, Stobo JD and Wells JV, eds, Appleton and Lange, Norwalk, CT, Chapter 4). The receptor-IgA complex is next translocated to the apical surface where IgA is secreted. The binding of dimeric IgA to the poly-Ig receptor is completely dependent upon the presence of a J chain (Brandtzaeg P (1985) Scan J Immunol 22, 111-146; Brandtzaeg P and Prydz H (1984) Nature 311:71-73; Vaerman J-P et al. (1998) Eur J Immunol 28, 171-182). Monomeric IgA will not bind to the receptor. The J chain requirement for IgM binding to the poly-Ig receptor is also true for this immunoglobulin (Brandtzaeg P (1985) Scan J Immunol 22, 111-146; Brandtzaeg P (1975) Immunology 29, 559-570; Norderhaug IN et al.

0028(1999) Crit Rev Immunol 19, 481-508). Because IgA and IgM bind to the poly-Ig receptor via their Fc domains, and because of a repeating Ig-like structure in the extracellular domains, the poly-Ig receptor classifies as a member of the Fc superfamily of immungobulin receptors (Krajci P et al. (1992) Eur J Immunol 22, 2309-2315; Daeron M (1991) Annu Rev Immunol 15, 203-234).

0029The poly-Ig receptor and the secretory component from human has been cDNA cloned and DNA sequenced (Krajci P et al. (1992) Eur J Immunol 11, 2309-2315; Krajci P et al. (1995) Adv Exp Med Biol 371A, 617-623; Krajci P et al. (1991) Hum Genet 87, 642-648; Krajci P et al. (1989) Biochem Biophys Res Commun 237, 9-20) as has the poly-Ig receptor from mouse (Kushiro A and Sato T (1997) Gene 204, 277-282; Piskurich JF et al. (1995) and bovine tissue (Verbeet MP et al. (1995) Gene 164, 329-333). Altogether, the human poly-Ig receptor coding sequence encompassed 11 exons. The extracellular five domains originate from exons 3 (Dl), exon 4 (D2) exon 5 (D3 and D4), exon 6 (D5), exon 7 (the conserved cleavage site to form the secretory component), exon 8 (the membrane spanning domain), exon 9 (a serine residue required for transcytosis), exon 9 (sequence to avoid degradation), exon 10, no known function) and exon 11 (sequence contains a threonine residue and the COOH terminus) (Krajci P et al. (1992) Eur J Immunol 11, 2309-2315). With the exception of domains 3 and 4 (both from one exon), the receptor structure follows the rule of one domain/one exon. The poly-Ig receptor binds IgA and IgM via their Fc domains, and more particularly, via a specific amino acid sequence (15→ 7) of domain 1 (Bakos M-A et al. (1991) J Immunol 147, 3419-3426). Of the other extracellular domains, only D5 is known for a specific function. It contains the disulfide bonds that covalently attach to IgA to for slgA during transcytosis. The role of this receptor in transcytosis of IgA IgM has been well studied with mucosal tissues and epithelial cells in culture (Naerman JP et al. (1998) Eur J Immunol 28, 171-182; Fahey JN et al. (1998) Immunol Invest 11, 167-180; Brandtzaeg P (1997) JReprod Immunol 36, 23-50; Loman S et al. (1997) Am J Physiol 111, L951-L958; Mostov KE et al. (1995) Cold Spring Harbor Symp Quant Biol 60, 775-781; Schaerer E et al. (1990) J Cell Biol 110, 987-998). During passage of IgA through the cell, its structure is modified. A M_r 80,000 fragment of the receptor containing all five of the extracellular domains becomes covalently attached to dimeric IgA to form secretory IgA (slgA) (Fallgreen-Gebauer E et al (1993) Biol Chem Hoppe- Seyler 374, 1023-1028). The receptor that mediates the translocation has been interchangeably called the "poly-Ig receptor" (poly-Ig receptor) or the "secretory component" (Krajci P et al. (1992) Eur J Immunol 22, 2309-2315). For the purposes of the present disclosure, however, the term "poly-Ig receptor" refers to the full length M_r 100,000 transmembrane protein and the term "secretory component" denotes only the M_r 80,000 extracellular five domains of the receptor that become covalently attached to IgA in forming the slgA structure (Fallgreen-Gebauer E et al (1993) Biol Chem Hoppe-Seyler 374, 1023-1028; Krajci P et al. (1992) Eur J Immunol 22, 2309-2315). Because of the unique structure of slgA, it is highly resistant to acid and proteolysis (Lindh E

0030(1975) J Immunol 114, 284-286) and therefore remains intact in secretions to perform extracellular immunological functions. IgM also binds secretory component, but not covalently (Lindh E and Bjork I (1976) Eur J Biochem 62, 271-278). However, IgM is less stabilized because of its different association with the secretory component, and therefore has a shorter functional survival time in acidic secretions (Haneberg B (1974) Scand J Immunol 3, 71-76; Haneberg B (1974) Scand J Immunol 3, 191-197). While the secretion mechanism for IgA and IgM are well described, conversely, a fundamental question surrounds IgG secretion. There is no "J" chain present in IgG land IgG2. From the known facts of transcytosis/secretion of immunoglobulins (Johansen FE et al. (2000) Scand J Immunol 52, 240-248), it is unlikely that IgG secretion is mediated by the poly-Ig receptor. An epithelial receptor specific for IgGl has been reported in bovine mammary gland (Kemler R et al. (1975) Eur J Immunol 5, 603-608). Apparently, it preferentially transports this class of immunoglobulins from serum into colostrum. Despite this 1975 report however, the receptor has not been chemically or structurally identified nor has the mechanism of transport of IgG monomers been satisfactorily defined. It is possible that this receptor is a member of a large group now designated as Fc receptors (Fridman WH (1991) FASEB J 5, 2684-2690), but there is one study with IgG showing that of 31 different long-term human carcinoma cell lines including breast "all lines were found to be consistently Fc receptor negative" (Kerbel RS et al. (1997) Int J Cancer 20, 673-679). One possible candidate for the epithelial transport of IgGl is the neonatal Fc receptor (Raghavan M and Bjorkman PJ (1996) Annu Rev Cell Dev Biol 12, 181-220). However, there is no indication yet of the presence of this receptor in adult mucosal tissues. Fc receptors are so named because they bind specific heavy chains (Fc domains).

0031However, before coming to this conclusion, it should be emphasized strongly that the Fc family represented by Fcγ (IgG), Fcα (IgA), and Fcμ (IgM) have traditionally been considered to be located on lymphoid series cells (Fridman WH (1991) FASEB J 5, 2684-2690; Raghavan M and Bjorkman PJ (1996) Annu Rev Cell Dev Biol 12, 181-220). There is only limited experimental support for the concept that these receptors are located on epithelial cells (Tonder O et al. (1976) Ada Pathol Microbiol Scand 84, 105-111). For the family of leukocyte IgG receptors, 12 transmembrane or soluble receptor isoforms are known. These are grouped into three classes FcγRl (CD64), Fcγ RU (CD32) and Fcγ RTJI (CD 16) (Valerius T et al. (1997) Blood 90, 4485- 4492). For IgA, there is one gene that encodes several receptors) (i.e. Fcα) by alternate splicing to yield forms from M_r 55,000 to 110,000 (Pleass RJ et al. (1996) Biochem J 318, 771-777; van Dijk TB et al. (1996) Blood 88, 4229-4238; Morton HC et al. (1996) lmmunogenetics 43, 246-247). One of these, FcαRl is constitutively expressed on monocytes and macrophages and other leukocytes. It binds IgAl and IgA2 with about the same affinity. The receptor for IgM (t.e.Fcμ) is less well defined, but still has been partially characterized as a M_r 60,000 protein present on activated B cells and other B series cells (Ohno T et al. (1990) J Exp Med 111, 1165-1175). The Fc superfamily has another very important aspect pertinent to this disclosure. Receptors of this family mediate negative effects on cells (Cambier JC (1997) Proc Natl Acad Sci USA 94, 5993- 5995). These receptors have an intracellular amino acid sequence motif I/NxYxxL described as an immunoreceptor tyrosine-based inhibitory motif (ITIM) that signals cell growth shutdown after ligand binding. These signals have been characterized in the FCγRIIBl receptors of human and mouse (Olcese L et al. (1996) J Immunol 156, 4531-4534). The hallmark of these ITIM receptors is that they shut off growth factor dependent growth.

0032Although the advances and teachings in the prior art have indicated that a serum-borne inhibitor of steroid hormone responsive tumor cell growth exists, until now there has been no adequate isolation or identification of such an inhibitor, and very little understanding of its mode of action has been gained. There is no satisfactory in vitro testing model presently available for demonstrating steroid hormone responsive cell growth that can be correlated to the in vivo situation, or for testing drugs, or other natural or synthetic substances for possible hormone- mimicking or anti-hormone effects.

SUMMARY OF THE INVENTION

0034The compositions, methods and models of the present invention overcome major shortcomings of the prior art and satisfy long-felt needs for, among other things, a sensitive way to screen substances for estrogenic and androgenic effects. It was discovered, and the embodiments herein demonstrate, that the immune system plays a major role in the growth of estrogen responsive breast and androgen responsive prostate cancers, as well as cancers of other steroid and thyroid hormone responsive mucosal epithelial tissues. IgA, IgM and certain IgGs provide negative regulation of steroid hormone responsive mucosal epithelial cancer cell growth, including breast, prostate, pituitary, kidney and other glandular cancer cells. For the purposes of this disclosure, "cell growth" means cell proliferation or an increase in the size of a population of cells through mitogenesis and cell division rather than an increase in cytoplasmic volume of an individual cell. Prior to the present disclosure, no growth regulating role was known for the secretory immune system, which produces predominantly immunoglobulin A (IgA) and immunoglobulin M (IgM) and lesser amounts of immunoglobulin G (IgG). The discovery that IgA and IgM are the major negative regulators of steroid hormone responsive cell growth arose out of the Inventor's work directed at purifying breast cancer regulatory factors from biological fluids, as described in the following Examples. This discovery and the present invention constitute a major breakthrough in the understanding of these cancers, and other glandular/mucosal tissues that secrete or are bathed by polymeric IgA, secretory IgA (slgA), IgM and certain IgGs. For the first time, a direct link has been established between the secretory immune system (IgA and IgM) and the most prevalent types of cancer that occur throughout the world. Binding of IgA and IgM to the poly-immunoglobulin receptor (i.e. poly-Ig receptor or poly IgR) is an important step in carrying out the regulatory function of IgA and IgM, and initial indications are that poly IgR mediates the negative regulation of steroid hormone dependent cell growth. Application of these scientific breakthroughs to the detection, diagnosis, prognosis, treatment and deterrence or prevention of cancer of mucosal epithelial tissues (e.g., breast, prostate, kidney, pituitary, thyroid and colon) is described in U.S. Patent App. No. (Arty. Dkt. No. 1944-00800)/PCT/US2001/ (Arty. Dkt. No. 1944-00801) entitled "Compositions and Methods for the Diagnosis, Treatment and Prevention of Steroid Hormone Responsive Cancers," which is hereby incorporated herein by reference.

0035No such serum-derived inhibitor has been previously isolated or identified that replicates the large magnitude estrogen reversible inhibitory effects demonstrated in the present investigations using hormone depleted full serum. The serum-borne inhibitor(s) are ubiquitous in mammals and lack species specificity. Their inhibitory activity is completely reversible by picomolar concentrations of steroid hormones when assayed in the new in vitro conditions. Moreover, before the surprising discovery that the serum-borne inhibitor(s) are secretory immunoglobulins, there has been no previous report that IgA, IgM or IgG play any role in the negative regulation of steroid hormone responsive (SHR) mucosal epithelial cell growth, or that binding of IgA and IgM to a polyimmunoglobulin receptor (poly-Ig receptor) is instrumental in carrying out such growth regulation. Prior to the present disclosure, a cell growth related function for the poly-Ig receptor transcytosis receptor, or a poly IgR-like receptor, has not been recognized, nor had such a role ever been attributed to an IgG Fcγ receptor. In accordance with certain embodiments of the present invention, an immunoglobulin inhibitor of in vitro steroid hormone-responsive steroid hormone responsive cancer cell growth or proliferation are provided. The cancer cells that are inhibited from proliferating in in vitro culture come from a cell line that is also capable of proliferating in vivo when implanted into a suitable host. This immunoglobulin inhibitor (e.g., one or more of the secretory immunoglobulins IgA, IgM and certain IgG subtypes) is a long sought after serum-derived negative regulator of steroid hormone responsive cancer cell growth that, in impure form, was previously referred to as a steroid hormone binding globulin like ("SHBG-like") fraction. For the first time it is disclosed that, surprisingly, certain immunoglobulins exert a steroid hormone reversible negative regulatory (inhibitory) effect on cancer cell growth that is distinct from their well-established immune functions. In the most preferred embodiments, the inhibitors) is/are dimeric IgA (non-sIgA), polymeric IgM, IgGlκ and IgG2.

0036Li some embodiments an isolated steroid hormone reversible inhibitor of steroid hormone- responsive cancer cell growth is provided, the inhibitor comprising a secretory immunoglobulin, such as IgA, IgM or IgG. Li some embodiments a steroid hormone irreversible cell growth inhibitor composition is provided that comprises at least one immunoglobulin inhibitor that is active with respect to the ability to inhibit steroid hormone-responsive cancer cell proliferation and inactive with respect to steroid hormone reversibility of the inhibition, and a carrier.

0037L some embodiments a method of making a steroid hormone irreversible cancer cell growth inhibitor composition comprising exposing an above-described inhibitor composition to calcium depleted conditions for a defined period of time sufficient to render the immunoglobulin irreversibly inhibitory of steroid hormone responsive cancer cell growth in vitro.

0038According to still other embodiments of the present invention an immunoglobulin inhibitor mimicking substance is provided. Li certain embodiments the mimicking substance is tamoxifen.

0039In certain other embodiments of the present invention a negative control serum is provided which contains steroid hormone depleted blood plasma or serum and is inactive with respect to the ability to inhibit steroid hormone-responsive cell proliferation in the absence of steroid hormone. Some embodiments of the invention provide a method of making a negative control serum, preferably comprising heat treatment at about 50-60°C for about 90 minutes to about 30 hours.

0040Also provided in accordance with certain other embodiments of the present invention is a control serum composition containing plasma or serum and containing a reactivatible immunoglobulin inhibitor that is inactive with respect to the ability to inhibit steroid hormone- responsive cell proliferation in the absence of the steroid hormone and in the absence of an activating amount of calcium. Li some embodiments the control serum is reactivated and contains calcium ion. These immunoglobulin inhibitors have many immediate and potential applications as reagents for cell growth assays and therapeutic agents. For example, they are useful for in vitro testing of substances for estrogenic effects (or other steroid hormone-like effects) on steroid hormone responsive cell growth, in a suitable assay system. They are useful for demonstrating steroid hormone reversible inhibition or arrest of cancer cell growth in a variety of in vitro cell culture

0041^' models employing cancer cell lines that are capable of in vivo tumor growth when implanted into a compatible host. The immunoglobulin inhibitors are also useful as an aid in assessing risk of cancer development or growth in a mucosal epithelial tissue (i.e., glands and tissues that secrete or are bathed by secretory immunoglobulins). Some of these tissues are breast, prostate, oral cavity mucosa, salivary/parotid glands, esophagus, stomach, small intestine, colon, tear ducts, nasal passages, liver and bile ducts, bladder, pancreas, adrenals, kidney tubules, glomeruli, lungs, ovaries, fallopian tube, uterus, cervix, vagina, and secretory anterior pituitary gland cells. The immunoglobulin inhibitors are also expected to be useful in the detection, diagnosis, prognosis, treatment and prevention of steroid hormone responsive cancers of the mucosal epithelial tissues.

0042In some embodiments of the present invention, a steroid hormone reversible, steroid hormone responsive cancer cell growth inhibitor composition is provided that contains at least one of the above-described immunoglobulin inhibitors together with a carrier, which preferably includes an inhibitor stabilizing medium. Such a composition is especially useful for storing and shipping preparations of the inhibitors without loss of activity. Preferably the stabilizing medium also contains an activity-stabilizing amount of calcium ion, a steroid hormone such as (DHT), and a substance that depresses the freezing point of the composition below about -20°C (e.g., glycerol). i some embodiments the composition contains steroid hormone depleted body fluid such as blood plasma or serum. In some embodiments of the present invention, an immunoglobulin inhibitor composition containing steroid-hormone depleted blood plasma or serum is provided. For many in vitro tests of steroid hormone responsive cancer cell growth, it is especially desirable to more closely approximate the in vivo condition by employing serum-containing assay medium instead of completely serum-free medium. In some embodiments this steroid hormone depleted serum-based immunoglobulin inhibitor composition is supplemented or "spiked" with a predetermined amount of certain inhibitors (e.g., IgA or IgM). Such serum-containing compositions are especially useful in assaying for estrogen-like cell growth stimulating effects by a substance of interest. These serum based compositions will also facilitate identification of substances that demonstrate a steroidogenic effect (e.g., estrogen-like stimulation of cell proliferation) in serum-free cell growth assays, but which do not demonstrate the same estrogenic effect in the presence of serum (i.e., when tested in a similar cell growth assay medium that contains serum.) The ability to determine whether a new drug, or other substance of interest, is likely to be non-estrogenic in vivo due to the presence or ameliorating effect of serum factors is of value to the medical profession and the pharmaceutical industry, in particular.

0043Accordingly, certain embodiments of the present invention provide methods of testing substances of interest, such as drugs or environmental chemicals, for their steroid hormone-like effects on cell growth stimulation employing one of the above-described immunoglobulin inhibitors or serum-based immunoglobulin inhibitor compositions with an appropriate steroid hormone responsive cell line and nutrient medium.

0044Certain embodiments of the present invention provide methods of testing substances of interest, such as drugs or environmental chemicals, to distinguish cytotoxic effects from anti- estrogenic effects on cell growth. These methods employ one of the above-described immunoglobulin inhibitors or serum-based immunoglobulin inhibitor compositions in an appropriate steroid hormone responsive cell line maintained in a suitable nutrient medium.

0045In still other embodiments of the present invention a non-inhibitory steroid hormone depleted serum composition is provided that contains steroid hormone-depleted blood plasma or serum, similar to certain of the above-described steroid hormone depleted serum-based immunoglobulin inhibitor compositions, except in this embodiment it contains either no immunoglobulin inhibitors) or it contains the immunoglobulin inhibitors) in inactive form with respect to ability of the immunoglobulin(s) to inhibit steroid hormone responsive cell proliferation in serum-free cell culture in the absence of a cell growth stimulating amount of steroid hormone. A non-inhibitory steroid hormone depleted serum composition is useful for many in vitro testing situations utilizing serum or plasma, in which the presence of steroid hormones is undesirable. For example, such a serum composition, prepared from a mature animal source, may be advantageously substituted for conventional fetal bovine serum to provide the in vitro growth promoting factors found in serum without introducing spurious amounts of steroid hormone.

0046In some embodiments, a non-inhibitory steroid hormone depleted serum composition comprises steroid hormone depleted blood plasma or serum plus an immunoglobulin inhibitor in a reactivatibly inactive form with respect to ability of the immunoglobulin to inhibit steroid hormone responsive cell proliferation in a suitable cell growth assay absent an inhibition-reversing amount of the steroid hormone. Li certain embodiments the non-inhibitory steroid hormone depleted serum composition contains less than an inhibitor activating amount of calcium ion. Li certain embodiments an active immunoglobulin inhibitor containing steroid hormone depleted serum composition is provided that is in reactivated form and contains an immunoglobulin inhibitor reactivating amount of calcium ion.

0047Li accordance with certain other embodiments of the invention, a method of making a steroid hormone-depleted serum extract comprising a steroid hormone reversible inhibitor of steroid hormone responsive cell growth is provided. Li some embodiments the method comprises (a) obtaining a non-heat-inactivated fresh or frozen serum specimen; (b) performing a first charcoal- dextran extraction on the specimen at about 30-37°C, preferably 34°C, to yield a first extract; and (c) performing a second 30-37°C, preferably 34°C, charcoal-dextran extraction on the first extract to yield a substantially steroid hormone-depleted serum extract. In another embodiment of the present invention, an alternative method of making a substantially steroid hormone-depleted serum extract comprising a steroid hormone reversible inhibitor of steroid hormone responsive cell growth is provided. Li certain embodiments this method comprises obtaining a non-heat-inactivated fresh or frozen serum specimen and performing an XAD™-4 extraction of the specimen. Li still another embodiment of the invention, a method of making a purified immunoglobulin inhibitor of steroid hormone responsive cancer cell growth is provided. This method includes (a) obtaining a substantially steroid hormone-depleted serum comprising an inhibitor of steroid hormone responsive cancer cell growth; (b) loading the depleted serum onto an agarose-based affinity matrix and eluting a fraction comprising the inhibitor; (c) loading the fraction onto a phenyl-Sepharose™ matrix and eluting a substantially purified inhibitor pool with a suitable buffer containing ethylene glycol; and concentrating the pool to yield a substantially purified inhibitor.

0048Certain embodiments of the present invention provide in vitro assay methods for detecting steroid hormone-like cell growth stimulation by a substance of interest. In some embodiments, the assay method comprises maintaining a predetermined population of steroid hormone-responsive cells in a nutrient medium comprising a quantity of an immunoglobulin cell growth inhibitor sufficient to inhibit cell growth in the absence of an inhibition-reversing amount of the steroid hormone. In some embodiments the medium is serum-free and the cells themselves are serum free and obtained from a stable steroid hormone-responsive cell line. The method also comprises adding a substance of interest to the cells and medium to yield a test mixture. The test mixture is then incubated for a predetermined period of time under cell growth promoting conditions. "Cell growth promoting conditions" refer to general environmental conditions, other than defined medium components, and include such things as favorable conditions of gaseous atmosphere, temperature and pH. Forexample, cell growth promoting conditions could include incubation at 37°C in a humid atmosphere of 5% (v/v) C0₂ and 95% (v/v) air in a defined nutrient medium at pH 7.4. After incubation for the desired period of time, it is determined whether the cell population in the test mixture has measurably increased, an increase indicating a steroid hormone-like cell growth stimulating effect by the substance of interest. An assay procedure such as this can be used for in vitro screening of drugs or other body-affecting substances for unwanted cell growth stimulating properties as an aid to avoiding undesirable side effects of such drug or substance in vivo. Li certain alternative embodiments, the assay method includes adding to the nutrient medium a defined amount of steroid-hormone depleted serum, which contains the inhibitors), and which is obtained from non-heat inactivated serum.

0049Li some embodiments of the assay method, in which the substance of interest contains or is suspected of containing proteolytic activity, the method includes selecting an immunoglobulin inhibitor such as IgA2, which resists protease degradation. In some embodiments of the assay method an inactive inhibitor-containing control serum is substituted for an active inhibitor-containing serum, to evaluate a substance of interest for cytotoxicity.

0050Li certain embodiments, the assay method comprises an assay procedure similar to the one previously described except that a defined amount of inactive immunoglobulin cell growth inhibitor (i.e., incapable of inhibiting steroid hormone responsive cell growth in the absence of an inhibition- reversing amount of the steroid hormone) is substituted for the active (inhibitory) immunoglobulin inhibitor. Li some embodiments a test substance is included in the test mixture. As assay of this type is particularly useful for determining a maximum (uninhibited) level of steroid hormone responsive cell growth stimulation by a test substance. Alternatively, this type of assay can be used to distinguish cytotoxic effects of a test substance from anti-estrogen activity, for example.

0051In accordance with certain embodiments of the invention, a method of detecting a steroid hormone antagonistic substance is provided. The method comprises (a) maintaining a predetermined population of steroid hormone responsive cancer cells in a nutrient medium comprising a quantity of immunoglobulin inhibitor sufficient to inhibit cell growth in the absence of an inhibition-reversing amount of the steroid hormone, the cells also being steroid hormone responsive for in vivo proliferation; (b) adding a defined amount of the substance of interest to the cells and medium; (c) adding to the cells and medium a defined amount of steroid hormone sufficient to stimulate cell growth in the presence of the inhibitor and in the absence of the substance of interest, to yield a test culture; (d) incubating the test culture for a predetermined period of time under cell growth promoting conditions; (e) testing the substance of interest for cytotoxic effects on the cells; and (f) determining the cell population in the test culture after the predetermined period of time, a lack of measurable increase in the cell population not attributable to cytotoxic effects of the substance indicating a steroid hormone antagonistic effect by the substance of interest.

0052Li accordance with certain embodiments of the invention, cell culture media are provided that comprise a basal nutrient fluid, such as D-MEM/F12, and are substantially devoid of unbound Fe(III), i.e., preferably containing less than 1 μM Fe (HI), and more preferably containing 0.15 μM or less. Li certain preferred embodiments, the amount of free, or active Fe (III) in the medium is less than a cell growth inhibiting concentration. The media also contain calcium ion, preferably about 0.6 mM to 1.0 M, and more preferably about 0.6 to 10 mM calcium. Li certain preferred embodiments, the concentration of calcium ion in the nutrient medium is preferably sufficient to maintain the inhibitory activity of any immunoglobulin inhibitors present in the media. Li certain embodiments, a cell culture medium that is especially suited for use in serum-free cell growth studies also includes a Fe(πT) chelating agent, preferably deferoxamine, and a cell attachment promoting protein, preferably fibronectin. L certain preferred embodiments the defined composition medium is DDM-2MF, CAPM, DDM-2A or PCM-9, the compositions of which are set out in the Examples below. Li preferred embodiments, the cell culture media comprise 100 ng/mL to 10 μg/mL insulin, 0.3 - 10 nM triiodothyronine, 2 - 50 μg/mL diferric transferrin, 5 - 100 μM ethanolamine, 0.2 - 5.0 mg/mL bovine serum albumin (BSA), 5 - 20 ng/mL selenium, 2 - 10 μM deferoxamine. Depending on the requirements of the selected cells to be cultures, the medium may also contain at least one of the following components: 1 - 50 ng/mL EGF, 0.2 - 20 ng/mL aFGF, 5 - 50 μM phosphoethanolaniine, 50 - 500 μg/mL linoleic acid-BSA, 1 - 50 μg/mL reduced glutathione, 0.5 - 2.0 mM glutamine, 1 - 10 ug/mL heparin, and 20 - 50 μg (per 35-mm diameter culture dish) human fibronectin. In some embodiments the cell culture medium also includes steroid hormone depleted serum.

0053According to other embodiments of the present invention an in vitro method of culturing steroid hormone responsive cancer cells or autonomous cancer cells is provided. The method comprises (a) maintaining a predetermined population of steroid hormone responsive cells or steroid hormone-independent cancer cells in a steroid hormone-free nutrient medium comprising an above- described cell culture medium and a quantity of immunoglobulin inhibitor sufficient to inhibit cell growth of steroid hormone responsive cancer cells in the absence of an inhibition-reversing amount of the steroid hormone, to provide an incubation mixture, the steroid hormone responsive cells also being steroid hormone responsive for proliferation in vivo when implanted into a suitable host, and the steroid hormone independent cancer cells also being steroid hormone independent for proliferation in vivo when implanted into a suitable host; (b) optionally, adding an inhibition- reversing amount of the steroid hormone to the incubation mixture; (c) incubating the incubation mixture under cell growth promoting conditions; (d) optionally, determining the cell population in the reaction mixture after incubation for a predetermined period of time.

0054According to certain embodiments of the invention an in vitro method of detecting a cell growth stimulatory or inhibitory effect of a substance of interest on steroid hormone independent cancer cells is provided. The method includes (a) maintaining a predetermined population of steroid hormone independent cancer cells in a nutrient medium as described above, optionally, devoid of the steroid hormone, and, optionally, containing a predetermined quantity of immunoglobulin inhibitor, the steroid hormone independent cells also being steroid hormone independent for proliferation in vivo when implanted into a suitable host; (b) adding a predetermined quantity of the substance of interest to the cells and medium to yield a test mixture; (c) incubating the test mixture for a predetermined period of time under cell growth promoting conditions; (d) optionally, assessing cytotoxicity of the substance of interest; and (e) determining the cell population in the test mixture after the incubation for the predetermined period of time, a measurable increase in the cell population indicating a cell growth stimulating effect by the substance of interest, and an absence of increase in the cell population, not attributable to cytotoxic effects, indicating a cell growth inhibitory effect by the substance of interest.

0055In accordance with still another embodiment of the present invention, an in vitro method of detecting an immunoglobulin inhibitor-like cancer cell growth inhibitory effect by a substance of interest is provided which comprises (a) maintaining a predetermined population of steroid hormone responsive cancer cells in a nutrient medium as described above, optionally, devoid of the steroid hormone, and, optionally, containing a predetermined quantity of inactivated immunoglobulin inhibitor, the steroid hormone responsive cells also being steroid hormone responsive for proliferation in vivo when implanted into a suitable host; (b) adding a predetermined quantity of the substance of interest to the cells and medium to yield a test mixture; (c) adding to the test mixture an amount of the steroid hormone that would be sufficient to induce cell growth in the absence of an active immunoglobulin inhibitor; (d) incubating the test mixture for a predetermined period of time under cell growth promoting conditions; (e) optionally, assessing cytotoxicity of the substance of interest; and (f) deteπ ining the cell population in the test mixture after the predetermined period of time, a measurable increase in the cell population indicating a lack of cell growth inhibitory effect by the amount of the substance of interest, and no increase in the cell population, not attributable to a cytotoxic effect, indicating a cell growth inhibitory effect by the amount of the substance of interest.

0056L accordance with another embodiment, a method of producing a quantity of a biomolecule, of interest such as a protein, peptide or polynucleotide. The method includes, in a serum-free nutrient medium as described above, culturing a population of cells expressing the biomolecule of interest, harvesting and recovering the biomolecule from the medium. Li certain preferred embodiments the protein is a monoclonal antibody. In accordance with another embodiment, a method of propagating a virus of interest is provided which comprises culturing a population of virus infected cells in an above-described serum- free nutrient medium, harvesting and recovering viruses from the medium.

0057Further provided in accordance with certain embodiments of the invention is an assay kit for detecting in vitro steroid hormone reversible steroid hormone-responsive cell growth by a substance of interest. Li some embodiments such a kit comprises a serum-free defined nutrient cell culture medium substantially free of unbound Fe(iπ) and containing calcium ion. The kit also contains a substantially steroid hormone-depleted serum comprising a steroid hormone reversible immunoglobulin inhibitor of steroid hormone responsive cell growth. Li certain preferred embodiments the extract is prepared by either a double charcoal-dextran extraction method or the XAD-4™ extraction method, described above. Li some embodiments the kit also includes a control serum composition comprising an inactivated immunoglobulin inhibitor. Li some embodiments the kit also includes a population of cultured steroid hormone responsive cancer cells that are also steroid hormone responsive for proliferation in vivo, preferably MTW9/PL2 rat mammary tumor cells. Li alternative embodiments, assay kits for detecting in vitro steroid hormone reversible steroid hormone-responsive cell growth by a substance of interest are provided. Li certain embodiments the kit, which is similar to the one described above, isolated immunoglobulin inhibitors (e.g., IgA, IgM and/or IgGl) are included in addition to, or instead of, the serum-based inhibitor composition(s). Use of this kit will be preferred when the user requires a totally serum-free assay system. In some situations both the steroid hormone depleted serum-containing and the serum-free assay systems are employed in order to detect serum factor effects or to distinguish the influence of serum on detection of cytotoxic effects of a chemical, for example.

0058Li some embodiments the kit also contains other components such as various steroid hormones, or agonists or antagonists thereof, that may be desired for adding to the medium in particular test situations.

0059In certain embodiments of the invention in vitro assay methods for detecting an immunoglobulin inhibitor of steroid hormone responsive cell growth in a sample of interest, such as a drug or environmental substance, blood serum or another body fluid, are provided. Li some embodiments the method comprises (a) maintaining a predetermined population of steroid hormone- responsive culture cells in a nutrient medium, the cells also being steroid hormone dependent for proliferation in vivo when implanted into a suitable host; (b) adding a quantity of steroid hormone to the medium sufficient to stimulate proliferation of the cells under cell growth promoting culture conditions; (c) adding a predetermined quantity of the sample of interest to the medium to yield a test mixture; (d) incubating the test mixture for a predetermined period of time under cell growth promoting culture conditions; (e) optionally, testing the sample for cytotoxic effects on the cells; and (f) determining the cell population in the test mixture after the predetermined period of time, a measurable decrease in the cell population not attributable to cytotoxic effects indicating inhibition by the amount of sample of steroid hormone responsive cell growth. '

0060In some embodiments the assay method also includes adding to the test mixture an amount of the steroid hormone in excess of the minimum amount necessary to maximally stimulate proliferation of the cells; and determining the cell population of the test mixture after the predetermined period of time, a measurable increase in the cell population indicating reversal by the excess amount of steroid hormone of steroid hormone responsive cell growth inhibition.

0061In accordance with still other embodiments of the present invention, in vitro cell culture models for predicting an in vivo steroid hormone-responsive cancer cell growth effect of a defined stimulus, such as an estrogen, an anti-estrogen, androgen, or other steroid hormone, or a steroid hormone mimicking compound, are provided. In certain embodiments the model includes steroid hormone-responsive cancer cells maintained in a growth medium containing a basal nutrient fluid substantially free of unbound Fe (JS), containing calcium ion, and containing an amount of steroid hormone reversible immunoglobulin inhibitor sufficient to arrest cancer cell growth in the absence of an inhibition-reversing amount of the steroid hormone. The cells are also steroid hormone responsive for proliferation in vivo, when implanted into a suitable host. The immunoglobulin inhibitor is chosen from among IgA, IgM and IgG, and combinations thereof. In some embodiments the nutrient medium is serum free, and in others it contains steroid hormone depleted blood plasma or serum. Li certain embodiments the steroid hormone responsive culture cells are MTW9/PL2 (rat mammary cancer), T47D (human breast carcinoma), MCF-7 (human breast carcinoma), MCF-7A (human breast carcinoma), MCF-7K (human breast carcinoma), LNCaP (human prostatic carcinoma), ZR- 75-1 (human prostatic carcinoma), H-301 (Syrian hamster kidney tumor), GHi or GH₃ (rat pituitary tumor), GH C] (rat pituitary tumor), or HT-29 (human colonic cancer).

0062Li still other embodiments of the present invention an isolated estrogen receptor gamma (ERγ) is provided. Li certain embodiments the (ERγ) has an estradiol binding affinity greater than that of estrogen receptor alpha (ERα) or estrogen receptor beta (ERβ), preferably having a K_dfor E₂ on the order of >10^"9 M. The ERγ also preferably has specificity for steroid hormone binding in the order estradiol » diethylstilbestrol » testosterone = dihydrotestosterone, and has a molecular weight of approximately 50 kDa. In certain embodiments a mediator of estrogen responsive cell growth comprises ERγ, and in certain embodiments a mediator of estrogen reversal of immunoglobulin inhibition of estrogen responsive cell growth comprises ERγ.

0063Also provided by the present invention are methods of detecting an estrogenic substance. According to certain embodiments, the method comprises (a) maintaining a predetermined population of estrogen responsive cancer cells in a steroid hormone-free nutrient medium comprising a quantity of immunoglobulin inhibitor sufficient to inhibit cancer cell growth in the absence of an inhibition-reversing amount of estrogen, the cells also being estrogen responsive for proliferation in vivo when implanted into a suitable host; (b) adding a defined amount of the substance of interest to the cells and medium, to yield a test culture; (c) incubating the test culture for a predetermined period of time under cell growth promoting conditions; and (d) determining the cell population in the test culture after the predetermined period of time, a measurable increase in the cell population indicating an estrogen-like cell growth stimulating effect by the substance of interest. In some embodiment the method also includes testing the substance of interest for binding to estrogen receptor gamma and/or testing for cytotoxic effects. Li certain embodiments, the method includes selecting estrogen responsive cancer cells containing estrogen receptor gamma. Also provided by the present invention are methods of detecting an anti-estrogenic substance, such as an antagonist. According to certain embodiments, the method comprises (a) maintaining a predetermined population of estrogen responsive cancer cells in a nutrient medium comprising a quantity of immunoglobulin inhibitor sufficient to inhibit cell growth in the absence of an inhibition-reversing amount of estrogen, the cells being capable of growing in vivo; (b) adding a defined amount of the substance of interest to the cells and medium; (c) adding a defined amount of an estrogen sufficient to stimulate cell growth in the presence of the inhibitor and in the absence of the substance of interest to the cells and medium, to yield a test culture; (d) incubating the test culture for a predetermined period of time under cell growth promoting conditions; (e) testing the substance of interest for cytotoxic effects on the cells; and (f) determining the cell population in the test culture after the predetermined period of time, a lack of measurable increase in the cell population not attributable to cytotoxic effects of the substance indicating a steroid hormone antagonistic effect by the substance of interest. Li some embodiments the method also includes testing the substance of interest for binding to estrogen receptor gamma and/or testing for cytotoxic effects. In certain embodiments, the method includes selecting estrogen responsive cancer cells containing estrogen receptor gamma.

0064Also provided in accordance with the present invention are methods of identifying an estrogen responsive cell that is capable of being inhibited or prevented from proliferating by an estrogen reversible inhibitor of estrogen responsive cell growth. Li certain embodiments the method comprises detecting estrogen receptor gamma in the cell. According to the present invention, methods of inhibiting in vitro cancer cell growth are provided. Li certain embodiments the method comprises (a) maintaining a predetermined population of cancer cells in an above-described nutrient medium; (b) adding an effective amount of an iron compound to the medium, to provide an incubation mixture comprising unbound Fe (III), preferably at least about 1 μM Fe (III); (c) incubating the incubation mixture for a predetermined period of time under cell growth promoting conditions; and (d) determining the cell population in the incubation mixture after the predetermined period of time, an increase in cell population indicating lack of inhibition by the Fe (HI), and the absence of an increase in cell population indicating inhibition of cell growth by the Fe (IS).

0065In certain embodiments of the present invention, a method of killing cancer cells in vitro is provided. In some of those embodiments a concentration of at least about 10 μM unbound Fe (III) is maintained in the nutrient medium. Alternatively, extended arrest of cancer cell growth by an immunoglobulin inhibitor can also serve to kill steroid hormone responsive cancer cells in culture.

0066Accordingly, in certain embodiments of the present invention, a method of killing steroid hormone responsive cancer cells in culture is provided which comprises (a) combining a predetermined population of steroid hormone responsive cancer cells with a nutrient medium comprising an above-described cell culture medium and a quantity of steroid hormone irreversible immunoglobulin inhibitor sufficient to inhibit cell growth of steroid hormone responsive cancer cells, to provide an incubation mixture, the steroid hormone responsive cells also being steroid hormone responsive for proliferation in vivo when implanted into a suitable host; (b) incubating the incubation mixture for a predetermined period of time under cell growth promoting conditions; and (c) optionally, determining the cell population in the reaction mixture after the incubation for the predetermined period of time. Li some embodiments the immunoglobulin inhibitor is irreversibly, or permanently inhibitory (i.e., the inhibitor is active with respect to the ability to inhibit steroid hormone-responsive cell proliferation and inactive with respect to steroid hormone reversibility of the inhibition.) Another embodiment of the present invention provides a method of killing a mixed population of steroid hormone responsive cancer cells and autonomous cancer cells. The method comprises contacting the mixed population of cells with an amount of an iron depleting substance sufficient to substantially deprive the autonomous cells of Fe (JS), and then maintaining the cells in an iron depleted environment for a sufficient period of time for the autonomous cells to die. The method also includes contacting the mixed population of cells with an amount of a Fe ( S) containing substance sufficient to inhibit cell growth and/or kill the steroid hormone responsive cells, and then maintaining the cells in a Fe (iπ)-enhanced environment for a predetermined period of time sufficient to inhibit cell growth and/or kill the steroid hormone responsive cancer cells. Li certain embodiments, the method also includes contacting the mixed population of cells with an amount of immunoglobulin inhibitor sufficient to inhibit proliferation of the steroid hormone responsive cells. Still other embodiments provided by the present invention are methods of determining the concentration of a steroid hormone in a defined amount of a body fluid. In certain embodiments the method comprises assaying the body fluid for binding of steroid hormone to an immunoglobulin inhibitor of steroid hormone responsive cancer cell growth. These and other embodiments, features and advantages of the present invention will become apparent with reference to the following description and drawings. BRTEF DESCRIPTION OF THE DRAWINGS

0067For the detailed descriptions of the preferred embodiments, reference will now be made to the accompanying figures which include graphs, charts, and test results:

0068Fig. 1. Effect of Temperature on ³H-E₂ Binding to MTW9/PL2 Cells. Fig. 2. Scatchard Analysis and Saturation Binding Analysis of ³H-E₂ Binding to

0069MTW9/PL2 Cells.

0070Fig. 3. Effect of Unlabeled Competitor Steroids on ³H-E₂ Binding to MTW9/PL2 Cells. (A) Competition with Unlabeled Androgens; (B) Competition with Unlabeled Progesterone and Cortisol. Fig.4. Effect of Temperature on ³H-Progesterone Binding to MTW9/PL2 Cells.

0071Fig. 5. Scatchard Analysis and Saturation Binding Analysis of ³H-Progesterone to MTW9/PL2 Cells.

0072Fig. 6. Effect of Unlabeled Competitor Steroids on ³H-Progesterone Binding to MTW9/PL2 Cells. Fig.7. Estrogen Induction of Progesterone Receptors with MTW9 PL2 Cells.

0073Fig. 8. Western immunoblotting Analysis of Androgen, Progesterone and Estrogen Receptors in MTW9/PL2 cells. Lanes 1 through 8 contain lOμg of the following cell extract protein: Lanes 1 and 2, cytosolic extracts of rat and human fibroblasts, respectively; Lanes 3 and 4, cytosolic and nuclear extracts, respectively, of MTW9/PL2 Cells; Lanes 5 and 6, cytosolic and nuclear extracts, respectively, of 47D Cells; Lanes 7 and 8, cytosolic and nuclear extracts, respectively, of LNCaP cells. Top, Middle and Bottom Panels are Androgen, Progesterone and Estrogen receptors, respectively.

0074Fig. 9. CDE-horse Serum Effect on MTW9/PL2 Cell Growth + 10 nM E₂ for 7 days. (A) Dose-response data expressed as cell numbers; (B) Dose-response data expressed as cell population doublings (CPD) per 7 days.

0075Fig. 10. Restoration of Growth by Addition of 10 nM E₂ on days 0, 2, 4 and 6 After Seeding the MTW9/PL2 cells into Fully Inhibitory Medium Containing 50% (v/v) of CDE-horse serum.

0076Fig. 11. Dose-Response Effects of Steroid Hormones on Growth of the MTW9/PL2 Cells in Medium Containing 50% (v/v) CDE-horse Serum.

0077Fig. 12. MTW9/PL2 Cell Growth ± E₂ in Medium with CDE Sera from Several Species. (A) CDE-porcine Serum; (B) CDE-pregnant Human Serum; (C) CDE-adult Rat Serum; (D) CDE- adult Bovine Serum; (E) CDE-fetal Bovine Serum; (F) CDE-fetal Horse Serum.

0078Fig. 13. CDE-horse Serum Effect on GHA Cell Growth + 10 nM E₂ for 10 days. Fig. 14. CDE-horse Serum Effect on ZR-75-1 Cell Growth + 10 nM Ej for 14 days.

0079Fig. 15. CDE-horse Serum Effect on MCF-7A Cell Growth + 10 nM Ez for 10 days. Fig. 16. Kinetics of T47D Cell Growth in CDE-horse Serum + 10 nM E₂. (A) Growth Kinetics in 20% CDE-horse ± E₂ versus 10% Fetal Bovine Serum; (B) Growth Kinetics in 50% CDE-horse Serum ± E₂.

0080Fig. 17. Rodent and Human ER⁺ Cell Growth in 50% CDE-human Serum ± E₂. (A) T47D Human Breast Cancer Cells; (B) LNCaP Human Prostate Cancer Cells; (C) MTW9/PL2 Rat Mammary Tumor Cells; (D) GH₃ Rat Pituitary Tumor Cells; (E) GEL Rat Pituitary Tumor Cells; (F) H301 Syrian Hamster Kidney Tumor Cells.

0081Fig. 18. Dose-Response of Steroid Hormones with T47D Cells in 50% CDE-horse Serum. Fig. 19. Dose-Response of Steroid Hormones with GEL Cells in 50% CDE-horse Serum.

0082Fig. 20. Dose-Response of Steroid Hormones with H301 Cells in 50% CDE-horse Serum. Fig. 21. Dose-Response of Steroid Hormones with LNCaP Cells in 50% CDE-horse Serum.

0083Fig. 22. T₃ Growth Effects with GH₃ Cells in Serum-free Medium (PCM). Fig. 23. E₂ Growth Effects with GH₃ Cells in Serum-free Medium (PCM) Minus E₂.

0084Fig. 24. T₃ Growth Effects with Three GH Cell Lines in 2.5% CDE-horse Serum. Fig. 25. T₃ Growth Effects with Two GH Cell Lines in 50% CDE-horse Serum. Fig. 26. Effect of 56°C Versus 34°C CDE-horse Serum on MTW9/PL2 Cell Growth. Fig.27. Effect of XAD-4 Resin Treated Horse Serum on MTW9/PL2 Cell Growth + E₂. Fig.28. Effect of XAD-4 Resin Treated Horse Serum on T47D Cell Growth + E₂.

0085Fig. 29. Effect of Phenol Red on Estrogen Responsive MCF-7 Cell Growth. (A) MCF- 7A Cell Growth in CDE-horse Serum + Phenol Red and + E₂; (B) Estrogenic Effects with MCF-7A Cells + Phenol Red; (C) MCF-7K Cell Growth in CDE-horse Serum ± Phenol Red and ± E₂ (D) Estrogenic Effects with MCF-7K Cells + Phenol Red. Fig. 30. Effect of Phenol Red on Estrogen Responsive T47D and ZR-75-1 Cell Growth.

0086(A) T47D Cell Growth in CDE-horse Serum ± Phenol Red and + E₂; (B) Estrogenic Effects with T47D Cells + Phenol Red; (C) ZR-75-1 Cell Growth in CDE-horse Serum + Phenol Red and ± Ε^; (D) Estrogenic Effects with ZR-75-1 Cells ± Phenol Red.

0087Fig. 31. Effect of Phenol Red on Estrogen Responsive MTW9/PL2 Cell Growth. (A) MTW9 PL2 Cell Growth in CDE-horse Serum + Phenol Red and + E₂; (B) Estrogenic Effects with MTW9 PL2 Cells ± Phenol Red.

0088Fig. 32. Dose-Response Effects of Phenol Red versus E₂ with Three ER⁺ Cell Lines. (A) Growth Effects of Phenol Red with MCF-7K, T47D and MTW9/PL2 Cells; (B) Growth Effects of E₂ with MCF-7K, T47D and MTW9/PL2 Cells. Fig. 33. Estrogen Induction of Progesterone Receptors by Phenol Red versus E₂. (A)

0089Induction by E₂ with T47D Cells; (B) Induction by Phenol Red with T47D Cells. Fig. 34. Effects of TGFβl on Cell Growth in 2.5% CDE-horse Serum ± E₂. (A) MCF-7K Cell Growth; (B) MTW9/PL2 Cell Growth.

0090Fig. 35. TGFβl Inhibition of ER⁺ Rodent and Human Cell Line Growth + E₂. (A) Inhibition Data ± E₂ Presented in Cell Number; (B) Inhibition Data ± E₂ Presented in CPD. Fig. 36. EGF and TGFα as Substitutes for the Effects of Ez in CDE-horse Serum. (A)

0091MCF-7A Cell Growth; (B) MCF-7K Cell Growth; (C) T47D Cell Growth; (D) ZR-75-1 Cell Growth.

0092Fig. 37. IGF-I as a Substitute for the Effects of E₂ in CDE-horse Serum. (A) MCF-7K Cell Growth MCF-7 A Cell Growth; (B) T47D Cell Growth.

0093Fig. 38. Growth of T47D Human Breast Cancer Cells in Standard and "low-Fe" D- MEM/F-12.

0094Fig. 39. Growth of LNCaP Human Prostate Cancer Cells in Standard and "low-Fe" D- MEM/F-12. Fig. 40. Growth of MDCK Dog Kidney Tubule Cells in Standard and "low-Fe" D-

0095MEM/F-12.

0096Fig. 41. Growth of AR⁺ LNCaP Cells in CAPM + DHT versus Growth in D-MEM/F-12 Containing 10% Fetal Bovine Serum.

0097Fig. 42 Growth of the AR^" DU145 and AR^" PC3 Cells in CAPM versus Growth in D- MEMZF-12 Containing 10% Fetal Bovine Serum.

0098Fig. 43. Dose-Response Effects of Individual Components of CAPM Serum-free Defined Medium on LNCaP Cell Growth.

0099Fig. 44. Effects of Deletion of Individual Components from CAPM Serum-free Medium on LNCaP, DU145 and PC3 Cell Growth ± DHT. Fig. 45. Effect of Fe (JS) on MCF-7A Cell Growth in DDM-2MF Serum-free Defined

0100Medium.

0101Fig. 46. Effect of Fe (JS) on T47D Cell Growth in DDM-2MF Serum-free Defined Medium .

0102Fig.47. Effect of Fe (JS) on LNCaP Cell Growth in CAPM Plus Apotransferrin. Fig. 48. Comparative Effect of Fe (JS) on LNCaP, DU145 and PC3 Cell Growth in

0103CAPM.

0104Fig. 49. Growth Restoring Effect of Fe (IS) Chelators in serum-free medium with T47D Cells.

0105Fig. 50. Growth Restoring Effect of Fe (JS) Chelators in serum-free medium with LNCaP Cells Fig. 51. Comparison of DU145 Cell Growth in "low-Fe" and "standard" D-MEM/F-12 Based Serum-free Defined Medium CAPM.

0106Fig. 52. Comparison of PC3 Cell Growth in "low-Fe" and "standard" D-MEM/F-12 Based Serum-free Defined Medium CAPM. Fig. 53. Growth of the DU145 Cells in CDE-horse Serum ± DHT.

0107Fig. 54. Growth of the PC3 Cells in CDE-horse Serum + DHT.

0108Fig. 55. Growth of the ALNA-41 Cells in CDE-horse Serum + DHT.

0109Fig. 56. Comparison of Estrogenic Effects in Serum-free Defined Medium and in D- MEM/F-12 Medium Supplemented with CDE-Horse Serum. (A) MCF-7K Cell Growth in Serum- free Defined Medium + Ez; (B) MCF-7K Cell Growth in D-MEM/F-12 with CDE-horse Serum ± E₂; (C) T47D Cell Growth in Serum-free Defined Medium + J^; (D) T47D Cell Growth in D-MEM/F- 12 with CDE-horse Serum + E₂; (E) LΝCaP Cell Growth in Serum-free Defined Medium + E₂; (F) LΝCaP Cell Growth in D-MEM/F-12 with CDE-horse Serum + Ea.

0110Fig. 57. Comparison of Estrogenic Effects in Serum-free Defined Medium and in D- MEM/F-12 Medium Supplemented with CDE-Horse Serum. (A) GHA Cell Growth in Serum-free Defined Medium ± E₂; (B) GHA Cell Growth in D-MEM/F-12 with CDE-horse Serum ± E₂; (C) MTW9/PL2 Cell Growth in Serum-free Defined Medium + E₂; (D) MTW9/PL2 Cell Growth in D- MEM/F-12 with CDE-horse Serum ± ^; (E) H301 Cell Growth in Serum-free Defined Medium ± E₂; (F) H301 Cell Growth in D-MEM/F-12 with CDE-horse Serum + JEfe. Fig. 58. Effect of CDE-horse Serum on LΝCaP Cell Growth in Serum-free CAPM ± E₂ and ± DHT.

0111Fig. 59. Comparison of the Inhibitor Reversing Effects of DHT, E₂, and DES on LΝCaP

0112Cell Growth in CDE-horse Serum Containing Medium. (A) Effect of DHT as an Inhibitor Reversing

0113Steroid; (B) Effect of E₂ as an Inhibitor Reversing Steroid; (C) Effect of DES as an Inhibitor Reversing Steroid; (D) Effect of Combinations of DHT, E₂, and DES as Inhibitor Reversing Steroids.

0114Fig. 60. Effect of Tris Buffer (pH 7.4) Dialysis on the Estrogen Reversible Inhibitor Activity of CDE-horse Serum Assayed with MTW9/PL2 Cell ± E₂.

0115Fig. 61. Ultrafiltration of CDE-horse Serum and Assay of the Filtrate and Retentate with MTW9/PL2 Cells ± E₂. Fig. 62. 50°C Treatment of CDE-horse Serum for 30 minutes and Assay with

0116MTW9/PL2 Cells ± E₂.

0117Fig. 63. Time Course of Heat Treatment of CCDE-horse serum at 50°C and Measurement of Estrogenic Effects with MTW9/PL2 Cells.

0118Fig. 64. 50°C Treatment of CDE-horse Serum for 20 hours and Assay with MTW9/PL2 Cells + E₂. Fig. 65. 60°C Treatment of CDE-horse Serum for 90 minutes and Assay with MTW9/PL2 Cells + E₂.

0119Fig. 66. Affi-Gel Blue Treatment of CDE-horse Serum and Assay with MTW9/PL2 Cells

0120Fig.67. Effect of 6 M Urea on the Estrogenic Activity of CDE-horse Serum Assayed with

0121MTW9/PL2 Cells + E₂.

0122Fig. 68. ED₅₀ Estimations for Purification Quantification of Beginning with Serum. (A) MCF-7K Cells ED₅₀ of CDE-horse Serum ± E₂; (B) ZR-75-1 Cells ED₅₀ of CDE-horse Serum ± E₂; (C) MTW9/PL2 Cells ED₅₀ of CDE-horse Serum ± E₂; (D) GH-A Cells ED₅₀ of CDE-horse Serum + E₂.

0123Fig. 69. Assay of Estrogenic Activity (ED₅₀) of Chromatographic Pools. (A) Ammonium Sulfate Active Fraction; (B)Affi-Gel BlueGel Albumin Rich Fraction; (C) DEAE Sepharose Pool IN Active Fraction Assay #1; (D) DEAE Sepharose Pool IV Active Fraction Assay #2.

0124Fig.70. Assay of Affi-Gel BlueGel By-Pass Fraction + E₂. Fig. 71. DEAE Sepharose Chromatography Elution Profile with Whole CDE-horse

0125Serum.

0126Fig. 72. Phenyl Sepharose Chromatography Elution Profile with DEAE Sepharose Pool

0127IN.

0128Fig.73. HTP Bio-Gel (hydroxylapatite) Elution Profile with DEAE Sepharose Pool IV. Fig. 74. MTW9/PL2 Cell Assay of CDE-horse Serum Estrogenic Activity after Dialysis in Tris-HCl, pH 7.4, plus 50 mM CaCl₂.

0129Fig.75. Effect of Calcium on the 50°C Heat Stability of the Estrogenic Activity in Chelex Treated CDE-horse Serum Assayed with MTW9/PL2 Cells.

0130Fig. 76. Effect of Zn, Mn, Mg and Ca on the 37°C Heat Stability of the Estrogenic Activity in Chelex Treated CDE-horse Serum Assayed with MTW9/PL2 Cells. Fig. 77. Binding Affinity (K_d) of ³H-DHT to CDE-horse Serum.

0131Fig. 78. Calcium Protection of both the Estrogenic Effect with MTW9/PL2 Cell and the Binding of ³H-DHT with Chelex Treated CDE-horse Serum. (A) Estrogenic Effect Protection by Calcium; (B) Calcium Protection of ³H-DHT Binding. Fig. 79. Lnmunoprecipitation of ³H-DHT Binding and Estrogenic Activity of CDE-horse

0132Serum by Anti-Human SHBG. (A) ³H-DHT Binding Reduction; (B) Estrogenic Activity Reduction. Fig. 80. Column Elution Profiles of the Two-step Cortisol finity and Phenyl Sepharose Elution of CA-PA-pool I and CA-PS-pool π.

0133Fig. 81. Identification of the Molecular Forms Present in Active CA-PS-pool II. (A) SDS-PAGE with Coomassie Blue Staining; (B) Western Analysis with Anti-human SHBG. Fig. 82. CA-PS-pool H Effect on ER⁺ Cell Growth in 2.5% CDE-horse Serum + E (A) GH] Cells; (B) GH₃ Cells; (C) GKA Cells; (D) H301 Cells; (E) MTW9/PL2 Cells; (F) MCF-7K Cells; (G) ZR-75-1 Cells; (H) T47D Cells.

0134Fig. 83. Cortisol Affinity Column Depletion of the Estrogenic Activity in CDE-horse Serum Assayed with ER⁺ Cell Lines ± E₂. (A) T47D Cells Pre-Column; (B) T47D Cells Post- Column; (C) GH₃ Cells Pre-Column; (D) GH₃ Cells Pre-Column; (E) H301 Cells Pre-Column; (F) H301 Cells Post-Column.

0135Fig. 84. Serum-free Growth of Cells in Four Different Defined Media ± E₂. (A) MTW9/PL2 Cells in DDM-2A; (B) T47D Cells in DDM-2MF; (C) GHA Cells in PCM-9; (D) H301 Cells in CAPM.

0136Fig. 85. Effects of CDE-horse Serum on Estrogen Responsiveness of Three ER ⁺ Cell Lines Growing in Serum-free Defined Media. (A)T47D Cells in DDM-2MF; (B) MTW9/PL2 Cells in DDM-2A; (C) GH₄d Cells in PCM-9.

0137Fig. 86. Effects of CA-PS-pool U on the Growth of Eight ER⁺ Cell Lines in Serum-free Defined Medium ± E₂.

0138Fig. 87. Protein Sequencing Results with CA-PS-Pool II Peptides and Homology to Human SHBG, Rabbit SHBG and Rat and Hamster Androgen Binding Protein.

0139Fig. 88. Western Analysis of CA-PS-pool I and CA-PS-pool II with the Antibody Raised to the 54kDa Band. Fig. 89. Effect of the Anit-54kDa Antiserum on the Inhibition of MWT9/PL2 Cell Growth by the Isolated Fraction CS-PS-Pool H

0140Fig. 90. Western Immunoblotting of Commercially Prepared Horse IgG, IgA and IgM with anti-54 kDa Antiserum.

0141Fig. 91. Effect of Horse IgG on MTW9/PL2 Cell Growth in 2.5% CDE-horse Serum ± E₂. Fig. 92. Effect of Horse IgM on MTW9/PL2 Cell Growth in 2.5% CDE-horse Serum ±

0142E₂

0143Fig. 93. Effect of Horse IgA on MTW9/PL2 Cell Growth in 2.5% CDE-horse Serum +

0144E₂

0145-Fig. 94. SDS-PAGE with Coomassie Staining and Western Analysis of Rat Purified "SHBG-like" Proteins. (A) SDS-PAGE of Purified Rat Preparations; (B) Western Analysis with Anti-rat IgG.

0146Fig. 95. Western Analysis of a Rat Purified "SHBG-like" Preparation. (A) Western with Anti-rat IgA with Purified IgA Control; (B) Western with Anti-rat IgGl with Purified IgGl Control; (C) Western with Anti-rat IgM with Purified IgM Control. Fig. 96. Protein Sequencing Results with Rat "SHBG-like" Peptides and Homology to

0147Human SHBG, Rabbit SHBG and Rat and Hamster Androgen Binding Protein. Fig. 97. Comparison of Rat IgG Subclasses. (A) SDS-PAGE with Coomassie Blue Staining; (B) Western Analysis with Rabbit Anti-Human SHBG.

0148Fig. 98. Effect of Rat IgG on MTW9/PL2 Cell Growth in Medium with 2.5% CDE-rat Serum ± E₂. Fig. 99. Effect of Rat IgA on MTW9/PL2 Cell Growth in Medium with 2.5% CDE-rat

0149Serum + E₂.

0150Fig. 100. Effect of Rat IgM on MTW9/PL2 Cell Growth in Medium with 2.5% CDE- rat Serum ± E₂.

0151Fig. 101. Mannan Binding Protein Isolation of Human Plasma/Serum IgM. Fig. 102. Jacalin Lectin Purification of Human Plasma/Serum IgA.

0152Fig. 103. Effect of Human IgM on MTW9/PL2 Cell Growth ± E₂ in Serum-free Defined Medium.

0153Fig. 104. Comparison of the Effects of Rat and Horse IgA and IgM on MTW9/PL2 Cell Growth ± E₂ in Serum-free Defined Medium Expressed in Cell Number and CPD. Fig. 105. Effect of Rat Myeloma IgA on GH] Cell Growth in Serum-free Defined Medium

0154+ E₂.

0155Fig. 106. Effect of Human Plasma IgA on GH] Cell Growth in Serum-free Defined

0156Medium ± E₂.

0157Fig. 107. Effect of Human Plasma IgM on GH] Cell Growth in Serum-free Defined Medium + E₂.

0158Fig. 108. Effects of slgA on GH] Cell Growth in Serum-free Defined Medium + E₂.

0159Fig. 109. Model of Mucosal Epithelial Cell Transport of IgA/IgM.

0160Fig. 110. Essential Structures of Human Plasma and Secretory IgA.

0161Fig. 111. Effect of Rat Myeloma IgA on GH₃ Cell Growth in Serum-free Defined Medium + E₂.

0162Fig. 112. Effect of Rat IgM on GH₃ Cell Growth in Serum-free Defined Medium ± E₂.

0163Fig. 113. Effect of Human Plasma IgA on GH₃ Cell Growth in Serum-free Defined

0164Medium + E₂.

0165Fig. 114. Effect of Human Plasma IgM on GH₃ Cell Growth in Serum-free Defined Medium ± Ez.

0166Fig. 115. Effect of Human Secretory IgA on GH₃ Cell Growth in Serum-free Defined

0167Medium ± E^.

0168Fig. 116. Effect of Rat Myeloma IgA on GHACell Growth in Serum-free Defined

0169Medium + E^. Fig. 117. Effect of Rat Plasma IgM on GHACell Growth in Serum-free Defined Medium Fig. 118. Effect of Human Plasma IgA on GHACell Growth in Serum-free Defined Medium ± E₂.

0170Fig. 119. Effect of Human Plasma IgM on GHACell Growth in Serum-free Defined Medium ± E_∑. Fig. 120. Effect of Human Secretory IgA on GHACell Growth in Serum-free Defined

0171Medium ± E₂.

0172Fig. 121. Effect of Mouse IgA on H301 Cell Growth in Serum-free Defined Medium ± E₂.

0173Fig. 122. Effect of Human IgA on H301 Cell Growth in Serum-free Defined Medium + F^ (A) Plasma IgA Effects; (B) Secretory slgA Effects. Fig. 123. Dose-Response Effects of Ea on H301 Cell Growth in Serum-free Defined

0174Medium Containing 40 μg/mL Human Plasma IgM.

0175Fig. 124. Effect of Human IgA on MCF-7A Cell Growth in Serum-free Defined Medium ± E₂. (A) Plasma IgA Effects; (B) Secretory slgA Effects.

0176Fig. 125. Effect of Human IgA on MCF-7K Cell Growth in Serum-free Defined Medium ± Ea. (A) Plasma IgA Effects; (B) Secretory slgA Effects.

0177Fig. 126. Effect of Human IgM on MCF-7A Cell Growth in Serum-free Defined Medium

0178Fig. 127. Effect of Human IgM on MCF-7K Cell Growth in Serum-free Defined Medium ± E,. Fig. 128. Dose-Response Effects of J . on MCF-7K Cell Growth in Serum-free Defined

0179Medium Containing 40 μg/mL Human Plasma IgM.

0180Fig. 129. Effect of Human IgA on T47D Cell Growth in Serum-free Defined Medium ± E₂. (A) Plasma IgA Effects; (B) Secretory slgA Effects.

0181Fig. 130. Effect of Human IgM on T47D Cell Growth in Serum-free Defined Medium ± E₂.

0182Fig. 131. Dose-Response Effects of E₂ on T47D Cell Growth in Serum-free Defined Medium Containing 40 μg/mL Human Plasma IgM.

0183Fig. 132. Effect of Human IgA on ZR-75-1 Cell Growth in Serum-free Defined Medium ± E₂. (A) Plasma IgA Effects; (B) Secretory slgA Effects. Fig. 133. Effect of Human IgM on ZR-75-1 Cell Growth in Serum-free Defined Medium

0184± E₂.

0185Fig. 134. Effect of Human IgM on HT-29 Cell Growth in Serum-free Defined Medium ± T₃.

0186Fig. 135. Effect of Human IgA on LNCaP Cell Growth in Serum-free Defined Medium ± E₂. (A) Plasma IgA Effects; (B) Secretory slgA Effects. Fig. 136. Effects of Human Plasma versus Human Myeloma IgM on LNCaP Cell Growth in Serum-free Defined Medium ± DHT.

0187Fig. 137. Summary of Estrogenic Effects with Various ER⁺ Cell lines and Different Ig Sources. Fig. 138. Effect of Tamoxifen on T47D Cell Growth in Serum-free Defined Medium

0188Fig. 139. Estrogen Reversal of Tamoxifen Inhibition of T47D cells in Serum-free Defined Medium

0189Fig. 140. Effect of Rat Immunoglobulins on Estrogen Responsive Growth of MTW9/PL2 Cells Li Serum-free Defined Medium. Fig. 141. Comparison of the Estrogenic Effects of Human Immungobulin with T47D Cells in Serum-free Defined Medium.

0190Fig. 142. Effect of Human IgG Isotypes on LNCaP Cell Growth in Serum-free Defined Medium ± DHT.

0191Fig. 143. Western Detection of the Secretory Component of Human Milk slgA. Fig. 144. Effect of Anti-Secretory Component on IgM Inhibition of T47D Cell Growth in

0192Serum-free Defined Medium.

0193Fig. 145. Effect of Anti-Secretory Component on plgA Inhibition of LNCaP Cell Growth in Serum-free Defined Medium.

0194Fig. 146. Western Analysis with Anti-Secretory Component to Detect the Poly-Ig Receptor in AR⁺ and AR^" Prostate Cancer Cells plus Control Cell Lines.

0195Fig. 147. Effect of Human plgA on DU145 Cell Growth in Serum-free Defined Medium +

0196DHT.

0197Fig. 148. Effect of Human plgA on PC3 Cell Growth in Serum-free Defined Medium +

0198DHT.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

0200To facilitate review of the detailed description of preferred embodiments, a Table of Contents is provided. The titles used for the various subsections and examples are not intended to be limiting and are only an aid to locating certain subject matter. Li addition, each Example begins with a short summary of that Example, which is intended merely to facilitate review and is not limiting on the disclosure contained in the full Example, and ends with a Discussion of some conclusions that may be drawn from that Example. Table of Contents Subsection Page No.

0201I. Introduction 38 H. General Materials and Methods 40 in. Examples

0202Example 1. Identification of Steroid Hormone Receptors in MTW9/PL2 Cells 51

0203Example 2. Three Preparations of Steroid Hormone Depleted Serum and Examples of

0204Support of Estrogen Responsive Cell Growth in Culture 58 A. Charcoal-dextran Extraction at 34°C 59

0205B. Charcoal-dextran Extraction at 56°C 60

0206C. Amberlite™ XAD™-4 Resin Treatment 60

0207Example 3. Cancer Cell Line MTW9/PL2 Exhibits Estrogen Responsiveness in 34°C Charcoal-dextran Extracted Serum 62 Example 4. Estrogen Responsive Growth of Additional Rodent and Human Cell Lines Li 34°C Charcoal-dextran Extracted Horse and Human Serum 69

0208Example 5. Thyroid Hormone Growth Effects in CDE-Horse Serum Prepared at 34°C 75

0209Example 6. Effect of 56°C Versus 34°C CDE-horse Serum on MTW9/PL2 Cell Growth 76 Example 7. Demonstration of Estrogenic Effects in XAD-4 Resin Treated Horse Serum 77

0210Example 8. Testing of Substances for Estrogenic Activity 78

0211Example 9. Testing of Substances for Inhibitor-like Activity 84

0212Example 10. Serum-free Defined Culture Medium Formulations 92

0213Example 11. Serum-free Defined Medium that Supports Hormone Sensitive and Autonomous Cancer Cell Growth 101

0214Example 12: Differential Effects of Fe (JS) on the Growth of Hormone Responsive and

0215Autonomous Human Breast and Human Prostate Cancer Cells 105

0216Example 13: Growth in Serum-free Defined Medium versus Growth in CDE-Serum + E2 108

0217Example 14: Action of DES on Human AR + LNCaP Prostate Cancer Cells 111 Example 15: Preparation of Inhibitor Depleted Serum for Control Studies and

0218Stability Properties of the Inhibitor 112

0219Example 16: Effects of Conventional Purification Methods and Properties of the Estrogen Reversible Serum-borne Inhibitor 114

0220Example 17: Calcium Stabilization and Correlation with 3H-DHT Binding and Immunoprecipitation by Antibodies Raised to Human SHBG 120 Example 18: Cortisol Affinity and Phenyl Sepharose Isolation of the "SHBG-like"

0221Estrogen Reversible Inhibitor from CDE-Horse Serum 122

0222Example 19: Serum-free Assay Systems for Measuring Large Magnitude Steroid Hormone

0223Mitogenic Responses with the Two-Step Purified Inhibitor 129 Example 20: Chemical and Immunological Properties of the Partially Purified

0224CA-PS-Pool II Inhibitors and Identification as IgA and IgM 132

0225Example 21 : Regulation of Steroid Hormone-responsive and Thyroid Hormone-responsive Cancer Cell Growth in Serum-free Defined Medium by Secretory and Plasma

0226Forms of IgA and Plasma and Cell Culture Derived IgM 138 A. MTW9/PL2 Cells - ER+ rat mammary tumor 138

0227B. GH1, GH3, and GH4Cl Cells - ER+ rat pituitary tumor 139

0228C. H301 Cells -ER+ Syrian hamster kidney tumor 140

0229D. MCF-7A and MCF-7K Cells - ER+ human breast cancer 141

0230E. T47D Cells - ER+ human breast cancer 141 F. ZR-75-1 Cells - ER+ human breast cancer 142

0231G. HT-29 Cells - Thyroid hormone responsive human colon cancer 142

0232H. LNCaP Cells - AR+ human prostate cancer 142

0233Example 22: Effect of Tamoxifen Antiestrogen in Serum-free Defined Medium 144 Example 23: IgGl and IgG2 as an Immunoglobulin Regulators of Estrogen and

0234Androgen Responsive Cancer Cell Growth 145

0235Example 24: Mediation of IgA/IgM Effects by the Poly-Ig Receptor 147

0236Example 25: Mediation of IgGlκ Effects by a Fc-like Receptor 151

0237Example 26. Immunoglobulin Inhibitors as Tools for Identifying the Receptors that Mediate the IgA/IgM/IgG Cell Growth Regulating Effects 153

0238Example 27: Conceptual Model for Cascading Loss of Cell Growth Inhibition in Cancer Cells.157 Example 28. IgA/IgM Based Test to Detect Lowered Levels of Steroid Hormone Reversible

0239Cell Growth Inhibitors in Plasma or Body Secretions 161

0240Introduction

0241Extracellular negative regulation is a key control mechanism of cell proliferation in steroid hormone responsive cancer cells. Sex steroid hormones (both estrogens and androgens) act to reverse the effects of a serum-borne inhibitor(s) that normally blocks target cell proliferation

0242(Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427; Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446; Moreno-Cuevas JE and Sirbasku

0243DA (2000) In Vitro Cell Dev Biol 36, 447-464, incorporated herein by reference). As demonstrated in the Examples that follow, these results were obtained with nine different estrogen receptor alpha (ERα) (Kumar V et al. (1987) Cell 51, 941-951) containing cell lines representing four target tissues and three species (Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446).

0244As mentioned in the Background of the Invention, the prior art fails to adequately address the issues of (i) whether there are one or more of the serum-derived inhibitors, (ii) what is/are the exact chemical composition of the inhibitor(s), and (iii) what conditions were required to yield the long term stable product(s) necessary for the commercial application of the testing methodology described. Methods and compositions are presented herein that are useful for testing and assessment of compounds and mixtures for estrogenic or androgenic activity as well as others possessing antiestrogenic and antiandrogenic activities. Li the Examples that follow, cell culture methodology and compositions are described that permit testing at concentrations lower than was previously possible using existing methodologies. Moreover, the new in vitro model assay systems obviate the need to conduct animal testing to predict in vivo responses. Some practical applications for the model include protecting the human population from unrecognized exposure to hormone-like compounds that present health hazards as well as developing new antihormone compounds to counterbalance these hazards. The testing of these compounds and mixtures are preferably conducted in serum-containing medium to mimic the conditions encountered by a blood borne agent. Testing can additionally be done under completely serum-free defined medium conditions to determine direct actions on cells without serum or non-essential proteins present. It has been discovered that the negative regulators of steroid hormone responsive cancer cell growth (estrogen reversible inhibitors) in serum are products from the secretory immune system, i.e., the immunologobulins A (IgA), M (IgM) and IgGl. These "immunoglobulin inhibitors" act as steroid hormone and thyroid hormone reversible inhibitors of mucosal cell growth. There has been no previous identification of secretory immune system immunoglobulins as regulators of epithelial (mucosal) cell growth, and this discovery is unique in the cell growth regulation field. Application of certain of the compositions and methods is expected to relate to 80% of all human cancers because this high incidence rate arises from mucosal tissues. There is no previously reported evidence directly linking the secretory immunoglobulins with regulation of mucosal cell growth.

0245With regard to applicability to several mucosal tissues, it is recognized that breast and prostate cancers are very similar diseases. Aside from tissue specific epidemiological and social factors, breast and prostate cancers have remarkable parallels (Grody WW et al. (1994) Am J Gin Pathol 102, S1-S67). The secretory immune system acts as a sex steroid hormone reversible inhibitor with target tumor cells from both of these cancers. Both are adenocarcinomas arising from sexually differentiate tissues. Certainly both cancers are very common in North America and northern Europe compared to the rest of the world. Both are strongly influenced by steroid hormones. Both increase in incidence with age. Both are thought to have at least some genetic component. Finally, both have very similar patterns of development when examined histologically. These facts also have implications with regard to colon, uterine and ovarian cancers. These cancers show familial clustering with breast cancer (Nelson CL et al. (1993) Genet Epidemiol 10, 235-244). The aggregation of colon, ovarian, endometrial and breast cancer in families has been described as a "cancer family" which now has the name Lynch Syndrome I and II (Lynch HT et al. (1978) Cancer 41, 1543-1549). Other studies have shown links between colorectal cancer and breast cancer (Rozen P et al. (1990) Cancer Lett 55, 189-194) and colorectal cancer and breast, uterine and ovarian cancer (Rozen P et al. (1986) Cancer 57, 1235-1239). It is clear that the incidence of these several mucosal origin cancers are linked and that this linkage has not been explained.

0246H. General Materials and Methods

0247Li the Examples below, which describe representative, preferred embodiments of the present invention, the following general materials and methods are employed, except as otherwise noted in the Examples.

0248Cell Culture Medium. The water used to prepare culture media and all other solutions was purified first by reverse osmosis followed by passage through a U. S. Filter Corporation system with a charcoal filter and two mixed bed ion exchangers. The effluent was distilled using a Bellco glass apparatus with quartz heating elements. The distilled water was stored in airflow restricted glass containers. No metal fittings are allowed in contact with the final purified water. This necessary precaution minimizes recontamination with metal ions. Standard phenol red containing Ham's F12-Dulbecco's modified Eagle's medium (D-MEM F-12), phenol red-free standard D- MEM/F-12 and a custom-prepared "low-Fe" D-MEM/F-12 medium were supplied by Gibco-BRL (Catalog No. 11330-032) or Bio^Whittacker (Catalog No. 12-719, liquid). The "low-Fe" medium was standard phenol red containing D-MEM/F-12 from which the usual additions of ferric nitrate and ferrous sulfate had been omitted (Eby E et al. (1992) Anal Biochem 203, 317-325; Eby JE et al. (1993) J Cell Physiol 156, 588-600). This medium was a special formulation purchased from Gibco- BRL as a powder and prepared in the highly purified water before 0.2μm pore filter membrane sterilization. A number of other stock solutions are required for cell culture in either serum containing or serum-free defined medium. Descriptions of each preparation are provided along with specific instructions for their use. The solutions used were designed to minimize the exogenous content of steroid hormone and to minimize the Fe (JS) content of the water. Steps are taken for the exclusion of all extraneous sources of steroid hormones and Fe (JS). Exclusion of Fe (JS) is highly preferred, and in most of the totally serum-free applications, it is considered essential. Wherever possible, disposable plastic ware or glassware is used to minimize potential contamination. It is important to note that excess solutions are preferably discarded after use with each individual cell line to avoid cross-contamination of cell types (Nelson-Rees WA and Fladermeyer RR (1977) Science (Wash DC) 195, 134-136). General Cell Culture - Serum. Adult and fetal horse, adult pig, adult sheep and adult and fetal bovine serum were obtained from Gibco-BRL. A mixture of adult male and female rat serum was purchased from Pel-Freez, Rodgers, AR. Human serum was purchased from Bio^Whittacker. Human plasma was a pool of samples collected from pregnant females during routine visits to a local clinic. All serum was stored frozen at -20°C until used. Repeated freeze-thaw of serum or plasma is avoided. Before charcoal extraction, the EDTA was removed by dialysis at 7°C for 24 hours against forty volumes of 0.05 M Tris-HCl, pH 7.4, containing 50 mM CaCl₂. Dialysis was done with Spectropor 1 membranes (Spectrum Medical Industries, molecular weight cut-off 6,000 to 8,000). The clotted material was removed by centrifugation. This preparation is termed plasma-derived serum. The serum or plasma was not heat pre-treated, or heat inactivated prior to use in the methods described below.

0249General Cell Culture - Normal Saline. Sterile normal saline (0.15 M NaCl) was prepared in 10 mL aliquots and stored at room temperature. Unused portions are discarded at the end of each experiment. A large supply is sterilized by autoclaving and used to prepare the solutions described below.

0250General Cell Culture - Trypsin/EDTA for Subculture. Sterile preparations were purchased from Irvine Scientific (Catalog No. 9341) or BioΦWhittacker (Trypsin-Versene EDTA Mixture) (Catalog No. 17-161F). This preparation contained 0.5g/L trypsin and 0.2 g/L EDTA in Hank's balanced salts solutions with lOmg/L phenol red. This preparation does not contain Ca or Mg salts nor does it have NaHC0₃. To trypsinize cells, 1.5 mL of this preparation was typically used. Aliquots (2 mL) were stored frozen until used and residual solution discarded at the end of each experiment or application to a cell line.

0251General Cell Culture - Soybean Trypsin Inhibitor (STR. STI was purchased from Sigma (Catalog No. T9128, Type H-2). An amount of 1.0 mg of this preparation will inactivate 1.0 mg of trypsin activity. The solution is prepared as 0.2% (w/v) in normal saline and sterilized using a 0.2 μm pore diameter filter membranes. Aliquots of 3.0 mL are stored at -20°C until used. This preparation is used to stop the action of trypsin during harvest of stock cultures for growth assays. STI ensures that all trypsin used to harvest cells for growth assays is inactivated and therefore will not damage the protein additions to serum-free defined medium. Also, use of STI ensures that no extraneous steroid hormones are introduced after harvest of cells from the stock culture dishes.

0252General Cell Culture - Crude Pancreatic Trypsin for Cell Counting. This trypsin preparation was used to harvest the cells for determining cell numbers. The cells are typically grown in 35-mm diameter dishes. This enzyme was purchased from ICN Biochemicals as the 1- 300 porcine pancreatic trypsin preparation (Catalog No. 103140). A stock solution is typically prepared by adding the contents of a preweighed bottle of IX Dulbecco's modified PBS medium without calcium or magnesium to 800 mL of water. This solution dissolves very gradually with adjustment to pH 7.3 using NaOH. After the solution was clear, 20 g of crude trypsin was added and this mixture stirred for 30 minutes at room temperature. The somewhat cloudy solution was diluted to 1000 mL with water and this volume was stored frozen in bulk overnight at -20°C to induce cold related precipitation that typically occurs when this preparation was frozen and thawed. After thawing at 37°C in a water bath, the preparation was filtered through 0.45 μm pore membranes. This preparation was stored at -20°C in useable portions.

0253General Cell Culture - EDTA for Cell Counting. The EDTA used is the disodium and dihydrate salt (Sigma Catalog No. E1644). A 0.29 M solution is prepared by adding 107.9 g to 800 mL of water with stirring and adjustment to pH 7.2 with NaOH. The volume is brought to one liter with water and the solution stored at room temperature. Because this solution is used only at the end of the experiments, it does not require sterilization.

0254General Cell Culture. In TABLE 1 the cell lines used in the described Examples are listed. The abbreviation "KCC" is the Karmanos Cancer Center, Cell Line Repository, Detroit, MI. The abbreviation "ATCC" is the American Type Culture Collection, Cell Line Repository,

0255Manassas, VA. Professor Armen Tashjian's address is Harvard University, Boston, MA. Dr.

0256William Rosner's address is Columbia University, New York. Dr. Sirbasku's address is The

0257University of Texas, Houston, Texas. The superscript designations in TABLE 1 for each of the cell lines indicate references that verify that the estrogen and androgen responsive cell lines used in this study are bonafide hormone responsive based on their tumor forming characteristics in host animals.

0258Those reports are clear demonstrations of the reliability of the models used in the present investigations to study sex hormone dependence in culture.

0259TABLE 1 Cell Lines Employed in the Examples. ER⁺ indicates receptor containing/E₂ sensitive

0260<img file="EP1328293A2_D0001.tif" />

0261The use of two strains of MCF-7 cells has been described (Sirbasku DA and Moreno-Cuevas (2000) In Vitro Cell Dev Biol 36, 428-446). Clonal variations of this line are known (Seibert K et al. (1983) Cancer Res 43, 2223-2239). Demonstration of estrogen responsive MCF-7 tumor formation in vivo (Huseby RA et al. (1984) Cancer Res 44, 2654-2659; Soule HD and McGrath CM (1980) Cancer Lett 10, 177-189; Welsch CW et al. (1981) Cancer Lett 14, 309-316).

0262² Estrogen responsive T47D tumors in vivo (Leung CKH and Shiu RPC (1981) Cancer Res 41, 546-551).

0263³ Estrogen responsive ZR-75-1 tumors in vivo (Osbome CK et al. (1985) Cancer Res 45, 584-589).

0264⁴ Estrogen responsive GH₄C₁ tumors in vivo (Riss TL and Sirbasku DA (1989) In Vitro Cell Dev Biol 25, 136- 142).

0265⁵ Estrogen responsive GH₃ tumors in vivo (Sorrentino JM etal. (1916) JNatl Cancer Inst 56, 1149-1154). Estrogen responsive MTW9/PL2 tumors in vivo (Sirbasku DA (1978) Cancer Res 38, 1154-1165; Danielpour

0266D and Sirbasku DA (1984) In Vitro 20, 975-980). η

0267Estrogen responsive H301 tumors in vivo (Sirbasku DA and Kirkland WL (1976) Endocrinology 98, 1260- 1272; Liehr JG et al. (1986) J Steroid Biochem 24, 353-356).

0268⁸ Androgen responsive LNCaP tumors in vivo (Sato N et al. (1997) Cancer Res 57, 1584-1589; Gleave M et al (1991) Cancer Res 51, 3753-3761; Horoszewicz JS et al. (1983) Cancer Res 43, 1809-1818; Pretlow TG et al. (1991) Cancer Res 51,3814-3817; Passaniti A et al. (1992) Int J Cancer 51, 318-324). General Cell Culture - Cell Passage Method. All stock cultures were grown in medium containing phenol red. Stocks of the cells were maintained at 37°C in a humid atmosphere of 5% (v/v) C0₂ and 95% (v/v) air in 17 to 20 mL of standard D-MEM/F-12 with 2.2 g per liter sodium bicarbonate, 15 mM HEPES (pH 7.4), and serum. With all cell lines except the rat pituitary cells, the serum used for stock culture was 10% (v/v) fetal bovine serum (FBS). For the three rat pituitary tumor cell lines GHA, GHi and GH₃, the medium contained 12.5% (v/v) horse serum and 2.5% (v/v) FBS. To passage the cells, the medium was removed and the dishes washed with 10 mL of saline. Next, the cells were dissociated by incubation at room temperature or at 37°C for 3 to 10 minutes with 1.5 mL of trypsin/EDTA. The action of the trypsin was stopped by addition of 8 mL of D-MEM/F-12 containing 10% (v/v) FBS or 8 mL of the horse serum/FBS combination. The cells were collected by centrifugation at lOOOx g for 5 minutes and suspended in 10 mL of fresh serum containing medium. Aliquots were diluted into Isoton II (Coulter Diagnostics) and cell numbers determined with a Model ZBI or Zl Coulter Particle Counter. The new dishes (100-mm diameter with 15 to 20 mL of fresh medium) were seeded with 2.0 x 10⁵ to 1.0 x 10⁶ cells on an alternating three-four day schedule or weekly as dictated by cell line growth rate. Cultures were used for growth assays between three and six days after passage. Acidic (yellow medium indicator color) cultures are not used for growth assays.

0269General Cell Culture - Media Types Used. The assays done in the presence of serum were initially in "low-Fe" D-MEM/F-12 containing phenol red (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427). The issue of the significance of the presence or absence of phenol red, a potential estrogen (Berthois Y et al. (1986) Proc Natl Acad Sci USA 83, 2496-2500), has been dealt with in considerable detail (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 447-464). The Fe (JS) content of this medium was < 0.2 μM (Eby JE et al. (1992) Anal Biochem 203, 317-325). Fe (JS) levels of > 1.0 μM interfere with thyroid hormone and estrogen responsive rat pituitary tumor cell growth in culture (Eby JE et al. (1992) Anal Biochem 203, 317- 325; Eby JE et al. (1993) J Cell Physiol 156, 588-600; Sato H et al. (1991) In Vitro Cell Dev Biol 21 A, 599-602; Sato H et al (1992) Mol Cell Endocrinol 83, 239-251). Although Fe (JS) might prevent estrogen responsiveness from being identified in culture with MTW9/PL2 cells, as shown herein and reported (Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446; Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 447-464), this is not the case when serum is present. Standard Fe (JS)lFe (U) containing D-MEM/F-12 was as effective as the low-Fe medium. It is clear that the apotransferrin in the serum effectively reduced the free Fe (JS) in the medium to less than cytotoxic levels. As stated above, apotransferrin binds Fe (JS) with very high affinity at pH 7.4 in plasma. The total concentration of fransferrin in serum is about 3 mg/mL. Usually, two-thirds of the total is apotransferrin. This amount is more than adequate to chelate Fe (JS) in culture medium (Eby JE et al. (1992) Anal Biochem 203, 317-325). However, in assays in serum-free defined medium, as described below, a Fe (IS) chelator (e.g. apotransferrin or DFX) is present in the serum-free defined medium at sufficient levels to neutralize the toxic iron.

0270General Cell Culture - Growth Assay Methods. Cell growth assays were initiated with stock cultures that were harvested by trypsin/EDTA treatment as described above with one exception. It was highly preferred to stop the action of trypsin with 3 mL of soybean trypsin inhibitor (0.5% w/v in saline) instead of medium containing serum. The use of trypsin inhibitor reduced the possibility of contamination of the subsequent assay media by serum-derived steroid hormones. The dissociated cells were collected by centrifugation as described above and washed three times with 10 mL volumes of serum-free standard D-MEM/F-12. After each wash, care was taken to aspirate all medium from the cell pellet and the walls of the centrifuge tubes. This minimized the carryover of steroid hormones into the experimental test dishes. By taking steps to avoid carryover of serum-containing medium, steroid hormones are prevented from being retained by the cells in culture. It is highly preferred to wash the cells in this way before assaying to measure various steroid hormone effects in culture. It has been reported that steroid hormones are retained long term by breast cancer cells in culture (Strobl JS and Lippman ME (1979) Cancer Res 39, 3319- 3327). The above-described wash procedure negates this problem. After the final wash, the cells were suspended in 10 mL of serum-free D-MEMF-12 and cell numbers determined. When cells were to be assayed in medium without phenol red discussed elsewhere herein and reported (Moreno- Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 447-464), the cells were washed and resuspended in phenol red free D-MEM/F-12 purchased from Gibco-BRL. The growth assays were initiated in 35-mm dishes containing a total of 2.0 mL of medium and the final concentration of all components except steroid hormones. The steroid hormone stocks were diluted to appropriate concentrations in serum-free D-MEM/F-12 and 20 μL aliquots added to each dish. For all growth assays, the medium was not changed after the initial inoculation. Because several of the cell lines described in TABLE 2 grow in serum containing medium and serum-free defined medium as mixtures of suspension and attached cells, removal or changing of the medium during the course of the assays causes substantial cell losses. For all cell growth assays, the initial seed densities ranged from 5,000 to 12,000 cells per 35-mm diameter dish.

0271General Cell Culture - Steroid Hormone Preparations. A number of hormone preparations are used to supplement the cell cultures. Unlabeled steroid hormones were obtained from Sigma or Steraloids. Stock solutions were prepared in sterile glass containers. The powder (non-sterile) steroid is added to the bottle along with 200 ml of 70% aqueous ethanol (ready as sterile). The steroids dissolve within an hour at room temperature, or when required were dissolved by gentle heating on a hot plate (hand temperature test — no boiling - no open flames). The stock solutions were stored at 4°C and renewed at six-month intervals. It is not necessary or desirable to filter sterilize these solutions because of steroid hormone loss on filter membranes. Stocks of 1.0 mM steroid hormones were prepared. To prepare diluted stocks for direct use in culture, 10 μL of 1.0 mM steroid hormone is diluted into 10 mL of D-MEM/F-12. This gives a stock of 1.0 μM. It is used in the assay dishes or diluted further in D-MEM/F-12 as needed. The diluted steroids are discarded after each use because they bind to the plastic with storage. The formula weight (FW) of each of the common natural and synthetic hormones used is listed below in TABLE 2 along the abbreviation used for each and the amounts required to prepare 200 mL of stock.

0272TABLE 2 Preparation of Steroid Hormone Stocks for Cell Culture and Hormone Binding Assays

0273<img file="EP1328293A2_D0002.tif" />

0274General Cell Culture - Harvest and Counting Cells. At the termination of the experiments, each plate received 0.4 mL of crude pancreatic trypsin dissolved in phosphate buffered saline was added along with 0.3 mL of 0.29 M EDTA. After 4 to 40 minutes incubation at room temperature or at 37°C, the action of the trypsin was stopped by addition of 0.6 mL of horse serum. The cell clumps were dissociated further by one passage through a 20 4 or 23-gauge needle and syringe. This suspension was then diluted to 10 mL with Isoton π and cell numbers determined with a Coulter Counter. The results are presented as the average of triplicate dishes for each test medium. To determine day zero cell numbers, at least triplicate 1.0 mL aliquots of the inoculum were collected for counting during the seeding of the test dishes. Coulter Counter standardization and monitoring were performed by the manufacturer.

0275General Cell Culture - Quantification of Growth. The cell number results are converted to cell population doublings (CPD) by the following calculation:

0276Logio Average Cell Number on Collection Day_/

0277'Logio Average Cell Number on Day Zero CPD =

0278Logio 2

0279For the purposes of this Disclosure, the mitogenic response to sex steroid hormones is designated the "steroidogenic effect." For example, the "estrogenic effect" is calculated as the difference between CPD measured in the presence of an estrogen minus CPD in the absence of the steroid. These values equal cell number increases of 2^CPD. The term "androgenic effect" has the same meaning except that it describes growth caused by androgens such as DHT and T. CPD is used herein as a measure of growth because it is a direct calculation of the number of times a cell population undergoes cell division. Furthermore, CPD use permits a direct measure of ED₅₀ and EDioo Concentrations in different and in replicate assays. The significance of differences between test dishes and controls was evaluated by the student's t test. Values of p &lt; 0.05 were accepted as significant. Standard deviations (± SD) are included when appropriate.

0280ATCC Cell Line Deposition. The rat breast cancer cell line designated as MTW9 PL2 (ATCC accession number ) was deposited with the American Type Culture Collection,

0281Rockville, Md. on , 2000. The hamster kidney tumor cell line designated as H301

0282(ATCC accession number ) was deposited with the American Type Culture Collection,

0283Rockville, Md. on , 2000. The subject cultures have been deposited under conditions that assure that access to the cultures will be available during the pendency of this patent application to one determined by the Commissioner of Patents and Trademarks to be entitled thereto under 37 CFR 1.14 and 35 U.S.C. 122. The deposits will be available as required by foreign patent laws in countries wherein counterparts of the subject application, or its progeny, are filed. However, it should be understood that the availability of a deposit does not constitute a license to practice the subject invention in derogation of patent rights granted by governmental action.

0284Cell Lines - Budapest Treaty Compliance. Further, the subject culture deposits will be stored and made available to the public in accord with the provisions of the Budapest Treaty for the Deposit of Microorganisms, i.e., they will be stored with all the care necessary to keep them viable and uncontaminated for a period of at least five years after the most recent request for the furnishing of a sample of each deposit, and in any case, for a period of at least thirty (30) years after the date of deposit or for the enforceable life of any patent which may issue disclosing the cultures. The depositor acknowledges the duty to replace the deposits should the depositor be unable to furnish a sample when requested, due to the condition of the deposit. All restrictions on the availability to the public of the subject culture deposits will be irrevocably removed upon the granting of a patent disclosing it.

0285Growth of Cells for Steroid Hormone Receptor Assays. Whole cells were assayed for the presence of steroid hormone receptors by a modification of described methods (Baxter JD et al. (1975) Methods Enzymol 36, 240-248). The cells (9 to 10 x 10⁶) were seeded in 850 cm² roller bottles (Corning) containing 200 mL of standard phenol red containing D-MEM/F-12 supplemented with 2.0% (v/v) charcoal-dextran extracted (CDE) horse serum and grown at 37°C for five to seven days. The cells were collected from the plastic surface and the medium and concentrated by centrifugation at lOOOx g for 15 minutes. The cells were washed with saline, resuspended in 20 mL of saline and treated with 2.0 mL of trypsin/EDTA at room temperature for one minute. The trypsin action was stopped with 6.0 mL of 0.5% (w/v) soybean trypsin inhibitor in saline. The dispersed cells were collected and suspended in phenol red-free D-MEM/F-12 to a density of 0.5 to 1.0 x 10⁶ per mL. The rationale for measuring steroid hormone binding with whole cells rests with the intent to replicate cell culture conditions. To derive complete information, the use of whole cells avoids the possible loss of a new receptor that might not withstand the cell extraction process or otherwise not be recovered.

0286Whole Cell Steroid Binding Assays. Total estrogen binding was measured with ³H-E₂ (2,4,6,7-³H-17β-estradiol) at specific activity 96 Ci per mmole (Amersham). Non-specific binding was assessed with ³H-E₂ plus a 100-fold molar excess of unlabeled DES. Specific binding was total binding minus non-specific binding. To assay specific progesterone binding, the medium contained either ³H-progesterone [1,2-³H (n) progesterone] at specific activity 92 Ci per mmole (ICN) or ³H-progesterone plus a 100-fold molar excess of the unlabeled synthetic progestin R5020 (DuPont NEN). Specific androgen binding was measured using [1,2 ³H (N)] DHT at specific activity 45 Ci per mmole (DuPont NEN) and the combination of ³H-DHT plus a 100-fold excess of unlabeled DHT. Glucocorticoid specific binding was assayed with [1,2 ³H (N)] hydrocortisone at specific activity 53 Ci per mmole (DuPont NEN) and ³H-hydrocortisone plus a 100-fold excess of unlabeled DEX. The steroid hormone binding incubations were done in phenol red free D- MEM/F-12 in a total volume of 1160 μL containing 1000 μL of cells, 100 μL of labeled steroid and 60 μL of unlabeled steroid hormone or medium. The incubations were done in glass tubes for two hours at 37°C with gentle agitation in an orbital shaker water bath followed by cooling to 0°C for 15 to 30 minutes. The cells were collected by centrifugation at 7°C as described above and washed three times with 2 mL portions of ice-cold phenol red-free D-MEM/F-12. The final collected cells were dissolved in 0.5 mL of 0.5N sodium hydroxide and the radioactivity quantified by liquid scintillation counting. All samples were duplicates or triplicates. To obtain dissociation (K_d) and association (K_a) constants, the data were analyzed by the method of Scatchard (Scatchard G (1949) Ann NY Acad Sci 51, 660-672).

0287Steroid Hormone Receptor Analysis by Western Immunoblotting. The following antibodies were obtained from Affinity Bioreagents: a rabbit polyclonal antibody against a bacterial fusion protein containing the N-terminal domain of the human androgen receptor and a mouse monoclonal antibody against the amino acid sequence 533 through 547 of the DNA binding domain of the progesterone receptor. An affinity-purified rabbit polyclonal antibody corresponding to the amino acid sequence 580 through 599 of the mouse estrogen receptor was obtained from Santa Cruz Biotechnology. To analyze steroid hormone receptor content, both cytosolic and nuclear extracts were prepared. To obtain the cytosol, 20 x 10⁶ cells were washed with serum-free D-MEM/F-12 and resuspended in 5 mL of 0.01 M Tris-HCl, pH 7.4, containing 0.15 M NaCl and 1 mM EDTA (Tris/EDTA). After cooling to ice-bath temperature, the cells were disrupted by three treatments for ten seconds with a Tekmar Polytron homogenizer. The homogenates were centrifuged at 800x g for 10 minutes followed by centrifugation at 150,000x g for one hour to obtain the cytosolic supernatants. To prepare nuclear extracts, the pellets from the 800x g centrifugation were homogenized again three times with 1.5 mL Tris/EDTA as described above. The centrifugation supernatants from the three homogenizations were combined to give the nuclear extract from each cell line. Protein concentrations were determined using the BCA kit from Pierce Chemical ® kit with bovine serum albumin as standard. When required, the samples were concentrated by precipitation with 20% (w/v) trichloroacetic acid. The precipitates were washed once with 500 μL of 70% (v/v) ethanol and twice with 500 μL of water. They were dissolved in 200 μL of 0.01 M Tris-HCl, pH 7.4, containing 1% (w/v) sodium dodecyl sulfate (SDS) by warming to 65°C. SDS-PAGE (Laemmli UK (1970) Nature (Lond) 227, 680-685) was done using 8 to 15% (w/v) acrylamide gradient gels with 3% (w/v) acrylamide stacking gels. Each sample was diluted with four volumes of buffer containing 0.3125 M Tris-HCl, pH 6.8, 10% (w/v) SDS, 50% (v/v) glycerol, 25% (v/v) mercaptoethanol and 0.0025% (w/v) bromophenol blue. After heating to 95°C for five minutes, the samples were applied to the gels and electrophoresis carried out at 7°C. The separated proteins were transferred to nitrocellulose membranes using a Milliblot Graphite Electroblotter I with a transfer buffer containing 1.0 mM (3-[cyclohexylamino]-l- propanesulfonic acid), pH 11, with 10% (v/v) methanol. Transfer was done for 45 minutes at 390 milliamps at room temperature. The receptors were detected by chemiluminescence using a kit from Tropix ®. The protocol used was that recommended by the manufacturer. For the detection of androgen receptors, the membranes were incubated at room temperature with a 1 : 100 dilution of the primary antiserum for one hour and a 1:10000 dilution of second antibody for one hour. To detect estrogen receptors, the primary antiserum was used at a 1:5000 dilution with incubation at room temperature for one hour followed by one hour with a 1:10000 dilution of second antibody. For estrogen and androgen receptors, the second antibody was an affinity purified anti-rabbit immunoglobulin conjugated to alkaline phosphatase. To detect progesterone receptors, the incubations were done with 5μg/mL primary antibody for 24 hours at 7°C, followed by incubation with a 1:1000 dilution of second antibody for eight hours at room temperature. The second antibody was an affinity purified anti-mouse immunoglobulin conjugated to alkaline phosphatase. Western Immunoblotting with other Primary Antibodies. The SDS-PAGE and Western Immunoblotting method described above was used throughout the Examples with the only significant modifications being changes in primary antibodies, and if required, changes in the secondary antibody. The changes are noted when introduced. Labeled Steroid Hormone Binding to Whole Serum and the Purified Inhibitor

0288Including Scatchard Analysis. The binding affinities of tritium labeled steroid hormones (purchased from DuPont NEN) to serum and the purified inhibitor were analyzed by the ion exchange filter method (Mickelson KE and Petra PH (1974) FEBS Lett 44, 34-38). Total DHT binding was measured with [1,2-³H(N)] DHT at specific activity 45 Ci/mmole. Nonspecific binding was assessed with ³H-DHT plus a 100-fold molar excess of unlabeled DHT. Specific binding was total binding minus nonspecific binding. For E₂ specific binding, the incubations contained either ³H-E₂ [2,4,6,7-³H-17β-estradiol] at specific activity 96 Ci/mmole or ³H-E₂ plus a 100-fold molar excess of DES. Glucocorticoid specific binding was assayed with [1,2-³H (N)] hydrocortisone at specific activity 53 Ci/mmole and ³H-hydrocortisone plus a 100-fold molar excess of unlabeled DEX. Progesterone specific binding was assayed with ³H-progesterone [1,2- ³H (n) progesterone] alone at specific activity 92 Ci mmole or ³H-progesterone plus a 100-fold molar excess of unlabeled synthetic progestin R5020. The use of 100-fold unlabeled steroids to determine nonspecific binding has been discussed (Chamness GC and McGuire WL (1975) Steroids 26, 538-542). Each assay contained 0.01 M Tris-HCl, pH 7.4, with 10 mM CaCl₂. The binding conditions were optimized for time and temperature. The incubations were done in glass tubes at 34 C in a total volume of 660 μL that included 50 μg/mL of the phenyl Sepharose pools and labeled and unlabeled steroid competitor. After two hours, the incubations were cooled to ice bath temperature and 50 to 200 μL aliquots applied to each DEAE-cellulose (DE-81) ion exchange filter (2.3-cm, Fisher) positioned in a Millipore Vacuum Filter Manifold ®. Thereafter, the filters were washed with ten one mL portions of ice-cold Tris/CaCl₂ buffer. With control incubations minus protein, &lt; 3% of the label was retained. Radioactivity was quantified by liquid scintillation methods (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427). To obtain K (dissociation constant) and K_a (association constant) values by Scatchard analysis (Scatchard G (1949) Ann NY Acad Sci 51, 660-672), the incubations began with 50 nM ³H-labeled steroid hormone and included five or six serial two-fold dilutions done both minus and plus 100- fold excesses of unlabeled steroid for each labeled hormone concentration. Data points were the averages of triplicate incubations. Ka/K_a were obtained from plots of (bound/free) versus (bound) hormone. Best-fit slopes were estimated with either Apple MAC computer software or with the PC based Graph Pad program.

0289Protein Assay and Quantification Methods. For cell growth assays, the protein sample volumes added to the culture medium were &lt; 20%. As required, chromatography samples were concentrated using Amicon Ultrafiltration with YM-10 (molecular weight cut-off 10,000) low protein binding membranes and nitrogen gas pressure (Sirbasku DA et al. (1991) Biochemistry 30, 295-304). Before assay, all fractions were dialyzed against 0.05 M Tris-HCl, pH 7.4, with 0.15 M NaCl using Spectropor 1 membranes (molecular weight cutoff 6,000 to 8,000). They were sterilized with 0.2 μ membrane filtration units. The protein concentrations of serum, the ammonium sulfate precipitation and all of the conventional chromatography fractions were estimated as one A_280BM equal to one mg/mL. For the cortisol affinity isolated proteins, concentrations were estimated either by the Pierce BCA® method according to the instructions supplied or the dye binding method of Bradford (Bradford MM (1976) Anal Biochem 11, 248- 254). Trichloroacetic acid (TCA) precipitation was used with BCA to eliminate the interfering cortisol and DHT. With TCA, deoxycholate was used to co-precipitate the protein and ethanol/water washes to eliminate the steroid. Human IgG was used as standard for the colorimetric protein determinations. The protein concentration of several lots of horse serum averaged of 30 ± 5 mg/ml. Ammonium sulfate precipitation was carried out as described (Sirbasku DA et al. (1991) Biochemistry 30, 295-304). Before further use, the protein was dissolved in 0.05 M Tris-HCl, pH 7.5, containing either 10 mM CaCl₂ or 0.15 M NaCl and dialyzed with Spectropor membranes against several four-liter volumes of the same buffer or a buffer appropriate to the next chromatography step.

0290HI. Examples

0291Example 1. Identification of Steroid Hormone Receptors in MTW9/PL2 Cells. L the course of searching for what regulates the growth of estrogen responsive breast cancer and of androgen responsive prostate cancer, an in vitro cell culture system was developed that would serve as an accurate model for predicting in vivo physiological effects. An estrogen responsive rat mammary tumor cell line, the MTW9/PL2 cell line had already been developed (Sirbasku DA (1978) Cancer Res 38, 1154-1165). The MTW9/PL2 population is the first highly steroid hormone-responsive rat mammary tumor cell line to be established in culture from a carcinogen-induced tumor. These cells have been shown previously to form estrogen responsive tumors in W/Fu rats (Sirbasku DA (1978) Cancer Res 38, 1154-1165; Danielpour D and Sirbasku DA (1984) In Vitro 20, 975-980; Riss TL et al. (1986) J Tissue Culture Methods 10, 133-150). Nonetheless, they were not estrogen responsive in culture (Sirbasku DA (1978) Proc Natl Acad Sci USA 75, 3786-3790). It was thought possible that the cells had lost the estrogen receptors (e.g. dedifferentiation). This Example presents evidence confirming that the cells are estrogen receptor positive and are suitable for use in in vitro and in vivo studies.

0292Identification of MTW9/PL2 Cell Estrogen Receptors by ³H-E₂ Binding Methods. Examining the MTW9/PL2 cell line anew, the MTW9/PL2 population was examined for ³H-E₂ binding to whole cells, to determine if estrogen receptors were present. First, the effect of temperature on ³H-E₂ specific binding was examined (Fig. 1). At 37°C, specific binding reached a maximum in two hours and thereafter decreased rapidly. At 23°C, specific binding reached the same maximum but at three hours. The decay in binding at 23°C was not as pronounced as at 37°C. At 7°C, the rate of specific binding reached a stable maximum at six hours. Similar temperature effects have been observed for the kinetics of ³H-E binding to MCF-7 breast cancer cells (Horwitz KB and McGuire WL (1978) JBiol Chem 253, 8185-8191; MacLidoe et al. (1982) Steroids 39, 245-258).

0293Effect of ³H-E₂Concentration on Binding. Next, the effect of the concentration of ³H- E₂ on binding at 37°C was characterized. Specific binding was saturated by > 5 nM ³H-E₂ (insert Fig. 2). One-half saturation occurred at 2 to 3 nM ³H-E₂. A Scatchard analysis (Scatchard G (1949) Ann NY Acad Sci 51, 660-672) of ³H-E₂ binding also was done at 37°C (Fig. 2) (N=2). It indicated a single class of E₂ binding sites with a dissociation constant (K_d) of 2.78 x 10^"9 M. This analysis indicated 38,400 estrogen receptors per cell. These values compared closely to a K_d of 1.89 x 10^"9 M and the estimated 34,000 sites per cell determined for ³H-E₂ binding to the original MTW9/PL cell population in 1982 (Leland FE et al. (1982) Li: Cold Spring Harbor Conferences on Cell Proliferation, Volume 9, Growth of Cells in Hormonally Defined Media, Sato G, Pardee AB and Sirbasku DA, eds, Cold Spring Harbor, New York, pp 741-750). Plainly, the estrogen receptor content of this permanent cell population has remained stable over several years.

0294Specificity of ³H-E₂ Binding. The specificity of ³H-E₂ binding to MTW9/PL2 cell receptors was examined. The effects of unlabeled DES, DHT or testosterone (T) on binding of 5 nM ³H-E₂ at 37°C were examined. The results from one of these experiments (N = 3) are shown in Fig. 3A. At 100-fold molar excess, unlabeled DES reduced ³H-E₂ total binding by 85%. Conversely, 100-fold molar excesses of either DHT or T did not displace ³H-E₂ total binding significantly. Even at 1000-fold excess, T or DHT only reduced ³H-E₂ total binding by 15%. Next, the effects of unlabeled progesterone and cortisol on ³H-E₂ binding to MTW9/PL2 cells were investigated under conditions similar to those used in Fig. 3B. A 100-fold excess of either progesterone or cortisol reduced ³H-E₂ binding by 30 to 50%. The results of the ³H-E₂ binding competition studies presented here are nearly identical to those done with cell extracts of the original MTW9/PL population in 1982. Comparison of the Labeled E₂ Binding Dissociation Constants (K_d) of Several

0295Estrogen Sensitive Cell Types. Clearly, the assays with extracts measured the same affinity binding sites as analyses with whole cells. This offers reasonable evidence that the standard binding technology employed in these studies is measuring the most common form of receptor present in cells, no matter whether whole cells are assayed or cell extracts. The affinity of the MTW9/PL2 estrogen receptor is that which is characteristic of the ERα. The K_d of the receptor measures the concentration of ligand that one-half saturates the sites. Li TABLE 3, the K_d values for labeled E₂ are presented as reported and presumably represent the ERα. Only when the measurements are specific for the β form is the designation (ERβ) included. TABLE 3 Comparison of E₂ Binding Affinities Expressed as Dissociation Constants (K )

0296<img file="EP1328293A2_D0003.tif" />

0297TABLE 3 presents only a fraction of the estrogen binding data available in the literature. However, the K_d values presented are representative and do show a discemable pattern. The lowest K_d from a literature search was in the range 5 x 10^"11 M to 1.0 x 10^"10 M for the ERβ and 7 x 10^"π M to 1.1 x 10^"10 M for the ERα. Li general, the binding affinities as estimated by K_d are lower for receptors from human cells than those from rodent lines. It is important to note that the results presented in TABLE 3 indicate that the lower limit of measuring estrogen receptor affinities most likely has been reached. The use of the highest specific activity tritium labeled steroids has been optimized and simply cannot be used to measure 10 to 100-fold lower K_d concentrations. This opens the possibility of an as yet unrecognized ER that mediates growth effects at lower concentrations of estrogen than either the ERα or the ERβ. Effect of Temperature on ³H-Progesterone Binding. Progesterone receptors in

0298MTW9/PL2 cells were sought using the same series of experiments done to identify estrogen receptors. The effect of temperature on progesterone specific binding with MTW9/PL2 cells is shown in Fig. 4. Maximum ³H-progesterone binding at 37°, 23°C and 7°C occurred at 2, 4 and 6 hours, respectively. After reaching an optimum, the binding decayed at 37°C and 23°C but not at 7°C.

0299Effect of ³H-Progesterone Concentration on Binding The saturability of ³H- progesterone binding was examined at 37°C. Labeled progesterone specific binding was saturated at > 5 nM (insert Fig. 5). One-half saturation occurred at 0.5 to 1 nM ³H-progesterone. Scatchard analysis (N = 2) identified a single class of binding sites with a K of 1.02 x 10^"9 M and yielded an estimated 26,800 sites per cell, as shown in Fig. 5. Previous studies in 1982 with extracts of the original MTW9/PL cell population had given a K_d of 3.29 x 10^"9 M for ³H-progesterone binding and an estimated 180,000 sites per cell (data not shown). Comparison of the number of progesterone sites then and now indicates a decrease. However, a sufficient number remain to expect progesterone specific gene expression or growth regulation (Alexander IE et al. (1989) Mol Endocrinol 3, 1377- 1386; Keydar I et al. (1979) Eur J Cancer 15, 659-670).

0300Effect of Other Steroid Hormones on ³H-Progesterone Binding. The effect of non- progestins on ³H-progesterone binding was investigated (Fig. 6). As control, the binding was studied in the presence of increasing concentrations of the synthetic progestin R5020. A 100-fold excess of the unlabeled R5020 reduced ³H-progesterone binding by 82%. A 100-fold excess of unlabeled DHT or T reduced binding by &lt; 20%. A 100-fold excess of unlabeled F^ reduced progesterone binding by a maximum of 20% (data not shown).

0301Assays for Androgen and Cortisol Receptors. Experiments (N = 3) were carried out to seek specific saturable binding sites for androgens using ³H-DHT. Li experiments not shown, incubation of MTW9/PL2 cells at 37°C for two hours with &lt; 20 nM labeled DHT did not reveal saturable specific binding sites. Studies using ≤ 20 nM ³H-cortisol did not identify specific saturable receptors for this corticosteroid.

0302Estrogen effects on Progesterone Receptor Expression in MTW9/PL2 Cells. Estrogens induction of progesterone receptors in target cells is generally taken as strong evidence of sex steroid responsiveness by the criteria of regulation of gene expression (Leavitt WW et al. (1977) Ann NY Acad Sci 286, 210-255; Toft DO and O'Malley BW (1972) Endocrinology 90, 1041-1045; Horwitz KB and McGuire (1978) JBiol Chem 253, 2223-2228; Haslam SZ and Shyamala G (1979) Biochem J 182, 127-131; Haslam SZ and Shyamala G (1979) Endocrinology 105, 786-795). L the next study, it was asked whether this was the case with MTW9/PL2 cells. The cells were grown for five to seven days in the absence of estrogens in standard phenol red containing D-MEM/F-12 plus 2% (v/v) charcoal-dextran extracted (CDE) horse serum. Thereafter, they were harvested and inoculated into phenol red free medium in 100-mm diameter dishes containing 1.0 x 10^"8 M E^ Beginning at day 0 (inoculation day) and for each of the next five days, the cells were assayed for progesterone receptors as described (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427). All results were normalized to "receptors per cell" to correct for growth over the duration of the experiment. The number of progesterone receptors increased 1.8-fold within two days after exposure to E₂ (Fig. 7). Li a replicate experiment, the induction was 2.0-fold in two days. The effect of estrogen concentration on progesterone receptor induction also was evaluated (insert Fig. 7). Maximum induction occurred at 1.0 nM E^ These results corifirm that MTW9/PL2 cells are E₂ responsive by a criterion separate from growth.

0303Western Analysis for Androgen, Estrogen and Progesterone Receptors. Western irnmunoblotting was used to analyze the MTW9/PL2 cells for the presence of steroid hormone receptors. Nuclear and cytosolic extracts were compared to those from negative control rat and human diploid fibroblasts and positive control T47D and LNCaP cells. The T47D cells have androgen (Keydar I et al. (1979) Eur J Cancer 15, 659-670; Horwitz KB et al. (1978) Cancer Res 38, 2434-2437), progesterone (Horwitz KB et al. (1978) Cancer Res 38, 2434-2437; Horwitz KB and Alexander PS (1983) Endocrinology 113, 2195-2201; Lessey BA et al. (1983) Endocrinology 112, 1267-1274) and estrogen (Horwitz KB et al. (1978) Cancer Res 38, 2434-2437; Keydar I et al (1979) Eur J Cancer 15, 659-670; Soto AM et al. (1986) Cancer Res 46, 2271-2275) receptors. The androgen receptors of LNCaP cells previously have been characterized by labeled hormone binding analysis (Veldscholte J et al. (1990) Biochem Biophys Res Commun 173, 534-540; Veldscholte J et al. (1990) Biochim Biophys Ada 1052, 187-194) and Western immunoblotting (Prins GS et al. (1991) Endocrinology 129, 3187-3199). Although the LNCaP cells were initially reported to not have progesterone or estrogen receptors (Schuurmans AL et al. (1988) Int J Cancer 42, 917-922; Brolin J et al. (1992) The Prostate 20, 281-295), more recent evidence indicates that they express significant levels of both (Castagnetta L et al. (1995) Endocrinology 136, 2309-2319). Western Analysis - Androgen Receptors. As shown in (Fig. 8, top panel), nuclear and cytosolic extracts of LNCaP cells gave an intense band at 101 kDa that was the expected mass of the androgen receptor as determined previously by irrimunoblotting (Prins GS et al. (1991) Endocrinology 129, 3187-3199; Berrevoets CA et al. (1993) J Steroid Biochem Mol Biol 46, 731- 736). It was also nearly the same mass as the 98.5 kDa predicated by molecular cloning (Trapman J et al. (1988) Biochem Biophys Res Commun 153, 241-248; Faber PW et al. (1989) Mol Cell Endocrinol 61, 257-262). The bands migrating at 79 kDa and 21 kDa may be degradation products, non-specific reactions between the antibody and unrelated proteins, or may represent other forms of the androgen receptor (Prins GS et al. (1991) Endocrinology 129, 3187-3199). The T47D cells showed the same androgen receptors as LNCaP cells. In contrast, there was little androgen receptor in the nuclear or cytosolic extracts of MTW9/PL2 cells. The faint band identified was the same intensity as an equivalent component seen in the extracts of fibroblasts (Fig. 8, top panel). Fibroblasts have been reported to have low levels of androgen receptors (Eil C et al. (1983) Gin Endocrinol 19, 223-230; Eil C et al. (1980) Steroids 35, 389-404).

0304Western Analysis - Progesterone Receptors. Similar experiments were done to identify progesterone receptors with MTW9/PL2 cells. Receptors of 79 kDa (A form) and 107 kDa (B form) were immunostained with both the cytosolic and nuclear extracts (Fig. 8, middle panel). Another possible form was identified at 44 kDa. The progesterone receptors of MTW9/PL2 cells were the same molecular mass as the A and B forms from chick oviduct (Dure LS et al. (1980) Nature (Lond) 238, 784-786; Birnbaumer M et al. (1983) JBiol Chem 258, 1637-1644). The forms in MTW9/PL2 cells also compared closely to the 85.6 kDa and 109.6 kDa masses reported for progesterone receptors of rat uterus (Henchuk TT et al. (1987) Endocrinology 120, 1449-1456). Furthermore, the 79kDa component (A form) was the more abundant of the two receptors in MTW9/PL2 cells. This was also the case for rat uterus (Henchuk TT et al. (1987) Endocrinology 120, 1449-1456). The T47D positive controls showed the same molecular masses of progesterone receptors although there was greater immunostaining with the cytosolic extracts than with the nuclear preparations. This receptor distribution between cytosol and nucleus was similar to that found when T47D cells were examined by labeled hormone binding (Horwitz KB et al. (1978) Cancer Res 38, 2434-2437). The T47D progesterone receptor masses observed in the present study were similar to those reported by others studying human breast cancer cells (Horwitz KB and Alexander PS (1983) Endocrinology 113, 2195-2201; Lessey BA et al. (1983) Endocrinology 111, 1267-1274; Horwitz KB et al. (1985) Recent Prog Hormone Res 41, 249-316). Additionally, the LNCaP cells showed intense nuclear extract staining for the same forms of progesterone receptors seen in extracts of MTW9/PL2 and T47D cells. No progesterone receptors were identified in the negative control rat or human fibroblasts.

0305Western Analysis - Estrogen Receptors. Li the final study, the estrogen receptors in MTW9/PL2 were sought (Fig. 8, bottom panel). The major form of estrogen receptor in MTW9/PL2 cells was molecular mass 67 kDa. Two presumed degradation products of 50 kDa and 17 kDa were also observed. The results with MTW9/PL2 cells were in agreement with the expected mass of the rat estrogen receptor estimated at 67 kDa by molecular cloning (Koike S et al (1987) Nucleic Acid Res 15, 2499-2513). The extracts from T47D and LNCaP cells showed a similar estrogen receptor pattern. Control fibroblasts showed no estrogen receptors. Initially, the band identified at 50 kDa was thought to be a degradation product of the 67kDa intact form of ERα (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427); however, it may be an ERβ because that new receptor has a mass of 54 kDa (Enmark E et al. (1997) J Gin Endocrinol Metab 82, 4258-4265). The 50 kDa band may also represent a variant form of ERβ (Gustafsson J-A and Warner M (2000) J Steroid Biochem Mol Biol 74, 245-248). Alternately, this band may represent a new estrogen receptor that regulates growth, hereby designated as estrogen receptor gamma (ERγ).

0306Thus, it is concluded that another positive acting ER exists in the MCF-7 and T47D cells and its function is dominant and sustains growth related gene expression even with the inhibitory ERα present. The existence of two ER receptors is also indicated in an older study of the growth of the GH Cι rat pituitary tumor cells in culture (Amara JF and Dannies PS (1983) Endocrinology 112, 1141-1143). These investigators demonstrated a biphasic effect of E₂ on these cells. At picomolar concentrations, E₂ caused growth. At higher concentrations, E₂ induced prolactin production secretion and inhibited growth. If two receptors are operating, the growth receptor is more sensitive to E₂ whereas the ER regulating gene expression (e.g. prolactin mRNA production) is activated by higher concentrations of estrogen. This same biphasic action of estrogen on the growth of T47D human breast cancers cells has also been noted (Chalbos D et al. (1982) J Gin Endocrinol Metab 55, 276-283). Low concentrations promoted growth, whereas higher levels were inhibitory. Indeed, a biphasic effect also was noted with the MCF-7 cell line (Soto AM and Sonnenschein C (1985) J Steroid Biochem 23, 87-94). When this observation is coupled with the clear statements of Soto et al (Soto AM et al. (1986) Cancer Res 46, 2271-2275) that "the free estradiol levels needed for maximum response are significantly lower than estrophilin (i.e. ERα) K_ds", there is further support for the view that an ER exists that regulates growth and is more estrogen sensitive (i.e. lower K_d) than the classical ERα. While those investigators conclude that the results exclusively supported their estrocolyone hypothesis, and excluded ERα as the positive growth regulator, there was no recognition of the possibility of a much higher affinity receptor different than ERα. Finally, there is one other issue that has perplexed endocrinologists and cancer biologists for several years. Breast cancer is sometimes treatable with high doses of estrogen (Segaloff A (1981) Banbury Report 8, 229-239). If the ERα is the only growth mediator, one is forced into many other postulates to explain this observation (Reese CC et al. (1988) Ann NY Acad Sci 538, 112-121). Indeed, this may actually represent evidence that full occupation of ERα is inhibitory and that another receptor is the positive signal.

0307Variant estrogen receptors have been identified previously by others. For example from the estrogen growth responsive T47D human breast cancer cell line, there have been three isoforms of the ERα identified in one study (Wang Y and Miksicek RJ (1991) Mol Endocrinol 5, 1707-1715) and another three in a different study (Graham ML et al (1990) Cancer Res 50, 6208- 6217). With another two estrogen growth responsive human breast cancer cell lines, the MCF-7 and ZR-75-1, another ERα variant was identified that lacked the entire exon 4 of the receptor (Pfeffer U et al. (1993) Cancer Res 53, 741-743). Variant receptors have also been identified from human breast cancer biopsy specimens (Murphy LC and Dotzlaw H (1989) Mol Endocrinol 3, 687-693). Another truncated variant of ERα acts as a natural inhibitor of the action of the wild- type ERα (i.e. unchanged receptor) (Fuqua SA et al. (1992) Cancer Res 52, 483-486). Another type of variant has received wide attention because it has constitutive transcriptional activity without the steroid hormone ligand bound (Fuqua SA et al. (1991) Cancer Res 51, 105-109). Even normal human breast epithelial cells show several natural variants of ERα (Yang J et al. (2000) Endocrine 12, 243-247). When all of these results are considered as a group, it is clear that different forms of the ERα are possible in cells. It is reasonable to conclude that an alternate form of ERα, possibly formed by alternate splicing, or possibly arising from an as yet unrecognized gene, may regulate estrogen dependent/responsive tumor cell growth. Upon further investigation ERγ may prove to be such a variant.

0308Whatever mechanism is proposed for the action of the steroid hormone (i.e. on growth), it can be seen from the data presented herein, and subsequently reported elsewhere (Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446), it takes a significant period to reverse the effects of the inhibitor. This process cannot be simply due to a rapid effect on transcription caused by steroid hormones (e.g. via a known estrogen receptor). Cellular metabolic events, including the transformation of E₂ to an active steroid metabolite, may provide the growth regulating ligand for one of the "orphan" estrogen receptors. An alternative possibility is that the receptor may be activated by metabolites formed from cholesterol metabolism (Gustafsson J-A (1999) Sczewee (Wash DC) 284, 1285-1286).

0309Discussion of Example 1. The evidence presented verifies that the MTW9/PL2 cells are estrogen receptor positive (ER¹) with a significant number of progesterone receptors and possibly low levels of androgen receptors. The estrogen receptor content and affinity characteristics of the MTW9/PL2 cells indicate appropriate stability for use as a testing standard and for commercial applications. The long-term stability of this cell line in culture, without alteration of its cell properties, is further discussed in Example 3. These results and the information provided above, show that this cell line is a unique asset for combination in vitro and in vivo modalities that can be applied to evaluate a multitude of compounds or preparations having, or potentially having, hormone or antihormone activities.

0310Example 2. Preparations of Steroid Hormone Depleted Serum

0311Three Methods. Li this example, evaluations of three methods for preparing steroid hormone depleted serum are described. The primary purpose was to prepare serum that supported large magnitude sex steroid growth effects in culture and to identify the dose-response concentrations that cause the effects, as demonstrated in Examples that follow. This meant preparing serum with &lt; 5pg/mL estrogen (and other steroid hormones). This concentration corresponds approximately to the lower Jimit of detection of steroids by radio immunoassay. The methods tested included (A) a two-step charcoal/dextran extraction of serum at 34°C, (B) a one- step charcoal extraction at 56°C, and (C) a one step treatment with Amberlite™ XAD™-4 resin at 37°C. One advantageous result established by this Example is that sera prepared according to the preferred methods contain significant amounts of active immunoglobulin inhibitors, in contrast to previously known steroid depleted sera.

0312(A) Charcoal-dextran Extraction at 34° C.

03131. Preparation of the charcoal/dextran mixture. Activated charcoal, untreated powder (100 to 400 mesh), was obtained from Sigma (Catalog No. C5260). This preparation was done at room temperature. The powder (30 g) was suspended in 600 mL of water and stirred for 20-30 minutes at room temperature. The water used to wash and suspend the charcoal was a purified source made by reverse osmosis/ ion exchange treatment/ charcoal filtration/ 0.2 μm pore diameter filtration/ distillation into glass (only) containers. Next, 3.0 g of Dextran T70 (Pharmacia) was dissolved in 300 mL of water, added to the charcoal suspension with stirring, and the mixture stirred for 30-60 min at room temperature, preferably 60 min. The mixture was then washed with about 6-8 liters of distilled water in a sintered glass funnel (2000 mL, ASTM 40-60, C#36060). This wash removes impurities as well as fine particles of charcoal that cannot be separated from serum by centrifugation. The charcoal-dextran retentate was suspended in a final volume of 300 mL of distilled water to yield a suspension of 100 mg/mL charcoal and 10 mg/mL dextran. Preferably the mixture is stirred vigorously for about an hour, and then stored at room temperature for no more than about 2-3 weeks prior to use.

03142. Charcoal-dextran extraction at 34°C of horse serum (CDE-horse serum). This serum in 500 mL sterile bottles was removed from the freezer (-17°C) and thawed at 4°C for 24 to 48 hours. Alternatively, fresh serum could be used. The thawed serum (still in the 500 mL sterile bottles) was placed in an orbital shaker water bath (Lab-Line Orbit Shaker Water Bath) equilibrated at 34°C. The serum was incubated at 140 RPM for 45-60 minutes to reach 34°C. Approximately 250 mL portions of the 34°C serum (total volume about 1 liter) were transferred to one-liter Erlenmeyer flasks and tightly capped with alvmiinum foil. These were incubated for 20-30 minutes (preferably 30 minutes) in the 34°C orbital shaker water bath at a medium-high rotation speed. Thereafter, 25 mL of the charcoal/dextran suspension was added to each flask. The charcoal-dextran suspension was stirred at room temperature while removing the 25 mL aliquot. The final charcoal concentration in each flask was 10 mg/mL, and the final concentration of dextran was 1 mg/mL. After addition of the charcoal-dextran mixture to each flask, the extraction mixtures were shaken at 140-160 RPM at 34°C for two hours. After this, the mixture was cooled on ice and the charcoal removed by centrifugation at 10,000x g for about 60 minutes at room temperature. Li some preparations the temperature of the mixture gradually warmed to about 40°C during centrifugation. The supernatants were pooled in a two-liter beaker and 275 mL portions of the supernatant (serum) transferred to fresh one-liter Erlenmeyer flasks. These were then incubated in the orbital shaker water bath at 34°C for 20-30 minutes (preferably 30 min) to re-equilibrate the temperature. A second extraction was done by addition of a fresh aliquot (about 14 mL) of the charcoal-dextran suspension. This re-extraction mixture was incubated with shaking for another 2 hours at 34°C. The final charcoal concentration during this extraction was about 5 mg/mL. Afterward the bulk of the charcoal was removed by centrifugation, as before. Li some preparations the temperature of the mixture reached about 41°C, without harming the quality of the serum. The supernatants were collected into a two-liter beaker and filtered through 5 μm pore diameter filters to remove residual charcoal. If it was considered necessary for particular preparations (for example, due to charcoal darkening serum) the serum was also filtered with 0.45 μm Millipore filters. These filtrations were done with plastic reusable filter holders and light vacuum. The steroid hormone depleted serum was then sterilized using 0.2 μm pore diameter filters. After sterilization, aliquots of about 26 mL were dispensed into sterile glass (50 mL) bottles or sterile 50 mL polypropylene tubes and stored frozen at -17°C. Although 34°C is preferred in the above-described regime, and provides the best results, satisfactory depletion of steroid hormones can be obtained over the temperature range of about 30 to 37°C. The 2 hour incubation times for the extraction and re-extraction mixture (at 34°C) are preferred, but a time range of 30 minute to 3 hours could also be used with success, employing longer incubation times for the lower temperatures within the 30-37°C range. A ± 25% variation in the charcoal concentration used for each extraction had no detrimental effects on the final product.

0315(B) Charcoal-dextran Extraction at 56°C. The preparation of 56°C charcoal-dextran extracted serum was done as described (Sirbasku DA and Kirkland WL (1976) Endocrinology 98, 1260-1272; Riss TL and Sirbasku DA (1989) In Vitro Cell Dev Biol 25, 136-142; Kirkland WL et al. (1976) J Natl Cancer Inst 56, 1159-1164). Frozen serum was thawed for four hours at 37°C followed by incubation in an orbital shaker water bath at medium-high speed at 56°C for 15 minutes. The same charcoal-dextran mixture described above was used in this extraction. One-tenth volume (warmed to 56°C) was added to the serum. Incubation was continued with shaking at 56°C for 15 minutes. Thereafter, the flasks were cooled in an ice bath and the charcoal removed by centrifugation at 8,000x g and filtration with 5.0 or 0.45μm pore membrane filters. The serum was 0.2 μ m pore filter sterilized and stored at -17°C.

0316(C) Amberlite™ XAD™-4 Resin Treatment. Li a different procedure carried out to free CBG of storage cortisol, XAD resin has been used to remove the steroid by incubation for 5 hrs at room temperature (A.M. Nakhla, et al. (1988) Biochem. Biophys. Res. Commun. 153, 1012-1018). Described as such, this method removed only about 80% of cortisol from the purified protein.

0317Careful application of that method failed to yield serum suitable for the purpose of this study. As an alternative to preparing steroid hormone depleted serum by charcoal-dextran extraction, horse serum was treated by incubation with Amberlite™ XAD™-4 nonionic polymeric absorbent (Aldrich, Catalog. No. 21,648-8; or Sigma Catalog No. XAD-4 37380-42-0). Specifically, a 500 mL bottle of horse serum was thawed at 37°C and divided into 250 mL portions that were each in a one-liter Ehlenmeyer flask. To each flask was added 25 grams of moist XAD-4 resin. The mixtures of serum and resin were then incubated with shaking in a rotary Labline Orbital Shaker water bath at 34°C at about two-thirds of the maximum rate for 24 hours (speed adjusted to control foaming). This extraction can be done at temperatures from 30°C to 37°C. At 30°C, the extraction requires 24 to 36 hours. At 37°C, it requires 18-24 hours to be complete. The 34°C and 37°C procedures are preferred. Each flask was tightly capped with aluminum foil and taped. After 24 hrs, the resin is allowed to settle by gravity, the supernatant decanted, and then vacuum filtered using a glass fiber filter in a Buchner funnel. The resulting serum was filter sterilized using 0.2 μm pore filter units. Aliquots of about 26 mL were frozen at -17°C in 50 mL sterile bottles or tubes.

0318The charcoal method described above is readily applicable to one to five liter volumes of serum per preparation. With use of < 50 mL per test substance, this is an adequate supply. To prepare larger volumes of serum (i.e. > 20 liters) for extensive testing programs or commercial applications, the charcoal-dextran methods will preferably employ industrial filtration or other separation equipment to remove the carbon after each extraction. The XAD-4 resin method as presented is adaptable to one to five liters for testing purposes. For industrial applications, where > 20 to 100 liter batches are customarily required, the resin method is preferred because of the need for only one separation after extraction. However, where "foaming" of the serum protein is to be avoided completely, charcoal extraction is superior. The materials cost for charcoal-dextran has an advantage when economy is a major consideration. It is less expensive than XAD-4 resin on a per liter basis, although the resin is commercially available at low cost when purchased in large amounts (i.e. > 50- 100 kilograms).

0319Discussion of Example 2. Each of the methods presented have advantages, depending on the particular needs and desires of the user. The scale procedures described are useful to prepared sufficient serum for testing of hormone activities or antihormone activities or evaluation of toxicity of compounds in cell culture assays. This 34°C method has been used to prepare CDE human serum, porcine serum, rat serum, hamster serum, ovine serum, fetal bovine serum, new born bovine serum (0 to 10 days old), young donor bovine serum (10 days to 6 moths old) young adult bovine serum (300 to 900 lbs), fetal horse serum, chicken serum, turkey serum, dog serum, goat serum, rabbit serum and monkey serum. Subsequent Examples demonstrate how these stripped sera are preferably employed. The results demonstrate the broad utility of the method of preparing charcoal-dextran extracted serum for testing of cell lines from many species using homologous serum assays. From these results it can also be readily appreciated that these methods are applicable to testing of veterinary medicine samples or compounds of significance to domestic animals, as well as any application where a steroid hormone stripped serum is used.

0320Example 3. Cancer Cell Line MTW9/PL2 Exhibits Estrogen Responsive Growth in 34°C Charcoal-dextran Extracted Serum

0321Estrogenic Effects in Cultures Supplemented with CDE-horse Serum. Unless otherwise stated, references in this and the following Examples to "CDE-horse serum" refer to the 34°C charcoal-dextran extraction process described in above. The MTW9/PL2 cells were assayed for E₂ responsiveness in cultures supplemented with increasing concentrations of CDE-horse serum (Fig. 9A). Concentrations < 5% (v/v) promoted growth. Typically within seven days cell numbers increased from 6,000 per dish to more than 200,000 in 2 to 5% serum. This most likely resulted from stimulation by serum-borne growth factors as well as the mitogenic effect of transferrin (Danielpour D et al (1988) In Vitro Cell Dev Biol 24, 42-52; Riss TL and Sirbasku DA (1987) In Vitro Cell Dev Biol 23, 841-849; Riss TL et al. (1986) J Tissue Culture Methods 10, 133-150). As serum concentrations exceeded 5% (v/v), the effects of the growth promoters were counteracted by a serum-borne inhibitor(s). At serum concentrations of 30 to 50% (v/v), growth was completely inhibited. Usually only seed density cell numbers were found after seven days in cultures containing 50% (v/v) CDE-horse serum. Li contrast, the presence of 1.0 x 10^"8 M E₂ completely reversed the serum dependent inhibition. Li cultures supplemented with 20 to 50% (v/v) CDE-serum plus 1.0 x 10^"8 M E₂, cell numbers were > 400,000 per dish. Logarithmic quantifying of cell growth was done by converting the cell number data in Fig. 9A into CPD. A plot of these values is shown in Fig. 9B. The estrogenic effect is also presented. Li Fig. 9B, the difference was maximum at 30% (v/v) CDE- horse serum. It was a 6.14 CPD or a 70-fold (i.e. 2 ^CPD or 2^{6 14}) increase in cell numbers in response to E₂. Li randomly selected replicate experiments (N = 9) done over a two-year period with different preparations of CDE-horse serum, the average maximum estrogen effect ± SEM was 6.43 CPD ± 0.49 (range 5.63 to 7.22). This was an 86-fold (2^6'43) estrogenic effect. The modal concentration of serum that promoted maximum E₂ effects was 40% (range 20 to 50%).

0322Morphology of MTW9/PL2 Cells Growing in CDE-horse Serum ± E₂. The morphology of the cells growing under the conditions was examined. The photomicrographs (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427) show cells growing under optimum conditions in medium with 2.5% (v/v) CDE-horse serum with and without E₂, respectively. The presence of the hormone had no effect on the appearance of the cultures. The cells grew in clusters in suspension and had the same morphology reported earlier for the parent MTW9/PL line grown in medium containing 10% (v/v) fetal bovine serum (Sirbasku DA (1978) Cancer Res 38, 1154-1165). When the concentration of CDE-horse serum was increased to 50% (v/v) without steroid, many fewer cells were present. Despite the near complete inhibition of growth, the morphology of the cells was the same as in rapidly growing cultures. Estrogen addition to this same medium caused substantial growth. The morphology of the estrogen-stimulated cultures in 50% serum was equivalent to that seen with or without estrogen in 2.5% serum. The inhibitor had no effect on the microscopic morphology of MTW9/PL2 cultures nor did it affect cell-cell adhesion or cell-surface adhesion.

0323Estrogen Reversibility of the Growth Inhibition Caused by CDE-horse Serum. It was examined whether inhibition caused by CDE-horse serum was reversible even after several days in culture (Fig. 10). The MTW9/PL2 cells were seeded into medium containing 50% (v/v) CDE-horse without E₂ and cell numbers monitored daily. Growth ceased within 48 hours; thereafter cell numbers remained static. In parallel cultures, addition of E^ on days two, four, and six after seeding caused resumption of growth (after a lag period) at nearly the same rate as cultures that received hormone at the time of inoculation. These results show that the cells survived in the presence of the inhibitor without E₂ for at least six days.

0324Storage Stability of CDE-horse Serum. TABLE 4 shows the effect of storage temperature on the estrogen mediating activity of CDE-horse serum. The assays were done with MTW9/PL2 cells as shown in Figs. 9A and 9B. Stability was assessed by four criteria: (i) the concentration of serum needed to give an estrogenic effect of 2.5 CPD (i.e. ED₂.₅), (ii) the percent serum needed for the maximum estrogenic effect, and the magnitude of the estrogenic effects (CPD) at (iii) 20% and (iv) 30% serum. CDE-horse serum was stable at 23°C for three weeks without loss of activity as assessed by all four criteria. Storage at 4°C was detrimental within 24 days as measured by the CPD at 20% and 30% (v/v) serum concentrations. Longer storage at 4°C was not advisable. Storage at -17°C was most effective; the activity was unchanged even after 90 days. In experiments not shown, repeated freeze-thaw cycles caused an appreciable loss of inhibitor activity.

TABLE 4

0326Summary of the Effects of Serum Storage Temperature on Activity.

0327<img file="EP1328293A2_D0004.tif" />

0328Dose-Response Effects of Steroid Hormones in CDE-horse Serum. The dose effects of a number of steroid hormones were evaluated with MTW9/PL2 cells in medium containing 50% (v/v) CDE-horse serum. The results of one of these studies (N=3) are presented in Fig. 11. Estrogens were the most effective mitogens. Their order of potency was E₂ > Ei > E₃. This relative potency was expected based on the affinities of these steroids for the estrogen receptors of other target tissues (Clark JH and Markaverich BM (1983) Pharmacol Ther 21, 429-453). The cell numbers in dishes containing 1.0 x 10^"13 M E₂ were 32-fold (p < 0.01) higher than in dishes without the hormone. Concentrations of 1.0 x 10^"!2 to 1.0 x 10^"11 M E₂ promoted 6.73 CPD that was a 110-fold estrogenic effect in seven days. The ED₅₀ of E₂ was about 0.5 to 1.0 x 10^"12 M. Using E_! and E₃, optimum growth was achieved at 1.0 x 10^"9 and 1.0 x 10^"8 M, respectively. In experiments not shown, the mitogenic potency of the synthetic estrogen DES was assessed. At 1.0 x 10^"8 M, it caused the same growth as saturating concentrations of the natural estrogens. The DES effect was 6.98 CPD in seven days that was a 126-fold (2^6,98) increase in cell number. The next most potent hormone was DHT. It caused significant (p < 0.05) growth at super physiologic concentrations ≥l.O x 10^"8 M. Progesterone also caused significant growth, but only at supraphysiological concentrations > 1.0 x 10^"7 M. Cortisol was ineffective at concentrations up to 1.0 x 10^"SM. Estrogen Mitogenic Effects with MTW9/PL2 cells in CDE-serum from Several Species.

0329Serum from species other than horse were examined to determine they also possessed estrogen reversible inhibitory activity with rat MTW9/PL2 cells. These experiments are shown in Fig. 12. All of the sera tested were charcoal dextran extracted at 34°C. CDE-porcine (Fig. 12A), and CDE- human serum (Fig. 12B) showed patterns nearly identical to that of CDE-horse serum. The maximum estrogenic effects with both sera were six to seven CPD (N = 3). CDE-rat serum also showed the same pattern of estrogen reversible growth inhibition (Fig. 12C). CDE-ovine serum showed estrogen reversible inhibition equivalent to CDE rat serum (data not shown). With serum from rats, the maximum estrogenic effect was four to five CPD (N = 4). CDE-bovine serum (adult donor herd) displayed the same pattern of activity as other sera (Fig. 12D). CDE-fetal bovine serum showed a different pattern (Fig. 12E). Even at 40% (v/v), there was no inhibition. With some batches of this serum, there was no inhibition even at 50% (N=2). With others (N=2), inhibition was found. Li these experiments, the estrogenic effects reached three to four CPD in 50% (v/v) CDE- serum. Even with this variability, fetal bovine serum has less activity than the serum from the adults of this species. The assays with CDE-fetal horse serum (N=3) showed inhibition at 50% (v/v) that was not reversible by 10 nM E₂ (Fig. 12F). The present study shows very clearly that estrogen growth effects were not found in medium with 5% (v/v) fetal bovine serum, as also reported (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427). In fact, charcoal-dextran treated fetal bovine serum at concentrations of &lt; 40% (v/v) does not cause inhibition of estrogen target cell growth in culture.

0330Technical Conditions for Demonstrating Estrogen Responsiveness in Culture and Evidence for a Serum-borne Inhibitor. Conditions that permit the observation of very large magnitude estrogen mitogenic effects with the permanent MTW9/PL2 cell line in culture are defined herein. As mentioned in the Background of the Invention, most existing rat mammary tumor cell lines are not suitable for use in evaluating hormone responsiveness in vivo because they are derived from outbred animals. This problem was overcome by developing the MTW9/PL2 rat mammary tumor cell line from a carcinogen-induced hormone responsive tumor (i.e. the MT-W9A tumor), first induced and transplanted in an inbred W/Fu rat as described (MacLeod RM et al. (1964) Cancer Res 75, 249-258). The MTW9/PL2 cells form hormone responsive tumors when implanted in these rats (Sirbasku DA (1978) Cancer Res 38, 1154-1165; Danielpour D and Sirbasku DA (1984) In Vitro 20, 975-980; Riss TL et al. (1986) J Tissue Culture Methods 10, 133-150). In culture, the MTW9/PL2 cells showed the same hormone responsiveness expected of rat and human breast epithelial cells, as shown herein and subsequently reported (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427; Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428- 446; Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 447-464). The effects of the steroid hormones in culture were the same as described for the growth of the original MT-W9A tumor in W/Fu rats (MacLeod et al. (1964) Endocrinology 75, 249-258) and tumor formation by the parental MTW9/PL cell line in this same strain of rats (Sirbasku DA (1978) Cancer Res 38, 1154-1165). The present embodiment is the first established cell line derived from a carcinogen induced rat mammary tumor that continues to show large magnitude growth responses to estrogens, progesterone and androgens even after extended periods in culture, preferably when cultured under the conditions disclosed herein. Thyroid hormone responsiveness has also been demonstrated for MTW9/PL cells (Leland FE et al. (1981) Li: Functionally Differentiated Cell Lines, Sato G, ed, Alan Liss, New York, pp 1-46). Of the other important hormones known to influence the growth of the original MT-W9A tumor, only prolactin remains to be investigated. Prolactin is not mitogenic for the parental MTW9/PL cells under serum-free defined conditions (Danielpour D et al. (1988) In Vitro Cell Dev Biol 24, 42-52). Continuing investigations are directed toward evaluating the possibility that prolactin also reverses the effects of the serum-borne inhibitor or otherwise acts as a cytokine to influence the production of immunoglobulins in breast and other mucosal tissues. The development of this cell line now permits not only sensitive steroid hormone growth analysis in culture, but also direct comparisons to the effectiveness of the same test substances in animals. No other such rat mammary system is currently available.

0331MTW9/PL2 Receptor Not Lost in Culture. The present results showing an average 86- fold MTW9/PL2 cell number increase in seven days in response to physiological concentrations of ---₂ have several important technical implications. Most notably, they contradict many earlier explanations for why estrogen stimulated cell growth has been difficult to demonstrate in culture. Originally, the lack of estrogenic effects in culture was thought to be due to a dedifferentiation of cells that resulted in a loss of functional receptors or some other aberration that disrupted the growth response. Li light of the present Disclosure, this explanation now seems very unlikely. The present results show the presence of similar levels of estrogen receptors in both the original MTW9/PL cell line reported in 1982 and the current MTW9/PL2 cells. Analyses made by others showing estrogen receptors in established cell lines in culture (Horwitz KB et al. (1978) Cancer Res 38, 2434-2437; Haug E (1976) Endocrinology 104, 429-437; Soto AM et al. (1988) Cancer Res 48, 3676-3680; Keydar I et al. (1979) Eur J Cancer 15, 659-670; Engel LW et al. (1978) Cancer Res 38, 3352-3364) also mitigate against this explanation. Furthermore, the estrogen receptors of the MCF-7 cells were functional based on the demonstration of estrogen inducibility of the progesterone receptor (Horwitz KB and McGuire WL (1978) JBiol Chem 253, 2223-2228). As with the human breast cancer cells, the MTW9/PL2 line was also significantly estrogen responsive by this criterion. When all of the available data is considered in light of the presently disclosed observations, the notion that long-term culture necessarily leads to loss of functional estrogen receptors is laid to rest. A major advantage of the MTW9/PL2 line is its long-term stability peπnitting series analyses over long periods of time without concern for cell property changes. Prolonged Steroid Hormone Retention by Culture Cells. It has been suggested that prolonged retention of estrogens might be the reason for a lack of responsiveness of target cells in culture (Strobl JS and Lippman ME (1979) Cancer Res 39, 3319-3327). Investigators have reported that the half-life of loss of specifically bound ³H-E₂ from MCF-7 cells was about 24 hours at 37°C (Strobl JS and Lippman ME (1979) Cancer Res 39, 3319-3327). Cells from stock cultures grown in untreated/steroid hormone containing serum were proposed to retain stimulating levels of estrogens. Even several washes over 78 hours did not attenuate the problem (Strobl JS and Lippman ME (1979) Cancer Res 39, 3319-3327). Conversely, the studies herein did not identify this problem. All the assays reported here were done with cells taken directly from cultures grown in steroid hormone containing serum (e.g. fetal bovine serum). After trypsinization of the MTW9/PL2 cells from stock culture, only three careful washes with serum-free D-MEM/F-12 were performed before initiating the growth assays. The results in Fig. 11 show clearly that 1.0 x 10^"12 M E₂ caused significant MTW9/PL2 cell growth. Also, the results in Fig. 10 show that MTW9/PL2 cells cease proliferation within 48 hours of starting a growth assay. These observations either support the conclusion that prolonged steroid hormone retention by cells is not as serious an issue as first proposed or are evidence that the technical processes described herein to prepare cells for assays have eliminated this problem. With regard to the present investigation, all cell lines studied showed this same property when prepared by the same technical process for growth assays.

0332Merits of Charcoal Extraction. Other investigators have challenged the use of charcoal extraction to deplete serum of steroid hormones. It has been stated that this procedure absorbs or otherwise alters serum to make it ineffective (Amara JF and Dannies PS (1983) Endocrinology 111, 1141-1143; Wiese TE et al. (1992) In Vitro Cell Dev Biol 28A, 595-602). To counter this problem, either individual lots of untreated serum were used to seek estrogenic effects (Wiese TE et al. (1992) In Vitro Cell Dev Biol 28A, 595-602), or serum was prepared from animals after endocrine ablation surgery (Amara JF and Dannies PS (1983) Endocrinology 112, 1141-1143). One of the best examples of use of surgically depleted serum came from the study of the GILA rat pituitary cells (Amara JF and Dannies PS (1983) Endocrinology 112, 1141-1143). They were highly F^ responsive in medium supplemented with the serum from a gelded horse (Amara JF and Dannies PS (1983) Endocrinology 111, 1141-1143). However, experience with serum derived by these methods has not been as positive. For example, this issue was inventigated in 1976 with the related GH₃Cι₄ rat pituitary tumor cell line (Kirkland WL et al. (1976) JNatl Cancer Inst 56, 1159-1164), and found that serum from ovariectomized sheep or adrenalectomized and ovariectomized sheep did not support estrogen effects. Furthermore, unextracted serum from different species can at times support limited estrogenic effects. However, the estrogenic effects are of lower magnitude than those in the CDE-serum described herein. Based on the observation that CDE-serum from a number of species was very effective, it seems highly unlikely that the now-disclosed preferred 34°C procedure is deleterious. However, it is clear that the 56°C charcoal method caused a temperature dependent loss of the inhibitor (Fig. 26). The presently described CDE-serum provides greater consistency and reproducibility than the other proposed approaches (Amara JF and Dannies PS (1983) Endocrinology 111, 1141-1143; Wiese TE et al. (1992) In Vitro Cell Dev Biol 28A, 595-602). Another advantage is that these results do not dependent significantly on the lot of serum purchased. Furthermore, CDE- serum consistently provides larger magnitude estrogenic effects than serum obtained by either of the other approaches discussed above.

0333Steroid Hormone Conjugates are Non-problematic. While charcoal treatment can be expected to remove the major classes of steroid hormones from serum, there is a question about its effect on the more soluble and potentially active conjugates. It has been reported that hydrolysis of estrogen sulfates provided free estrogens in human breast cancer cell cultures (Vignon F et al. (1980) Endocrinology 106, 1079-1086). This abrogated the effects of exogenous F^. Although the previous investigations did not address estradiol sulfate, it was shown that more than 95% of estrone sulfate and estradiol glucuronide were removed from serum by a single 56°C charcoal extraction (Sirbasku DA and Kirkland WL (1976) Endocrinology 98, 1260-1272). Additionally, in previous studies MTW9/PL cells were incubated with tritium labeled estradiol glucuronide for up to 24 hours under cell culture conditions and found no organic solvent extractable free steroid. Both past and current results indicate that the impact of the estrogen conjugates has been overestimated. In the present study, no precautions were taken to remove the conjugated forms of estrogens from any of the sera tested. Despite this, it was found that many different types of serum were effective after charcoal extraction at 34°C. Thus, it is concluded that removal of steroid conjugates by digestion or any procedure beyond charcoal treatment is unnecessary. This is a further advantage of the new CDE method because the additional treatment to remove the steroid conjugates could be prohibitive for larger scale production than a few liters. Plastic Product Use for Cell Culture. The present studies were done with plastic ware made of polystyrene. Plastic is manufactured using alkylphenols (Platt AE (1978) Li: Encyclopedia of Chemical Technology, Kirk RE, Othmer DF, eds, 3^rd Edition, Volume 26, Wiley, New York, pp 801-847). One of these compounds, p-nonyl-phenol, has been reported to be estrogenic for MCF-7 cells in culture (Soto AM et al. (1991) Environ Health Persped 91, 167-173). This xenobiotic most likely is present in the cultures used in these studies. No precautions were taken to exclude compounds leaching from plastic. Li fact, the bioassay procedures herein are done with polystyrene plastic ware and culture dishes almost exclusively. If there had been a significant contamination of the medium by such compounds, the estrogenic effects reported in this study should not have been seen or should have been markedly attenuated. An advantage of the assay systems described herein is that they have no need for expensive and or exotic substitutes for the common plastic ware used in cell culture laboratories to conduct bioassays. Also, the CDE-serum can be stored and shipped for commercial use in plastic containers without concern for creation of plastic-induced artifacts. Discussion of Example 3.

0334Using the present 34°C CDE and XAD-4 serum, estrogen responsiveness can be demonstrated in rat tumor cells, where no such responsiveness had previously been demonstrated. Further, because estrogen responsiveness and binding affinity can now be compared, data indicating the existence of an heretofore unknown estrogen receptor have been generated.

0335Example 4. Estrogen Responsive Growth of Additional Rodent and Human Cell Lines In Charcoal-dextran Extracted Horse and Human Serum

0336Li addition to the above-described studies using the MTW9/PL2 rat mammary tumor cell line, several other cell lines were employed to define the conditions for demonstrating estrogen and androgen responsive cell growth. Established cell lines from a number of different steroid hormone target tissues were selected for growth regulation analysis under those defined conditions. Estrogen Mitogenic Effects with Established ER⁺ Rodent Tumor and Human

0337Carcinoma Cells in CDE-horse Serum. Li the first study of this series, the three GH rat pituitary tumor cell lines were examined for estrogenic effects in CDE-horse serum. This was considered important in light of their clear responsiveness to many hormones (Tashjian AH Jr (1979) Methods Enzymol 58, 527-535). Furthermore, these cells are from a tissue that grows coordinately with mammary tissue in castrated rats administered exogenous estrogens. As described above, this suggested a common regulation mechanism. Fig. 13 shows an estrogenic effect > 5 CPD with GHA cells in 10 days. The results with GH₃ and GHi cells ranged between 4.0 and 5.2 CPD in 10 to 14 day assays (data not shown). The same progressive estrogen reversible CDE-serum inhibition was demonstrable with both rat mammary and rat pituitary tumor cells in CDE-horse serum. To confirm the effectiveness CDE-horse serum with human cells, the ZR-75-1 breast cancer line was selected because of previous attempts to demonstrate its estrogen responsiveness in culture (Allegra JC and Lippman ME (1978) Cancer Res 38, 3823-3829; Darbre PD et al. (1984) Cancer Res 44, 2790-2793; Darbre P et al. (1983) Cancer Res 43, 349-355). The ZR-75-1 cells showed the same CDE-serum caused estrogen reversible inhibition as seen with rodent cell lines in this serum. L 14 days, there was a 3.65 CPD (i.e. 12.5-fold) estrogenic effect (Fig. 14). This was a greater response than recorded in the ZR-75-1 cell studies cited above. Of all of the cell lines studied, the MCF-7A was the least estrogen responsive even in 50% CDE-horse (Fig. 15). The estrogenic effect was 2.65 CPD in 10-12 days. This was still significant (p < 0.01) as a 2²'⁶⁵ or 6.3-fold increase in cell number caused by estrogen. The present serum-derived inhibitor exhibited biological activity exactly opposite the estrogen reversible inhibitors described by M Tanji et al. (Tanji M et al. (2000) Anticancer Res. 20, 2779-2783; Tanji M et al. (2000) Anticancer Res. 20, 2785-2789). Additional Cell Lines Evaluated. Evidence is presented herein that the MCF-7K, T47D, LNCaP, and H301cells are highly sex steroid hormone responsive in CDE-horse serum.

0338Kinetics of Estrogen Responsive Growth in CDE Serum Containing Medium. Li the experiments presented in Figs. 16A and 16B, ER⁺ cell growth was measured daily for 15 days to determine cell growth kinetics ± E₂. The results with the T47D line are presented as characteristic of human cells. When evaluated in medium with partially inhibitory 20% (v/v) CDE horse serum, the effect of Ε₂ on cell number increase was not apparent until after 4 days (Fig. 16A). Increasing the concentration of CDE serum to 50% (v/v) further delayed the effect of E₂ (Fig. 16B). Clearly, whatever mechanism is proposed for the action of the steroid hormone, it takes a significant period to reverse the effects of the inhibitor. This process cannot be simply due to a rapid effect on transcription caused by steroid hormones. The interaction of ³H-Ε₂ with intracellular estrogen receptors saturates in < 1 hour at 37°C (Horwitz KB and McGuire WL (1978) J Biol Chem 253, 8185-8191; Maclndoe JH et al. (1982) Steroids 39, 245-258; Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427), while de novo hormone induced protein synthesis requires only a few hours (Beato M (1989) Cell 56, 335-344). Based on a growth lag of > 4 days, it is likely that steroid hormones initiate a cascade of signaling events that are more complex than recognized today. To demonstrate that the lag period was related to the inhibitor, T47D growth was monitored daily in D-MEM/F-12 supplemented with 10% (v/v) fetal bovine serum (Fig. 16A). This concentration of fetal bovine serum shows no inhibitor (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427). Cell growth in medium with fetal bovine serum showed at most a one or two day lag period.

0339Effect of CDE-human Serum on Estrogen Responsive Cell Growth. The next study examined whether human serum was a source of inhibitor for steroid hormone sensitive cell lines from different species and tissues. The results confirm that CDE-human serum contains approximately the same level of inhibitor as CDE-horse serum. Results are shown with T47D human breast cancer cells (Fig. 17A), LNCaP human prostatic carcinoma cells (Fig. 17B), MTW9/PL2 rat mammary tumor cells (Fig. 17C), two GH rat pituitary tumor cell lines (Figs. 17D and 17E), and the Syrian hamster H301 kidney tumor cells (Fig. 17F). All lines showed the same biphasic response to CDE-human serum. Low concentrations (i.e. ≤ 10%) promoted growth whereas higher concentrations (i.e. > 20%) progressively inhibited growth. Only the absolute magnitudes of the estrogenic effects varied. Replicate assays with MCF-7 A, MCF-7K and ZR-75-1 cells gave the same outcomes (data not shown). The experiments reported thus far herein support the conclusion that the inhibitor is ubiquitous in mammals and is not species specific, also subsequently reported (Sirbasku DA and Moreno-Cuevas (2000) In Vitro Cell Dev Biol 36, 428-446). Dose-response Effects of Steroid Hormones with Human Breast Cancer Cells in CDE

0340Serum. The studies presented thus far have assessed estrogen effects using 10 nM E₂. Although 10 nM saturates growth, it is decidedly at the high boundary of physiological. It is important to note that circulating estrogens in non-pregnant females are generally thought to be in the range of 10^"8 to 10^"10 M (Clark JH et al. In: Williams Textbook of Endocrinology (1992), Saunders, Philadelphia, pp 35-90). Tissue concentrations are generally conceded to be lower due to SHBG that reduces the "free" or "active" form of sex steroid hormones (Rosner W (1990) Endocr Rev 11, 80-91). The next studies with T47D cells determined the effective concentration ranges for the three most common estrogens and compared these to non-estrogen steroid hormones. Fig. 18 shows an analysis with T47D cells in D-MEM/F-12 containing 50% (v/v) CDE horse serum for 14 days. Estrogens were the only physiologically relevant activators of T47D growth. As expected from previous studies with breast cancer cells (Lippman ME et al. (1977) Cancer Res 37, 1901-1907; Jozan S et al. (1979) J Steroid Biochem 10, 341-342; Katenellenbogen BS (1980) Annu Rev Physiol 42, 17-35) and other estrogen target tissues (Clark JH and Markaverich BM (1983) Pharmacol Ther 11, 429-453), their order of effectiveness was E^ > E_! > E₃. E^ caused significant (p < 0.05) growth when present at 1.0 x 10^"14 M and optimum growth at 1.0 x 10^"10M. Higher concentrations were not inhibitory. The ED₅₀ concentration of E₂ was < 1.0 x 10^"13 M. It is noteworthy that even E₃ was remarkably potent. Others also had commented that E₃ was more potent than expected (Lippman ME et al. (1977) Cancer Res 37, 1901-1907). This observation may have special significance because breast cancers that appear during pregnancy can be particularly life threatening. Human maternal plasma has greatly elevated levels of E₃ during the last trimester of pregnancy. Testosterone and DHT promoted growth but only at supraphysiological concentrations (Fig. 18). Other investigators have suggested that supraphysiological concentrations of androgens act through the ER of human breast cancer cells (Zava DT and McGuire WL (1978) Science (Wash DC) 199, 787-788). However, another group has reported no effect of androgens on human breast cancer cell proliferation (Soto AM and Sonnenschein C (1985) J Steroid Biochem 23, 87-94). Li the present study, progesterone and cortisol were completely ineffective with T47D cells (Fig. 18). Others have also reported negative results with these hormones and human breast cancer cells (Schatz RW (1985) J Cell Physiol 124, 386-390; Soto AM and Sonnenschein C (1985) J Steroid Biochem 23, 87-94). The data presented in this Disclosure support the conclusion that the new CDE serum culture conditions yield physiologically relevant information. Dose-response Effects of Steroid Hormones with Rat Pituitary Tumor Cells in CDE

0341Serum. The GH family of related cell lines responds to a number of different classes of hormones (Amara JF and Dannies PS (1983) Endocrinology 111, 1141-1143; Tashjian AH Jr et al. (1970) J Cell Biol 47, 61-70; Tashjian AH Jr (1979) Methods Enzymol 58, 527-535; Haug E (1979) Endocrinology 104, 429437; Schonbrunn A et al. (1980) J Cell Biol 85, 786-797; Sorrentino JM et al. (1976) JNatl Cancer Inst 56, 1159-1164; Ramsdell JS (1991) Endocrinology 128, 1981-1990;

0342Hayashi I et al. (1978) In Vitro 14, 23-30; Faivre-Bauman A et al. (1975) Biochem Biophys Res

0343Commun 67, 50-57). These cells also form steroid hormone responsive tumors in W/Fu rats (Sorrentino JM et al. (1976) JNatl Cancer List 56, 1149-1154). The GH strain was selected as an example for this next study because of its marked E₂ responsiveness in culture (Amara JF and Dannies PS (1983) Endocrinology 111, 1141-1143; Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446; Sato H et al. (1991) In Vitro Cell Dev Biol 11 A, 599-602) and estrogen requirement for tumor formation in rats (Riss TL and Sirbasku DA (1989) In Vitro Cell Dev Biol 25, 136-142). The dose-response effect of steroid hormones with GELA rat pituitary tumor cells in 50% CDE-horse serum was analyzed next. Fig. 19 shows the results of these experiments. All three major estrogens promoted growth. The potencies of E₂ and Ei were equivalent whereas E₃ was substantially less effective. Even at supraphysiologic concentrations, E₃ did not promote the saturation densities seen with E₂ and Ei. The lowest concentration of E₂ and Ei that gave significant (p &lt; 0.05) growth was 1.0 x 10^"12M. The ED₅₀ of E₂ was &lt; 1.0 x 10^"" M. Optimum growth required supraphysiological concentrations (i.e. 1.0 x 10^"8M) of E₂ and Ei. Li the present studies, the biphasic effect of E. reported by Amara and Dannies (Amara JF and Dannies PS (1983) Endocrinology 112, 1141-1143) was not found. This may be explained by the different conditions used to conduct the assays. The matter of assay culture conditions with ER⁺ cells has been discussed (Zugmaier G et al. (1991) J Steroid Biochem Mol Biol 39, 681-685). Certainly however, the low E₂ concentration for ED₅₀ still speaks to a problem with ERα as the mediating receptor. Furthermore, the pattern reported in this Example is consistent with physiological facts. Tumor formation by GH cells was greater in W/Fu rats treated with 25 mg estrogen pellets than in untreated intact sexually mature females (Sorrentino JM et al. (1976) JNatl Cancer List 56, 1149-1154). Without a doubt, supraphysiological levels of estrogens were most effective in vivo. In contrast to estrogens, progesterone and cortisol had no effect on GH-A growth in culture (Fig. 19). These steroids also did not promote GH cell tumor growth in vivo (Sorrentino JM et al. (1976) JNatl Cancer L st 56, 1149-1154). The findings with androgens and GHA cell growth shown in Fig. 19 revealed another important contribution made by the work in CDE serum supplemented cultures described herein. It has been shown before that T promoted GH tumor growth in vivo (Sorrentino JM et al. (1976) JNatl Cancer List 56, 1149- 1154). It was proposed at that time that T was effective because it was metabolized to estrogens in the rat. Therefore, it was expected that T would be ineffective in culture. The results in Fig. 19 confirm this expectation. Li this case, the new culture methods permitted resolution of an issue arising from previous in vivo observations. The dose-response results in Fig. 19 fortify a conclusion arrived at earlier that cell culture can be used to uncover physiologically important new information not accessible by in vitro methods (McKeehan WL et al. (1990) In Vitro Cell Dev Biol 26, 9-23).

0344Dose-response Effects of Steroid Hormones with Hamster Kidney Tumor Cells in CDE Serum. To explore the utility of the new culture conditions further, steroid hormone effects on the H-301 Syrian hamster kidney tumor cells in D-MEM/F-12 containing 50% (v/v) CDE-horse serum were investigated. This cell line has two unique characteristics. First, tumors form from

0345H301 cells in Syrian hamsters only in response to exogenous estrogens (Sirbasku DA and Kirkland WL (1976) Endocrinology 98, 1260-1272). It is very important to note that normal physiologic levels in intact adult female hamsters do not support tumor formation (Sirbasku DA and Kirkland WL (1976) Endocrinology 98, 1260-1272). It is thought that progesterone from the normal estrus cycle suppresses growth in response to physiological levels of estrogen (Kirkman H and Robbins M (1959) Li: National Cancer Institute Monograph No. 1, National Institutes of Health, Bethesda, MD). Second, these cells only form tumors in response to estrogens. The other major classes of steroid hormones are ineffective in vivo. The relative effectiveness of the three estrogens with H301 cells was investigated (Fig. 20). Their potency was E₂ > Ei > E₃. As with rat tumor cells, E₃ was markedly less effective than E^ or E]. E^ and Ei required 1.0 x 10^"11 M and 1.0 x 10^"10 M, respectively, to achieve significant (p < 0.05) growth. The ED₅₀ concentration of E₂ is about 5 to 9 x 10^"11 M. As expected from in vivo results (Sirbasku DA and Kirkland WL (1976) Endocrinology 98, 1260-1272), this concentration was higher than for the rat pituitary tumor cells (Fig. 19) or rat mammary tumor cells (Fig. 11). Li fact, they were as much as 100 to 1000-fold higher than for human breast cancer cells (Fig. 18). Li other tests shown in Fig. 20, progesterone, cortisol, T and DHT were all inactive. The higher estrogen concentrations required for significant growth of the H- 301 cells in culture, coupled with the marked estrogen specificity, indicate that the medium conditions used in this study yielded physiologically germane results.

0346Dose-response Effects of Steroid Hormones with Human Prostatic Carcinoma Cells in CDE Serum. In the final dose-response study, the potency of several classes of steroid hormones with the LNCaP cells was analyzed. This was done in D-MEM/F-12 containing 50% (v/v) CDE horse serum. Due to a point mutation which permits bmding of both androgen and non-androgen hormones to the AR of LNCaP cells (Veldscholte J et al. (1990) Biochem Biophys Res Commun 173, 534-540; Veldscholte J et al. (1990) Biochim Biophys Ada 1052, 187-194), the Inventor expected several classes of steroids to promote growth, albeit at concentrations compatible with their known affinities for the mutated receptor. This proved to be the case, as shown in Fig. 21. DHT and E₂ were the most potent steroids. Li fact, they were equipotent. Both caused significant (p < 0.05) growth at 1.0 x 10^"12 M. Contrary to other reports (Schuurmans AL et al. (1988) The Prostate 12, 55-64; Sonnenschein C et al. (1989) Cancer Res 49, 3474-3481; de Launoit Y et al. (1991) Cancer Res 51, 5165-5170; Lee C et al. (1995) Endocrinology 136, 796-803; Kim I et al. (1996) Endocrinology 137, 991-999), the present study did not find that high concentrations of DHT inhibited LNCaP growth. The potency of the steroid hormones tested was DHT = E₂ > T > Ei > progesterone > E₃ > cortisol. As potencies declined, saturation densities also decreased. The observed relative steroid potencies agreed with those of others (Belanger C et al. (1990) Ann NY Acad Sci 595, 399-402), and correlated with the expected binding of the various classes of steroids to the mutated AR of the LNCaP line. Additionally, the presently disclosed methods offered the advantage of greater growth responses. The results in Fig. 21 not only lend support to the view that cultures containing a high concentration of CDE serum yield physiologically relevant information, but they also demonstrate that the new charcoal extraction method disclosed herein effectively depletes several classes of steroid hormones.

0347Comparisons of ED_S0 and K as Evidence Supporting a New ER Designated ERγ. As mentioned in the Background of the Invention, it is important to recognize that if a given estrogen receptor is in fact a mediator of estrogen-induced growth, then the steroid hormone concentrations required for one-half maximum growth (i.e. ED₅₀), or for optimum growth (i.e. ED_]0o), should be about the same. According to the theory of hormone binding, the K value represents the steroid concentration that one-half saturates the existing receptors. The following TABLE 5 summarizes the ED₅₀ concentrations required for a one-half maximum growth and the corresponding lowest K_d measured for the same or closely related cell lines:

0348TABLE 5 Comparisons of ED₅₀ and ICj as Evidence Supporting a New ER Designated ERγ

0349<img file="EP1328293A2_D0005.tif" />

0350Clearly, to seek the new ERγ, the rat mammary or human breast cells will be the best sources based on the differences between the ED₅₀ growth concentrations and the K values for ERα or ERβ. Because the ERβ was first obtained from rat tissues, the MTW9/PL2 cells will be the preferred source of ERγ.

0351One preferred application supported by the data in TABLE 5 is the use of the ERγ for diagnosing and/or screening for breast cancer. Measurement of the ERγ specifically will provide a more accurate determination of estrogen receptor status and therefore permit more precise modeling of the therapy for each patient. ERγ will be identified by immunohistochemical methods, labeled ligand binding with very high specific activity isotopes, and by PCR and other molecular biology analyzes. Other methods will also be applied. Similar analyses are expected to be applicable to other estrogen receptor related or estrogen receptor containing mucosal cancers including ovarian, uterine, vaginal, cervical, colon, lung, stomach, pituitary, liver, pancreas, skin and kidney, as described in co-owned, concurrently filed U.S. Patent App. No. (Atty.

0352Dkt. No. 1944-00800)/PCT/US2001/ (Atty. Dkt. No. 1944-00801) entitled "Compositions and Methods for the Diagnosis, Treatment and Prevention of Steroid Hormone Responsive Cancers," which is hereby incorporated herein by reference. The dose-response results presented in Figs. 11, and 18 through 21 demonstrate the usefulness of the extracted sera, assays and cell lines with regard to assessment of estrogenic activity or androgenic activity in industrial, commercial, environmental, medicinal, or other medical samples where activity measurement is required at concentrations below the usual levels detectable by radioimmunoassay. The sensitivity of this bioassay is unique.

0353Benign prostatic hypertrophy (BPH) is among the most common afflictions of older men. About 50% of 60-year old men have BPH. At 85 years about 90% of men have BPH (Berry SJ et al. (1984) J Urol 132, 474-479). There is a general view that estrogens may be important in BPH (Henderson D et al. (1987) Steroids 50, 219-229; Nakhla AM et al. (1994) Proc Natl Acad Sci USA 91, 5402-5405). The paradox involved is that as men age androgen levels fall and SHBG rises. These work in concert to further limit available androgen (Davidson JM et al. (1983) J Gin Endocrinol Metab 57, 71-77; Tenover JS et al. (1987) J Gin Endocrinol Metab 65, 1118-1126). Furthermore, as part of the weight gain with age, estrogens become more prominent in older men. Although it has been suggested that estrogens cause LNCaP cell growth via an estrogen receptor, it remains to be proven conclusively. Nonetheless, the ERγ may be expressed in BPH and prostatic cancer and therefore its use as a diagnostic tool and a site for development of new antihormone treatments of these diseases has great potential.

0354Discussion of Example 4. The results presented in this Example have special significance with regard to support for the conclusion that a new ERγ regulates growth and is activated by more than 10-fold at lower concentrations of E₂ than expected of the classical EDα. Example 4 also demonstrates the utility of assays using 34°C CDE serum for demonstrating estrogen responsive cell growth in a variety of tissues.

0355Example 5. Thyroid Hormone Growth Effects in CDE-Horse Serum Prepared at 34°C

0356This Example demonstrates that not only steroid hormone but also thyroid hormone growth effects can be demonstrated in cell growth assays using the present 34°C CDE serum. Thyroid Hormone Responsive Pituitary Tumor Cell Growth in CDE-Serum

0357Prepared at 34° C. GH rat pituitary tumor cells are highly thyroid hormone responsive in serum- free defined medium (Eby JE et al. (1992) Anal Biochem 203, 317-325; Eby JE et al. (1992) J Cell Physiol 156, 588-600; Sato H et al. (1991) In Vitro Cell Dev Biol 21 A, 599-602). An example of this responsiveness with the GH₃ line is shown in Fig. 22. However, in serum-free defined medium, these cells are not E₂ responsive when T₃ is omitted from the medium (Fig. 23). During evaluation of the role the GH cell lines in CDE-serum, in D-MEM F-12 with 2.5% (v/v) CDE- horse serum, T₃ caused substantial growth of the GH-A, GHi and GH₃ rat pituitary tumor cell lines (Fig. 24). However, at 50% (v/v) CDE-horse serum, only supraphysiologic concentrations of thyroid hormone showed growth effects (Fig. 25). Nonetheless, the 34°C CDE method described in the preceding Examples is clearly functional to demonstrate both steroid hormone and thyroid hormone growth effects in culture. It is known that the thyroid hormone receptor is a member of a superfamily of receptors that also includes the steroid hormone receptors (Evans RM (1988) Science (Wash DC) 240:889-895). Testing of substances expected to have thyroid hormone like activity can be performed with the GH cell lines in the presence of low concentrations of CDE- serum. Discussion of Example 5. The removal of thyroid hormones from serum has been described before using the Bio-Rad ™ AG-1 X8 ion exchange resin (Samuels HH et al. (1979) Endocrinology 105, 80-85). Removal of T₃/T₄ by the AG-1 X8 method relies on their negative carboxylic acid charge at neutral pH. However, ion exchange does not remove the uncharged/hydrophobic steroid hormones. This Example demonstrates that the 34°C CDE method described herein is more effective than the AG-1 X8 method previously known.

0358Example 6. Effect of 56°C Versus 34°C CDE-horse Serum on MTW9/PL2 Cell Growth.

0359Previously, unsuccessful attempts were made to identify estrogen responsive tumor cell growth in cultures supplemented with serum depleted of steroid hormones by a 56°C charcoal extraction procedure (Kirkland WL et al. (1976) JNatl Cancer Inst 56, 1159-1164; Sirbasku DA and Kirkland WL (1976) Endocrinology 98, 1260-1272; Sirbasku DA (1978) Proc Natl Acad Sci USA 75, 3786-3790; Leland FE et al. (1982) In: Cold Spring Harbor Conferences on Cell Proliferation, Volume 9, Growth of Cells in Hormonally Defined Media, Cold Spring Harbor, New York, pp 741- 750; Liehr JG and Sirbasku DA (1985) Li: Tissue Culture of Epithelial Cells, Taub M, ed, Plenum, New York, pp 205-234; Riss TL and Sirbasku DA (1989) In Vitro Cell Dev Biol 25, 136-142). Li light of the data presented in the foregoing Examples, it appears that the 56°C method was the major problem. The high temperature may have inactivated the inhibitor. Alternatively, because the 56°C method was done for only a brief period, it may not have sufficiently removed the steroid hormones. Clearly, from the results presented above, even modest levels of residual estrogens can promote growth. This latter possibility seemed likely because the 56°C method removed only somewhat more than 90% of the serum steroid hormones (Kirkland WL et al. (1976) J Natl Cancer Inst 56, 1159-1164; Sirbasku DA and Kirkland WL (1976) Endocrinology 98, 1260-1272). To reevaluate this problem, the E₂ effects on MTW9/PL2 cell growth were compared in medium supplemented with either 34°C or 56°C CDE-horse serum. As expected, the assay with control 34°C treated serum, prepared as described in Example 2, showed maximum estrogenic effects of 6.01 CPD (Fig. 26). By comparison, the same lot of serum that had been charcoal extracted at 56°C showed a maximum estrogenic effect of only 2.96 CPD (Fig. 26). When 34°C CDE-serum was either charcoal extracted again at 56°C, or heated at this temperature for 20 minutes without charcoal, E^ induced growth was reduced to only 1.47 and 2.01 CPD, respectively (Fig. 26). A typical assay from which these results were calculated is shown in (Fig. 26, insert). This experiment demonstrates that 56°C treatment results in the loss of the inhibitory activity in serum. It should be noted that many investigators routinely "heat inactivate" serum at 56°C for 20 to 30 minutes to destroy complement. The results indicate that this heating should be avoided when the serum is to be used in cell culture experiments testing steroid hormone growth responsiveness.

0360Discussion of Example 6. This example makes clear some major differences between the serum-borne inhibitor presently disclosed and those previously described. Specifically, exposure to heat can inactivate or alter the effect of inhibitors. For example, U.S. Patent Nos. 4,859,585 (Sonnenschein) and 5,135,849 (Soto) describe an inhibitor that was derived from heat inactivated (i.e. 56°C treated) serum and thereafter depleted of its endogenous estrogens and androgens by a 37.5°C single step charcoal-dextran procedure. The facts of that method are also stated in a publication (Soto AM and Sonnenschein C (1984) Biochem Biophys Res Commun 122, 1097-1103). Li light of the results presented in Fig. 26, it is likely that the inhibitor described by Sonnenschein and Soto is a different molecular entity, as further illustrated in Examples which follow. Among other differences, the serum used to isolate the present active inhibitor has not been inactivated by exposure to heat.

0361Example 7. Demonstration of Estrogenic Effects in XAD-4 Resin Treated Horse Serum

0362Horse serum depleted of steroid hormones by XAD-4™, prepared as described in Example 2.C, was assayed to determine if it demonstrated estrogen reversible inhibition of ER⁺ cancer cell growth in culture. Fig. 27 shows the effects of XAD-4 treated horse serum + 10 nM E₂ with the MTW9/PL2 cell line. Unmistakably, the pattern of cell response was the same as seen with CDE-horse serum prepared as described in Example 2. A.. At 50% XAD-4 serum (v/v), an estrogenic effect of 5.2 CPD was observed in 7 days. Fig. 28 shows a similar experiment with T47D cells after 14 days. At 50% (v/v) XAD-4 treated serum, an estrogenic effect with T47D cells of 5.3 CPD was observed. The magnitudes of the estrogenic effects with both cell lines were the same as observed with CDE-horse serum. Because both MTW9/PL2 and T47D cells are sensitive to picomolar concentrations of estrogen, it was evident that the XAD-4 resin treatment effectively removed the endogenous sex steroids present in serum.

0363Discussion of Example 7. There is no previous report of the preparation of steroid depleted serum by this resin treatment method. As indicated in Example 2, the XAD-4™ treatment method has particular applicability for the industrial preparation of large volumes of steroid hormone depleted serum, and will allow the commercial supply of steroid depleted serum at reasonable cost. A preferred application for this steroid hormone stripped serum is in the biotechnology industry, in which cell culture is used to produce medically and otherwise commercially significant proteins and cellular products. Steroid hormone depleted serum has applicability beyond the ER⁺ and AR⁺ cells described in this report. For example, hybridoma cells are the sources of many important monoclonal antibodies. Depletion of steroids from the serum used to grow these cells will increase cell viability (cortisol is a potent cytotoxic agent) and therefore increase product yield. These and other applications of the XAD-4™ treated serum for both commercial and diagnostic testing as well as for industrial production of valuable cellular products are foreseen.

0364Example 8. Testing of Substances for Estrogenic Activity.

0365The purported estrogenic effects of phenol red were tested and proven to be unfounded. Further, the methods described in this Example exemplify methods that are generally effective for assessingthe steroidogenic activity of any substance.

0366Phenol Red as an Estrogen. It is widely believed that the phenol red indicator in tissue culture medium acts is a weak estrogen (Berthois Y et al. (1986) Proc Natl Acad Sci USA 83, 2496- 2500). Li the first report describing the phenol red problem, the indicator itself was thought to act as an estrogen (Berthois Y et al. (1986) Proc Natl Acad Sci USA 83, 2496-2500). At the concentration in standard culture media (e.g. D-MEM/F-12 is 8.1 mg/mL or 22.9 μM), it was believed to stimulate ER⁺ cell growth nearly as well as natural estrogens. Simply stated, this meant that exogenous estrogens would have no effect because the cells were already nearly completely stimulated. Further work by the original investigators later demonstrated that it was the lipophilic impurities in phenol red that were the true culprits (Bindal RD et al. (1988) J Steroid Biochem 31, 287-293). The chemical structure of one was detei-rnined to be bis(4-hydroxyphenyl)[2- (phenoxysulfonyl)phenyl]methane (Bindal RD and Katzenellenbogen JA (1988) J Med Chem 31, 1978-1983). Interfering amounts of the impurities were identified in many different commercially available preparations of phenol red (Bindal RD et al. (1988) J Steroid Biochem 31, 287-293; Bindal RD and Katzenellenbogen JA (1988) J Med Chem 31, 1978-1983). These investigators concluded that many, if not most, phenol red containing culture media had sufficient contaminants to at least partially mask estrogenic effects. Despite the wide acceptance of phenol red as an estrogen, experience has shown differently. Instead, large estrogen mitogenic effects have been observed in phenol red containing culture medium with ER⁺ MCF-7 human breast cancer cells, T47D human breast cancer cells, MTW9/PL2 rat mammary tumor cells, GH rat pituitary tumor cells, H301 Syrian hamster kidney tumor cells and the androgen receptor positive (AR*) and ER⁺ LNCaP human prostatic carcinoma cells, as shown herein and subsequently reported (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427; Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446, incorporated by reference). Li these studies, even when phenol red was present, estrogen-inducible cell number increases of 8 to 80-fold were observed. These responses were as large or larger than any previously reported. They exceeded any reported in phenol red free medium. Also, growth without the natural hormone was very limited even with phenol red present (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427; Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446). Phenol Red Indicator is a "Red Herring". Phenol red was further evaluated. Head-on comparisons of the growth of nine different ER⁺ cell lines representing four target tissues and three species in medium with and without phenol red were performed (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 447-464). These studies were designed to specifically test various aspects of published reports that phenol red is estrogenic. Considering the results of these head-on comparisons, new conclusions have been reached about the effects of phenol red in culture, especially as they are relevant to experimental conditions available today to most investigators. Even more important, the test assays show the methods that can be used to determine if any commercial preparation or other source material possesses estrogenic activity. To do this, nine cell lines were employed in the tests. Five different experimental protocols were used to investigate phenol red. First, F^ responsive growth of all nine ER⁺ cells lines was compared in medium with and without the indicator. Second, using representative lines it was asked if phenol red was mitogenic in indicator free medium. The dose-response effects of phenol red were compared directly to those of E₂. Third, it was asked if tamoxifen inhibited growth equally in phenol red containing an indicator-free medium, which would also confirm or refute a report indicating that antiestrogen effects should be seen only in phenol red containing medium. Fourth, it was asked if phenol red displaced the binding of ³H-E₂ using ER⁺ intact human breast cancer cells. Fifth, E₂ and phenol red were compared as inducers of the progesterone receptor using a human breast cancer cell line. All of the experiments reported (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 447-464) support the conclusion that the concentration of phenol red contaminants in a standard culture medium available today is not sufficient to cause estrogenic effects. The real issue of how to demonstrate estrogenic effects in culture resides elsewhere than phenol red. Demonstration of sex steroid hormone mitogenic effects in culture depends upon conditions that maximize the effects of a serum-borne inhibitor, as described in foregoing Examples. When the effects of the inhibitor are optimized, the presence or absence of phenol red makes no everyday difference to the demonstration of estrogen mitogenic effects with several target cell types from diverse species.

0367Phenol Red Testing for Estrogenic Activity with MCF-7A Cells. The original reports of the effect of phenol red or its impurities had used the MCF-7 human breast cancer cells to assess estrogenic activity (Berthois Y et al. (1986) Proc Natl Acad Sci USA 83, 2496-2500; Bindal RD et al. (1988) J Steroid Biochem 31, 287-293; Bindal RD and Katzenellenbogen JA (1988) J Med Chem 31, 1978-1983). The initial study began with the MCF-7A strain of this population. As shown in Fig 29A, growth was measured in the presence of increasing concentrations of CDE-horse serum with and without phenol red in the medium and ± E₂. Concentrations of &lt; 10% (v/v) CDE- horse serum supported more than 5 CPD. Higher concentrations progressively inhibited in both indicator containing and indicator free medium. Li both types of medium, Ε₂ was required to reverse the serum inhibition. To confirm that E₂ was equally effective in phenol red free and phenol red containing medium, the estrogenic effects shown in Fig. 29A were compared in both types of medium and at each serum concentration. The results of this analysis are presented in Fig. 29B. The maximum estrogenic effect at 50% (v/v) serum was 2.38 CPD (i.e. 2^2'38 or 5.2-fold) in medium without indicator and 2.56 CPD (i.e. 2^2'56 or 5.9-fold) with phenol red. This difference was not significant. Only at 5% (v/v) serum was there a significantly (p &lt; 0.05) greater estrogenic effect in phenol red free medium. However, in replicate experiments this &lt; 1.0 CPD effect was inconsistent. At all other serum concentrations, the growth differences between plus and minus phenol red were not significant.

0368Test of Phenol Red Effects with MCF-7K Cells. The MCF-7K strain was routinely more estrogen responsive than the MCF-7A line (Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446). The MCF-7K cells also showed a serum concentration dependent growth inhibition (Fig. 29C). The final degree of inhibition at 50% (v/v) serum was independent of phenol red. Only in the presence of 2.5, 5, 10 and 20% (v/v) CDE-horse serum were the estrogenic effects significantly greater in phenol red free (Fig. 29D). It is important to note that while these differences were identified more often with the MCF-7K strain than the MCF-7A line, they were invariably small. Plainly, no serum concentration supported > 1.0 CPD estrogenic effects in phenol red free medium compared to indicator free medium (Fig. 29D). Li fact, deletion of phenol red improved estrogen responsiveness by an average of only 0.6 CPD with the MCF-7K line. When judged by the maximum estrogenic effects achievable with MCF-7K cells in 50% (v/v) CDE-horse serum, plus and minus phenol red gave indistinguishable results of CPD 3.01 (8.0-fold) and CPD 2.99 (7.9-fold), respectively (Fig.29D). Phenol Red Testing for Estrogenic Activity with T47D and ZR-75-1 Cells. The same experiments just described above with the MCF-7 cell strains were repeated with T47D and ZR-75-1 cells. These lines were substantially more estrogen stimulated in CDE-serum than MCF-7 cells (Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446) and hence were expected to be more sensitive to phenol reάVcontaminants. Phenol Red and T47D Cells. T47D cells were grown in medium with CDE-horse serum both with and without phenol red (Fig. 30A). Low concentrations of serum (i.e. < 2%) promoted growth. Higher concentrations progressively inhibited growth irrespective of indicator content. Li both media, E₂ was required to reverse the inhibition (Fig.30A). Li 50% (v/v) CDE-horse serum, the maximum E₂ responses were 2⁵'³⁵ (41 -fold) and 2⁵'²⁹ (39-fold) in phenol red containing and indicator free medium, respectively (Fig. 30B). Only at low serum concentrations were phenol red effects observed in any experiment. Li some replicates, the phenol red effect was opposite to that expected. For example, in the experiment shown in Fig. 30B, 0.5 to 2.5% serum showed significantly (p < 0.05) greater estrogenic effects in the presence of phenol red. These results graphically illustrate the hazards of inte reting 1.0 CPD responses either in favor of phenol red/contaminants as estrogens or in opposition to this proposal.

0369Phenol Red and ZR-75-1 Cells. ZR-75-1 cells showed similar results as the T47D line. Serum caused an inhibition of growth that was undoubtedly unrelated to phenol red (Fig. 30C). Li both types of medium, and at every serum concentration tested, E₂ was required to reverse the inhibition (Fig. 30C). Li 50% (v/v) serum, ZR-75-1 cells showed maximum estrogenic effects of 2³³⁹ (10.5-fold) and 2³⁴⁹ (11.2-fold) in medium with and without indicator, respectively (Fig. 30D). As seen with T47D cells, the ZR-75-1 line showed greater estrogenic effects in medium with phenol red than in medium without indicator when the serum was 0.5, 5 or 10%> (v/v) (Fig.30D).

0370Phenol Red Testing for Estrogenic Activity with MTW9/PL2 Cells The next experiments were done with MTW9/PL2 rat mammary tumor cells (Fig. 31A). They were inhibited by high concentrations of CDE-horse serum with and without indicator. E₂ was required to reverse the inhibition in both types of medium (Fig. 31A). The maximum estrogenic effects in 50% serum were 2^{5 82} (56-fold) and 2⁵⁶⁹ (52-fold) with and without phenol red, respectively (Fig. 31B). Li the experiment shown in Fig. 31B, estrogenic effects were unpredictably greater in phenol red free medium than in medium with indicator. This was observed at low serum concentrations (i.e. 0.5 and 1.0 %) and again at higher levels (i.e. 20 and 30%). Although suggesting a phenol red effect, these results in fact only serve to emphasize the pitfalls of accepting small changes as meaningful even though they are significant at p &lt; 0.05. When estrogenic effects were found with MTW9/PL2 cells in phenol red free conditions, they invariably were &lt; 1.0 CPD. The sum of the studies with MTW9/PL2 cells did not yield a predictable correlation between estrogenic effects in the absence of the indicator and serum concentrations.

0371Other Cell Lines Tested for Growth + Phenol Red and + E₂. The results presented above were replicated with the GHi and GHA rat pituitary tumor cell lines as well as with the H301 cells and the LNCaP cell line (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 447-464). Again, the presence or absence of the indicator in the medium containing CDE- horse serum had no effect whatever on the demonstration of the usual high estrogenic effects with these cells. Direct Test of Phenol Red Estrogenic Activity. Three cell lines were selected for a direct test of phenol red as a mitogen. The MCF-7A line was used because it most closely approximated the origin and passage age of the cells used to conduct the original study of phenol red as a weak estrogen (Berthois Y et al. (1986) Proc Natl Acad Sci USA 83, 2496-2500). The T47D cells were chosen because they are the most estrogen responsive human breast cancer cell line available today (Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446). The MTW9/PL2 cells were chosen as an example of a highly estrogen responsive rodent origin line (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427; Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446). The assays were done in phenol red free D-MEM/F-12 supplemented with 30% CDE-HS. This concentration was chosen even though it is not as inhibitory as 50% (v/v) serum. This selection was made to reduce possible interactions of the phenol red/contaminant with serum proteins while still retaining a significant inhibitory effect. Phenol red concentrations of up to 16 mg/L were added to this medium. This highest level was twice that in standard commercially formulated Gibco-BRL D-MEM/F-12. Several different manufacturing lots of aqueous phenol red gave equivalent results. The preparations used in this study ranged in age from newly obtained to more than ten year old laboratory stocks. These experiments gave unmistakable results. There was no increase in the growth of any of the cell lines in response to phenol red (Fig. 32A). By comparison, parallel cultures receiving E₂ showed sizable 2 to 5 CPD responses to the natural hormone (Fig. 32B). E₂ at 1.0 x 10^"10 M optimized growth of all three cell lines. The ED₅₀ concentrations of E₂ were 3.0 x 10^"12 M. Significant (p &lt; 0.05) estrogenic effects were observed at 1.0 x 10^"12 M. The results presented in Fig. 32 indicate that the culture conditions used in this study could reasonably be expected to detect responses due to contaminants present at the concentrations indicated in the original reports (Berthois Y et al. (1986) Proc Natl Acad Sci USA 83, 2496-2500; Bindal RD et al. (1988) J Steroid Biochem 31, 287-293; Bindal RD and Katzenellenbogen JA (1988) JMed Chem 31, 1978-1983).

0372Comparison of E₂ Potency in Medium with and without Phenol Red. As described above in TABLE 5, the T47D and MTW9/PL2 cells grow significantly in response to 1.0 x 10^"12 M E₂. The D-MEM/F-12 used in those studies also contained about 23 μM phenol red. When the results of those studies were compared to the experiments in Fig.32B, done in D-MEM/F-12 without indicator, the estrogen dose response curves were very similar. The conclusion is straightforward. E₂ dose-responses were not affected by phenol red. If phenol red lipophilic contaminants were present at the concentrations originally suggested (Berthois Y et al. (1986) Proc Natl Acad Sci USA 83, 2496-2500; Bindal RD et al. (1988) J Steroid Biochem 31, 287-293; Bindal RD and Katzenellenbogen JA (1988) JMed Chem 31, 1978-1983) they should have masked the observation of picomolar effects of exogenous estrogens.

0373Effect of Phenol Red on Binding of ³H-E₂ to Intact Cells. For the next study, intact T47D cells were used to measure the effects of phenol red on estrogen receptor binding. The cells were incubated with 5 nM H-E₂ and the effects of addition of increasing concentrations of unlabeled E> assessed (TABLE 6). A 100-fold excess of unlabeled E₂ displaced 75% of the binding of ³H-E>. By this criterion, 75% of the binding of ³H-E₂ was specific to estrogen receptors (Chamness GC and

0374McGuire WL (1975) Steroids 26, 538-542). The same analysis was conducted with aqueous preparations of phenol red. Even at 16 mg/L, the indicator did not reduce the binding of ³H-E₂

0375(TABLE 6). This was true no matter which batch of indicator was analyzed (results not shown).

0376The phenol red used for the experiment shown in TABLE 6 was approximately the same age

0377(purchased in 1986) as the date of the original report (Berthois Y et al. (1986) Proc Natl Acad Sci

0378USA 83, 2496-2500). These results raise the question how often preparations of phenol red purchased at that time as an aqueous membrane filtered product contained a sufficient level of contaminants to elicit an estrogenic effect.

TABLE 6

0380Displacement of ³H-E₂ Binding to Intact T47D Cells by Unlabeled E₂ or Unlabeled Phenol Red Indicator Free and Serum-free D-MEM/F-12 for Two Hours at 37°C

0381<img file="EP1328293A2_D0006.tif" />

0382Comparison of the E₂ and Phenol Red Induction of Progesterone Receptors. Another putative function of phenol red was to induce progesterone receptors in estrogen sensitive cells. An investigation was made as to whether the indicator induced an increase in the progesterone receptors of T47D cells which contain these sites (Horwitz KB et al. (1978) Cancer Res 38, 2434-2437). Li a first study, the kinetics of progesterone receptor induction versus estrogen concentration in phenol red free medium were investigated (Fig.33A). E₂ levels as low as 1.0 x 10^"12 M caused a significant two-fold increase in receptor content in four days. At 1.0 x 10^"8 M, E₂ induced a four-fold increase in progesterone receptors in four days. Clearly, E^ induced a time and concentration dependent increase in the progesterone receptors with T47D cells. Next, this same analysis was done with phenol red over a concentration range of 1 to 16 mg/L (Fig. 33B). Phenol red induced a small increase in progesterone receptors at 8 and 16 mg/L after four days. This induction was about the same as caused by 1.0 x 10^"14 M E₂ (Fig. 33 A). These results indicate that if estrogenic contaminants are present in phenol red, they are most likely in the 10^"14 M range even assuming equal receptor binding capacity to E₂. This point is important because the active agent is thought to be only a trace impurity in many batches of phenol red (Bindal RD et al. (1988) JMed Chem 31, 1978-1983). The impurities bind to the estrogen receptor with only 50% of the affinity of E₂. The impurity was expected to be 0.002% of the phenol red concentration. Based on test results that employed many different batches of Gibco-BRL D-MEM/F-12, this concentration of the impurity seems highly unlikely in the medium commercially available today.

0383Discussion of Example 8. The studies of the effects of phenol red or its lipophilic impurities demonstrate the usefulness of the presently disclosed methods for the assessment of estrogenic and androgenic activity of commercially prepared materials, substances present or extracted from environmental or food sources or other sources that are thought to contain such activities. The testing can be approached by three separate methods as shown by examples with phenol red. (1) Compounds or other preparations and substances can be tested for growth activity with human or rodent cell lines depending upon the information sought. Potency can be established as UNITS based on Ε₂ or any other estrogen or androgen required. This permits direct expression of the estrogen like activity or androgen like activity per volume or mass of the substance under evaluation. Levels can be measured without regard for expensive development of a radio immunoassay that in the end still does not yield evidence of biological activity as a sex steroid hormone analog (agonist or antagonist). The use of rodent cell lines opens the possibility of direct comparison to in vivo activity if required. (2) Another form of analysis is direct measure of potency by ³H-E₂ or ³H-DHT binding displacement analysis from whole cells or extracted estrogen receptors. An example with ³H-E₂ and whole cells is shown in TABLE 6. The two different binding assays offer different information. Whole cells have a predominance of hydrophobic sites (i.e. membranes) that absorb lipophilic substances and therefore may attenuate their activity. Use of cell extracted sex steroid hormone receptors permits direct measure of the potential of a substance to act as a hormone independent of its biological effects. (3) Finally, use of the progesterone receptor analysis permits evaluation of substances and preparations by a method entirely independent of growth. This is a gene expression based analysis that permits evaluation that can be used to supplement growth data or be used in place of growth analysis. The MTW9/PL2 cells have been shown above to be suitable for this purpose.

0384Example 9. Testing of Substances for Inhibitor-like Activity

0385Li studies described in this Example, TGFα, TGFβl, EGF, IGF-I, IGF-II and insulin were tested for inhibitor-like acitivity, using the cell growth assay described in the General Materials and Methods section,and in the foregoing Examples, substituting those proteins for the serum-borne inhibitor contained in the preferred CDE serum.

0386TGFβl as a Substitute for the Serum-borne Estrogen Reversible Inhibitor. Normal mouse mammary (Silberstein GB and Daniel CW (1987) Science (Wash DC) 237, 291-293; Silberstein GB et al. (1992) Dev Biol 152, 354-362) and normal human breast epithelial cell growth is inhibited by TGFβ (Bronzert DA et al. (1990) Mol Endocrinol 4, 981-989). Additionally, human breast cancer cells are inhibited by TGFβ (Knabbe C et al. (1987) Cell 48, 417-428; Arteaga CL et al. (1988) Cancer Res 48, 3898-3904; Arteaga CL et al. (1990) Cell Growth Diffl, 367-374). TGFβ also inhibits the GELA rat pituitary tumor cells (Ramsdell JS (1991) Endocrinology 128, 1981- 1990) and the LNCaP human prostatic carcinoma cells (Schuurmans AL et al. (1988) The Prostate 12, 55-64; Wilding G et al. (1989) Mol Cell Endocrinol 61, 79-87; Carruba G et al (1994) Steroids 59, 412-420; Castagnetta LA and Carruba G (1995) Ciba Found Symp 191, 269-286; Kim IY et al. (1996) Endocrinology 137, 991-999). In studies presented next, replacement of the serum-borne inhibitor with TGFβ was attempted. A number of related forms of this inhibitor are known (Clark DA and Coker R (1998) Int J Biochem Cell Biol 30, 293-298; Massague J (1998) Annu Rev Biochem 67, 753-791). TGFβl and TGFβ2 are most often studied and commonly have similar potencies. For example, they are equipotent with human breast cancers cells (Zugmaier G et al. (1989) J Cell Physol 141, 353-361). TGFβl was chosen for the instant study. Without a doubt, a number of the key cell lines used throughout the Examples were inhibited by TGFβ. It was therefore considered essential to ask if TGFβ was the estrogen reversible inhibitor.

0387TGFβl and MCF-7 Cells. Because MCF-7 cells are probably the most studied human breast cancer line today, this next work began with those cells. TGFβ has been described as a hormone regulated autocrine inhibitor of the ER⁺ MCF-7 human breast cancer cell growth (Knabbe C et al. (1987) Cell 48, 417-428). In the present study, to test if TGFβl substituted for the serum- borne inhibitor with these cells, they were grown in D-MEM/F-12 containing 2.5 % (v/v) CDE-horse serum plus increasing concentrations of fransforming growth factor and + E₂. The results in Fig. 34A show that even 50 ng/mL of TGFβl caused only a modest inhibition of MCF-7K cell growth. Cell numbers were reduced from 350,000 to 200,000 per dish. This difference was significant (p &lt; 0.05). Nevertheless, the estrogen reversal of the inhibition was no larger than the F^ effect observed in D-MEM F-12 containing 2.5% (v/v) horse serum without TGFβl Fig. 34A. Furthermore, when the cell number data were expressed as CPD (insert Fig.34A), it was definite that TGFβl was at best a very modest inhibitor and that there was no TGFβl related estrogenic effect.

0388TGFβl and MTW9/PL2 Cells. The next study was performed because the MTW9/PL2 cells are the only known estrogen growth responsive rat cell line derived from a hormone responsive carcinogen induced tumor. A similar analysis was done with the MTW9/PL2 rat mammary tumor cells (Fig. 34B). TGFβl reduced cell numbers from 350,000 to 100,000 per dish. This was significant (p &lt; 0.05). However, the presentation of cell number results only tends to exaggerate the effects of TGFβl. When the results were converted to CPD (Fig. 34B, insert), the actual inhibition was 1.5 CPD. This was at most a 25% decrease in growth rate. As shown, there was no estrogen reversal of the TGFβl inhibition with MTW9/PL2 cells. TGFβl and other ER⁺ Cell Lines. The effects of TGFβl at 50 ng/mL ± E₂ were also investigated with the other cell lines used in this study. The MCF-7 A, T47D and ZR-75-1 human breast cancer cells were inhibited by TGFβl (Fig.35A). From these results, and those in Fig.34A, it was clear that the MCF-7 cells were the most sensitive of the ER⁺ human breast cancer lines tested. Irrespective of the line, E₂ had no influence on the TGFβl mediated inhibitions (Fig. 35A). The same experiments were done with the LNCaP cells and the GHA pituitary line (Fig. 35A). They were more sensitive to TGFβl than breast cancer cells. Nonetheless, the TGFβl effects were not reversed by F^. When the cell number decreases presented in Fig. 35A were converted to CPD, it was clear that the TGFβl effects were negligible and that E₂ was of no significant consequence (Fig. 35B). Thus, TGFβl did not substitute for the estrogen reversible inhibitors) in CDE serum with any of the sex steroid sensitive ER⁺ cell lines tested.

0389TGFα and EGF as Substitutes for the Estrogen Reversible Inhibitor in CDE Serum. The EGF family of mitogens and receptors has been linked to breast cancer proliferation, invasion and progression (Dickson RB and Lippman ME (1987) Endocr Rev 8, 29-43; Normanno N et al. (1994) Breast Cancer Res Treat 19, 11-27; Ether SP (1995) JNatl Cancer Inst 81, 964-973; de Jung JS et al. (1998) JPathol 184, 44-52 and 53-57). Most prominent among these polypeptide mitogens has been the EGF analogue, TGFα (Dickson RB and Lippman ME (1987) Endocr Rev 8, 29-43; de Jung JS et al. (1998) JPathol 184, 44-52 and 53-57). Estrogen induced secretion of TGFα is thought to create an autocrine loop that promotes breast cancer cell growth (Dickson RB et al. (1985) Endocrinology 118, 138-142; Dickson RB et al. (1986) Cancer Res 46, 1707-1713; Dickson RB et al. (1986) Science (Wash DC) 232, 1542-1543; Dickson RB and Lippman ME (1987) Endocr Rev 8, 29-43; Derrick R (1988) Cell 54, 593-595; Arrack BA et al. (1990) Cancer Res 50, 299-303; Kenney NJ et al. (1993) J Cell Physiol 156, 497-514; Normanno N et al. (1994) Breast Cancer Res Treat 29, 11-27; Dickson RB et al. (1987) Proc Natl Acad Sci USA 84, 837-841; Salomon DS et al. (1984) Cancer Res 44, 4069-4077; Liu SC et al. (1987) Mol Endocrinol 1, 683-692). TGFα is also thought to potentiate estrogen action in uterus (Nelson KG et al. (1992) Endocrinology 131, 1657-1664) as well as to regulate the EGF receptor in this tissue (DiAugustine RP et al. (1988) Endocrinology 111, 2355-2363; Huet-Hudson YM et al. (1990) Mol Endocrinol 4, 510-523; Mukku VR and Stancel GM (1985) J Biol Chem 260, 9820-9824). The culture conditions described herein offer a new opportunity to test the autocrine growth model under conditions not previously available. Application of the new cell growth assays allowed a direct test to determine if an autocrine/intacrine growth factor loop explains the estrogen reversal of the serum inhibition. EGF and TGFα as Substitutes for E₂. Growth of the MCF-7A, MCF-7K, T47D and ZR- 75-1 cells was measured in D-MEM/F-12 containing increasing concentrations of CDE horse serum with and without exogenous EGF or TGFα. The results with the four cell lines are shown in Figs. 36 A, 36B, 36C, and 36D, respectively. As expected, CDE horse serum was progressively inhibitory at concentrations > 5% (v/v). The addition of growth saturating concentrations (Karey KP and Sirbasku DA (1988; Cancer Res 48. 4083-4092) of EGF or TGFα did not reverse the effects of the serum-borne inhibitor. Li control cultures without added polypeptide mitogens, E₂ completely reversed the serum inhibition. These results again confirm the same conclusion arrived at earlier using an entirely different approach (Karey KP and Sirbasku DA (1988) Cancer Res 48. 4083-4092). Direct evidence for obligatory EGF/TGFα autocrine loops in estrogen responsive cell growth simply has not yet been established. Li fact, there is solid in vivo evidence to challenge EGF/TGFα autocrine loop participation in the action of estrogens (Arteaga CL et al. (1988) Mol Endocrinol 2, 1064-1069).

0390IGF-I, IGF-H and Insulin as Substitutes for Estrogen Action. Lisulin-like growth factors I and II (IGF-I and IGF-IT) promote breast cancer cell growth (Furlanetto RW and DiCarlo JN (1984) Cancer Res 44, 2122-2128; Myal Y et al. (1984) Cancer Res 44, 5486-5490; Dickson RB and Lippman ME (1987) Endocr Rev 8, 29-43; Karey KP and Sirbasku DA (1988) Cancer Res 48, 4083- 4092; Ogasawara M and Sirbasku DA (1988) In Vitro Cell Dev Biol 24, 911-920; Stewart AJ et al. (1990) JBiol Chem 265, 2172-2178). IGF-I related proteins (Huff KK et al. (1986) Cancer Res 46, 4613-4619; Huff KK et al. (1988) Mol Endocrinol 2, 200-208; Dickson RB and Lippman ME (1987) Endocr Rev 8, 29-43; Minute F et al. (1987) Mol Cell Endocrinol 54, 17-184, as well IGF-II (Yee D et al. (1988) Cancer Res 48, 6691-6696; Osborne CK et al. (1989) Mol Endocrinol 3, 1701-1709), are thought of as possible autocrine/paracrine mitogens. Their secretion in response to hormones has been proposed (Dickson RB and Lippman ME (1987) Endocr Rev 8, 29-43; Huff KK et al. (1988) Mol Endocrinol 2, 200-208; Osborne CK et al. (1989) Mol Endocrinol 3, 1701-1709). Insulin itself is likely an endocrine mediator. Li the instant study, it was investigated whether exogenous IGF-I addition to cultures containing CDE-horse serum substituted for the inhibition reversing effects of estrogens with human breast cancer cells. Fig. 37A and 37B show the results with the MCF-7K and MCF-7A cells, respectively. Clearly, 1.0 μg/mL IGF-I did not reverse the serum inhibition. This was true despite the fact that this concentration of added IGF-I was much more than growth saturating (Karey KP and Sirbasku DA (1988) Cancer Res 48, 4083-4092). Duplicate studies with the T47D cells gave the same results (Fig. 37C). It should be noted that IGF-I is active with breast cancer cells even in the presence of serum (Furlanetto RW and DiCarlo JN (1984) Cancer Res 44, 2122-2128; Myal Y et al. (1984) Cancer Res 44, 5486-5490; Osborne CK et al. (1989) Mol Endocrinol 3, 1701-1709; Stewart AJ et al. (1990) JBiol Chem 265, 2172-2178; Cullen KJ et al. (1990) Cancer Res 53, 48-53) that contains specific growth factor binding proteins (Re der M et al. (1980) Endocrinology 107, 1451-1459). Human breast cancer cells also secrete binding proteins for the insulin-like growth factors (Yee D et al. (1991) Breast Cancer Treat Res 18, 3-10). Binding of the insulin-like factors to carrier proteins may attenuate activity (Zapf J et al. (1978) J Gin Invest 63, 1077-1084), have both inhibiting and activating effects (De Mellow JS et al. (1988) Biochem Biophys Res Commun 156, 199-204), or enhance biological action (Elgin R et al. (1987) Proc Natl Acad Sci USA 84, 3254-3258; Blum WF et al. (1989) Endocrinology 125, 766-772). In parallel studies (data not shown), the effects of IGF-II were assayed with the same breast cancer lines under the conditions used with IGF-I. Even at 500 ng/mL, IGF-II did not reverse the inhibitory effects of 10 to 50% (v/v) CDE serum. In another related test, insulin at 10 ng/mL to 10 μg/mL did not reverse the inhibition caused by 50% (v/v) CDE serum. The results with insulin, IGF-I and IGF-II were mutually supportive because these mitogens promote growth via a common receptor (Rechler M et al. (1980) Endocrinology 107, 1451-1459; Karey KP and Sirbasku DA (1988) Cancer Res 48, 4083- 4092; Osborne CK et al. (1989) Mol Endocrinol 3, 1701-1709; Stewart AJ et al. (1990) JBiol Chem 265, 2172-2178). The insulin results were also important in another way. This hormone does not interact with binding proteins and hence their presence in medium will not influence insulin action. These results again confirm the same conclusion arrived at earlier using an entirely different approach (Karey KP and Sirbasku DA (1988; Cancer Res 48. 4083-4092). Direct evidence for obligatory IGF-1/IGF-II autocrine loops in estrogen responsive cell growth simply has not been confirmed yet. Li fact, there is solid in vivo evidence to the challenge IGF-1/IGF-JJ autocrine loop participation in the action of estrogens (Arteaga CL et al. (1989) J Clin Invest 84, 1418-1423).

0391Conceptual Derivations from this Study. These results also have a direct bearing on a number of hypotheses advanced to explain how estrogens cause target tissue cell growth. The development of the new methods herein provided a unique opportunity to reevaluate the most widely cited proposals under consideration. It was concluded that serum contains an inhibitor that effectively blocks ER⁺ and AR⁺ cell growth. Furthermore, physiologic concentrations of sex steroid hormones reverse this inhibition. The results were uniformly the same no matter from which species the cell lines were derived or which species was the source of the serum. Li every case, the effects of the various classes of steroid hormones on the different cell lines were consistent with their known tumor forming/growth properties in vivo or published responses in vitro. These results provide new insights into the following proposed mechanisms.

0392Serum Factor Regulation - Demonstration of Estrogen Responsiveness. The literature describing positive sex steroid hormone growth effects is notably weighted in favor of the use of serum-supplemented cultures. Li fact, a review made of the literature (Briand P and Lykkesfeldt AE (1986) Anticancer Res 6, 85-90; Wiese TE et al. (1992) In Vitro Cell Dev Biol 28A, 595-602) indicates that most past studies have used medium containing &lt; 20% (v/v) steroid hormone depleted serum. Although other investigators have reported estrogenic effects in "serum-free defined culture", these studies actually used conditions that included a prolonged preincubation in the presence of serum (Allegra JC and Lippman ME (1978) Cancer Res 38, 3823-3829; Briand P and Lykkesfeldt AE (1986) Anticancer Res 6, 85-90; Darbre PD et al. (1984) Cancer Res 44, 2790-2793). The results presented in preceding Examples demonstrate clearly that large magnitude effects are readily demonstrable in medium with CDE-serum and that as the CDE-serum concentrations increase to a maximum useable level of 50%, cell growth is inhibited and estrogens invariably reverse these effects. In light of those results, it was clear that the presence of serum, or a factor(s) contained in serum, made possible the demonstration of sex hormone dependent growth in culture.

0393The Endocrine Estromedin Hypothesis - Positive Indirect Control. Li 1978 it was proposed (Sirbasku DA (1978) Proc Natl Acad Sci USA 75, 3786-3790) that growth of estrogen target tissues was not mediated directly by these hormones, but was instead controlled indirectly by steroid inducible circulating growth factors (i.e. endocrine estromedins). Estromedins were proposed to be secreted by target tissues such as uterus, kidney and pituitary, and to act in concert to simultaneously promote the growth of all ER⁺ target tissues (Sirbasku DA (1978) Proc Natl Acad Sci USA 75, 3786-3790; Sirbasku DA (1981) Banbury Report 8, 425-443; Ikeda T et al. (1982) In Vitro 18, 961-979). The estromedin hypothesis arose from the observation that reproducible in vitro direct estrogen mitogenic effects were not identifiable (Sirbasku DA (1978) Proc Natl Acad Sci USA 75, 3786-3790; Sirbasku DA (1981) Banbury Report 8, 425-443; Ikeda T et al. (1982) In Vitro 18, 961- 979). It must be emphasized that the original estromedin hypothesis rested entirely upon the failure to demonstrate large magnitude estrogen mitogenic effects in culture with cell lines confirmed to form steroid hormone responsive tumors in host animals. When estrogen effects were clearly observed with the MTW9/PL2 rat mammary tumor cells in culture, as described herein and reported (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427; Sirbasku DA and Moreno-Cuevas JE (2000) In Vitro Cell Dev Biol 36, 428-446), it was apparent that the endocrine estromedin model required further evaluation. It was reasoned that extension of these results to additional ER⁺ cell lines, including those from other species and diverse target tissues, would either provide important support for the earlier hypothesis or disprove it. In the work disclosed herein, this reassessment has been accomplished. All of the ER⁺ cells tested, as well as one androgen sensitive AR⁺ human cancer line, manifested substantial growth in response to the appropriate steroid hormones in cultures containing inhibiting concentrations of CDE serum. There can be no doubt that steroid hormones act positively to promote target tumor cell growth. The results presented in this report plainly nullify the previous endocrine estromedin model of steroid hormone responsive cell growth. The disproval of the earlier endocrine estromedin model reopened the question of how estrogens and other factors regulate sex steroid responsive growth.

0394The Autocrine and Paracrine Models - Positive Indirect Control. Li the studies described in this Example, it was asked if exogenous growth factors mimic the inhibitor reversing effects of estrogens. The EGF/TGFα and insulin-like families were focused on because of their high biological potencies and physiologic relevance. These growth factors were expected to substitute for steroid hormones based on the autocrine loop mechanisms proposed earlier. Despite this expectation, polypeptide growth factors did not substitute for the estrogens. They were inactive in the presence of the serum-borne inhibitor. Li point of fact, deduction indicates that it makes no practical difference whether the growth factors were autocrine or paracrine in origin. The presence of the serum inhibitor in effect blocks all mitogenic action except that exerted by the steroid hormones. This is a preferred feature of the serum-borne inhibitor(s) disclosed herein, and is further described in Examples which follow, when the use of serum-free defined culture is described. These results also indicate that the search for the regulatory mechanism controlling estrogen dependent growth must seek new directions. Since the estrogenic effects seen in CDE-serum are the largest yet recorded, CDE is the preferred source of the regulator in the cell growth assays.

0395Culture Parallels in vivo Growth Regulation. The results shown in this Example have another important implication. Usually normal in vivo tissues are bathed in growth factor containing fluids. Mitogens within tissues may be of local origin or may be derived from the circulation (Gospodarowitz D and Moran JS (1976) Annu Rev Biochem 45, 531-558; Goustin AS et al. (1986) Cancer Res 46, 1015-1029). If growth factors have unrestricted freedom to stimulate cell proliferation, normal formation and architecture of the tissues would not develop nor could they be maintained. Manifestly, tissue architecture would be disrupted. Li fact, this is one definition of cancer (Sonnenschein C and Soto AM (2000) Mol Carcinog 29, 205-211). The properties of a serum-borne inhibitor that counterbalances unrestricted growth merit serious further consideration with regard to how cancers develop in steroid hormone sensitive tissues. Others researchers have also arrived at this conclusion (Soto AM and Sonnenschein C (1985) J Steroid Biochem 23, 87-94).

0396The Estrocolyone Hypothesis - Negative Indirect Regulation. The estrocolyone model (Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52) is an indirect negative mechanism based on regulation of sex steroid hormone dependent cells via a serum-borne inhibitor. The inhibitor blocks growth promoted by non-steroidal mitogens such as growth factors and diferric transferrin. Sonnenschein and Soto first proposed that estrocolyone acted at the cell surface via specific receptors. The effects of sex steroid hormones were to bind estrocolyone and prevent it from associating with the cells. Only low physiologic concentrations of sex steroid hormones were needed for this function. The special emphasis of this model was that sex steroid hormones did not act through intracellular located DNA binding receptors (i.e. cytosolic or nuclear sites). These intracellular sites had no growth function. Hence, this was an indirect negative mechanism (Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52). The results presented in this disclosure are in agreement with the serum-borne mediator aspect of the estrocolyone hypothesis. There is no doubt that serum from several species contains a steroid hormone reversible inhibitor and that its isolation and molecular characterization will be a major advance with both practical and conceptual applications. With regard to the action site of the steroid hormones, these results differ from the estrocolyone hypothesis as described (Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52). As discussed in the Background of the Invention, the tentative identification of several estrocolyone candidates have been described, and in U.S. Patent Nos. 4,859,585 (Sonnenschein) and 5,135,849 (Soto), the issue of properties was raised again, but with different conclusions than published earlier. The Positive Direct Model — Steroid Hormone Receptor Mediation. The one mechanism most widely accepted regarding steroid hormones and growth involves the nuclear located DNA binding ERα receptor (Gorski J and Hansen JC (1987) Steroids 49, 461-475). Growth is thought to be mediation by specific cytosolic and/or nuclear located receptors that ultimately alter DNA transcription to regulate gene activity. Results from many laboratories support this mechanism (Jensen EV and Jacobson HI (1962) Recent Prog Horm Res 18, 387-414; Gorski J et al. (1968) Recent Prog Horm Res 24, 45-80; Jensen EV et al. (1968) Proc Natl Acad Sci USA 59, 632-638; Jensen EV and DeSombre ER (1973) Science (Wash DC) 182, 126-134; Anderson JN et al. (1974) Endocrinology 95, 174-178; O'Malley BW and Means AR (1974) Science (Wash DC) 183, 610-620; Lippman ME (1977) Cancer Res 37, 1901-1907; Harris J and Gorski J (1978) Endocrinology 103, 240-245; Markaverich BM and Clark JH (1979) Endocrinology 105, 1458-1462; Katzenellenbogen BS (1980) Annu Rev Physiol 42, 17-35; Katzenellenbogen BS (1984) J Steroid Biochem 20, 1033- 1037; Clark JH and Markaverich BM (1983) Pharm Ther 21, 429-453; Darbre P et al. (1983) Cancer Res 43, 349-355; Darbre PD et al. (1984) Cancer Res 44, 2790-2793; Huseby RA et al. (1984) Cancer Res 44, 2654-2659; Gorski J and Hansen JC (1987) Steroids 49, 461-475; Katzenellenbogen BS et al. (1987) Cancer Res 47, 4355-4360; O'Malley BW (1990) Mol Endocrinol 4, 363-369). As also discussed in Example 1, the preferred positive action of estrogens is activation of a new ERγ that saturates/activates at lower steroid concentrations than the ERα or the ERβ.

0397Serum Proteins with Estrocolyone Steroid Binding Characteristics. If the estrocolyone mechanism is in fact correct, one must be able to identify at least one serum protein with very high affinity binding (i.e. K_d picomolar) for sex steroids. There is, however, a major unresolved problem with that hypothesis. Other than sex hormone binding globulin (SHBG), additional high affinity estrogen binding in CDE human serum has not been found. SHBG has K_d of 1.7 x 10^"9 M for E₂ at 37°C (Rosner W and Smith RN (1975) Biochemistry 14, 4813-4820). This affinity does not qualify as the high binding expected of estrocolyone. Also, a search for estrocolyone in human serum only resulted in identification of SHBG (Reny J-C and Soto AM (1992) J Gin Endocrinol Metab 68, 938- 945). No higher affinity binding site/protein was found. The binding of labeled steroid hormones with CDE-horse and CDE-rat serum was studied (results presented in an Example which follows), and ³H-E₂ specific binding at Kd of 20 to 50 nM was found. This is a significant matter because estrogenic effects are demonstrated in this disclosure at 1 to 10 picomolar. As further support for this point, the estrocolyone authors found estrogenic effects at 10 to 30 picomolar E₂ (Soto AM and Sonnenschein C (1985) J Steroid Biochem 23, 87-94; Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52). The lack of correlation between the concentration of steroid that promotes growth and affinity of sex steroids for serum components raises serious questions about this aspect of the estrocolyone hypothesis. These observations also suggest that a very high affinity intracellular ERγ regulates growth.

0398A New Model of Steroid Hormone Responsive Cell Growth. A new model best fits the available data. It brings together aspects of both the direct positive mechanism and indirect negative control. According to this model, regulation of steroid hormone target tumor cell growth is a balance between positive and negative control signals. This balance dictates either growth (i.e. cell division) or quiescence (i.e. cell metabolism and tissue specific function but without cell division). The positive mediators are the steroid hormones acting mediated by a high sensitivity intracellular DNA binding sex steroid receptor that ultimately activates gene expression via intracellular located receptors; whereas negative regulation is exerted by a serum-borne inhibitor that acts at the cell surface. The results disclosed herein support the view that growth is controlled directly by both negative and positive mediators. Li a subsequent Example, this model of negative and positive response control mechanisms is further described and the mediators are shown to be the secretory immunoglobulins acting on cell surface (membrane) receptors.

0399TGFβ and Relevant Inhibition. The results presented further define the molecular properties of the serum-borne inhibitor by eliminating TGFβl as a candidate. This is an important issue because of the well-known effects of TGFβ on normal breast epithelial cells (Hosobuchi M and Stampfer MR (1989) In Vitro Cell Dev Biol 25, 705-713) and ER^" estrogen insensitive breast cancer cells (Arteaga CL et al. (1988) Cancer Res 48, 3898-3904). The results herein continue to confirm a previously unrecognized entity that serves as the estrogen reversible inhibitor in serum. Inhibitors that lack estrogen reversibility can be eliminated from consideration.

0400Discussion of Example 9. From this series of experiments, it can be readily appreciated that any other natural or synthetic protein or other substance can be similarly tested for cancer cell growth inhibiting activity akin to the serum-derived inhibitor in the CDE horse serum. Also, the same XAD™-4 and CDE extraction protocols may also be applied to body fluids and secretions other than serum, and the extracted fluids may be assayed as described for inhibitor activity. Such fluids or secretions include plasma, urine, seminal fluid, milk, colostrum, mucus and stool. An XAD™-4 column is especially suited for preparing a steroid hormone depleted specimen from a small sample of body fluid.

0401Example 10. Serum-free Defined Culture Medium Formulations.

0402Li this Example, formulations of various serum-free defined culture media are discussed. Among other features, the preferred embodiments of the present media provie useful tools for detecting estrogenic effects. During the course investigations leading to the present invention, serum-free defined medium was used to identify IgA, IgM and IgGs as estrogen and/or androgen reversible inhibitors of target tumor cell growth in culture, as demonstrated in subsequent Examples. However, before the full effects of the immunoglobulins could be measured in serum-free defined medium, another issue had to be resolved. The growth of hormone responsive cancer cell types in serum-free medium based on standard preparations of D-MEM/F-12 was not as vigorous as expected. Despite the extensive purification of the water used for cell culture procedures (Sirbasku DA et al. (1991) Biochemistry 30, 295-304; Sirbasku DA et al. (1991) Biochemistry 30, 7466-7477) and careful management of technical issues (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 410-427), it was still apparent that the problem persisted. It was found that thyroid hormone dependent cell growth in culture was being inhibited by a normal component in standard D-MEM/F-12 medium and that a serum-borne factor corrected the problem. This work was done with established rat pituitary tumor cell lines (Tashjian AH Jr (1979) Methods Enzymol 58, 527- 535) in serum-free defined medium (Sirbasku DA et al. (1991) 77, C47-C55; Sirbasku DA et al. (1991; Biochemistry 30, 295-304; Sirbasku DA et al. (1991) Biochemistry 30, 7466-7477). As this work developed, it was recognized that the serum factor was apotransferrin and that its addition to serum-free defined medium permitted observation of thyroid hormone dependent pituitary tumor cell growth (Sirbasku DA et al. (1992) In Vitro Cell Dev Biol 28A, 67-71; Sato H et al. (1992) Mol Cell Endocrinol 83, 239-251). Apotransferrin is a M_r 80,000 bilobular serum protein that binds one Fe (JS) in each lobe, albeit with slightly different affinities (Aisen P and Liebman A (1972) Biochim Biophys Ada 257, 314-323; Chasteen ND (1983) Trends Biochem Sci 8, 272-275; Evans RW and Williams J (1978) Biochem J 173, 543-552). When Fe (JS) saturated, the protein is called diferric transferrin. This form of transferrin is the major iron delivery system for the body (Young SP and Aisen P (1981) Hepatology 1, 114-119; Ciechanover A et al. (1983) J Biol Chem 258, 9681-9689). Because apotransferrin possesses very high affinity for Fe (JS) (i.e. ~10²⁰ at pH 7.4), there is no significant free iron in blood. Considering the extraordinary specificity of apotransferrin for Fe (JS), it was concluded that the presence of the ferric (FeJXi) form of iron in culture medium was deleterious to hormone responsive rat pituitary tumor cell growth (Sato H et al. (1991) In Vitro Cell Dev Biol 21 A, 599-602). Additional work with apotransferrin and other Fe (JS) chelators, along with direct addition of Fe (JS) to culture medium, confirmed that this toxic metal was inhibiting thyroid hormone dependent growth of rat pituitary tumor cells in culture (Eby JE et al. (1992) Anal Biochem 203, 317-325; Eby JE et al. (1992) J Cell Physiol 156, 588-600). In neutralizing studies, the very specific Fe (JS) chelator deferoxamine mesylate (a.k.a. deferoxamine or desferrioxamine) stood out because of its very high affinity for the iron (i.e. ~ 10³⁰'⁶) (Eby JE et al. (1993) J Cell Physiol 156, 588-600) and its lack of toxicity to cells in culture. Its addition to cell culture, at concentrations in small excess of the few μM levels of Fe (JS) in medium, essentially neutralized the toxic metal (Eby JE et al. (1992) Anal Biochem 203, 317-325). Because deferoxamine is a low molecular weight bacterial product, it is relatively inexpensive compared to serum-derived apotransferrin. But without doubt, it is equally effective (Eby JE et al. (1992) Anal Biochem 203, 317-325). Before these studies, deferoxamine had never been used in serum-free defined medium to protect hormone responsive cell growth from the toxic effects of Fe (JS). Prior to the present invention, the broad applicability of that information had not yet been discovered nor had it been discovered that some of the tools developed (e.g. a deferoxamine-Sepharose® affinity matrix) were applicable to estimation of the concentrations of biologically active Fe (JS) in chemical, industrial, environmental and biological samples. Deferoxamine mesylate is a U.S. FDA approved drug used to treat iron overload and iron toxicity in humans. It is marketed by Novartis Pharmaceuticals, East Hanover, NJ, under the trade name DESFERAL®. Deferoxamine mesylate is sold to researchers by Sigma Chemicals (St. Louis, MO).

0403Serum-free Defined Mammalian Cell Culture - Development Background. The use of serum-free defined medium to grow diverse cell types in culture gained national and international recognition with the publication by Hayashi and Sato (Hayashi I and Sato GH (1976) Nature (Lond) 259, 132-134). They demonstrated a breakthrough. The serum supplement commonly used in cell culture medium could be replaceable entirely by mixtures of nutrients and hormones in serum-free medium. This observation was expanded to include cell types from many mammalian tissues (Barnes D and Sato G (1980) Anal Biochem 102, 255-270; Barnes D and Sato G (1980) Cell 22, 649-655; Bottenstein J et al. (1979) Methods Enzymol 58, 94-109; Rizzino A et al. (1979) Nutr Rev 37, 369-378). Further development and application of this technology has been reported (Barnes DW, Sirbasku DA and Sato GH (Volume Editors) (1984) Cell Culture Methods for Molecular Biology and Cell Biology, Volume 1: Methods for Preparation of Media, Supplements, and Substrata for Serum-free Animal Cell Culture; Volume 2: Methods for Serum- free Culture of Cells of the Endocrine System; Volume 3: Methods for Serum-free Culture of Epithelial and Fibroblastic Cells; Volume 4: Methods for Serum-free Culture ofNeuronal and Lymphoid Cells, Allan R. Liss/John Wiley, New York). A national symposium organized and directed by Drs. Gordon Sato, Authur Pardee and David Sirbasku was held at the Cold Spring Harbor Laboratory to address the unfolding technology required for serum-free defined medium growth of cells in culture and to discuss its applications (Sato GH, Pardee AB and Sirbasku DA (1982) Volume Editors, Cold Spring Harbor Conferences on Cell Proliferation, Volume 9, Books A and B, Growth of Cells in Hormonally Defined Media, Cold Spring Harbor, New York).

0404Serum-free Defined Culture - Nutrient Additions. A number of nutrient additions to D-MEM/F-12 are needed to grow the cells used in the presently described studies. The formulations of serum-free defined medium employed are specific optimizations, modifications, or necessary changes of earlier media that have been described (Riss TL and Sirbasku DA (1987) Cancer Res 47, 3776-3782; Danielpour D et al. (1988) In Vitro Cell Dev Biol 24, 42-52; Ogasawara M and Sirbasku DA (1988) In Vitro Cell Dev Biol 24, 911-920; Karey KP and Sirbasku DA (1988) Cancer Res 48, 4083-4092; Riss TL et al. (1988) In Vitro Cell Dev Biol 24, 1099-1106; Riss TL et al. 25, In Vitro Cell Dev Biol 25, 127-135; Riss TL and Sirbasku DA (1989) In Vitro Cell Dev Biol 25, 136-142; Riss TL et al. (1986) J Tissue Culture Methods 10, 133-150; Sirbasku DA et al. (1991) Mol Cell Endocrinol 11, C47-C55; Sirbasku DA et al. (1991) Biochemistry 30, 295-304; Sirbasku DA et al. (1991) Biochemistry 30, 7466-7477; Sato H et al. (1991) In Vitro Cell Dev Biol 21 A, 599-602; Sirbasku DA et al. (1992) In Vitro Cell Dev Biol 28A, 67-71; Sato H et al. (1992) Mol Cell Endocrinol 83, 239-251; Eby JE et al. (1992) Anal Biochem 203, 317-325; Eby JE et al. (1993) J Cell Physiol 156, 588-600; Sirbasku DA and Moreno-Cuevas JE (2000) In vitro Cell Dev Biol 36, 428-446). Serum-free Defined Medium Nutrient Supplements - Bovine Serum Albumin.

0405Bovine serum albumin (BSA) (Sigma Catalog No. A3912) was made by "initial fractionation by heat shock and Fraction V", minimum purity 98% (electrophoresis), according to the supplier. A 50mg/mL stock solution of BSA was prepared in normal saline and was sterilized using 0.2 μm pore membrane filters. Aliquots are stored at -20°C in plastic tubes. As will be discussed below, the "heat shock" step that was used in most albumin preparation methods inactivates the estrogen reversible inhibitor disclosed herein.

0406Serum-free Defined Medium Nutrient Supplements - Linoleic Acid - Albumin (Lin-Alb). This preparation was purchased from Sigma as Linoleic Acid Albumin Conjugate (Catalog No. L8384). The conjugate is supplied as a powder sterilized by irradiation. The fatty acid content is 1% linoleic acid by weight. A stock solution was typically prepared by dissolving the contents of a 500 mg bottle in 10 mL of sterile normal saline to give a final concentration of 50 mg/mL. Aliquots are stored at 4°C in polystyrene tubes. This solution is never frozen. Mammalian cells cannot produce polyunsaturated fatty acids. They must be supplied in a soluble form. Fatty acids are carried physiologically bound to albumin. Serum-free Defined Medium Nutrient Supplements - Ethanolamine (ETN). ETN was purchased from Sigma (Catalog No. A5629) (FW 61). This liquid has a density of 1.0117 grams/mL. Using 0.610 mL in 100 mL of water, a 100 mM stock solution was prepared which was sterilized using the 0.2 um pore membrane filters. The ETN was stored at -20°C in polystyrene tubes. This nutrient is required to sustain phospholipid metabolism required for all membrane biosynthesis.

0407Serum-free Defined Medium Nutrient Supplements - Phosphoethanolamine (PETN). This solid material was purchased as o-phosphoryl-ethanolamine (FW 141) (Sigma Catalog No. P0503). A 10 mM stock of PETN was prepared by dissolving 141 mg in 100 mL of water and sterilizing with 0.2 μm pore membrane filters. Aliquots were stored at -20°C in polystyrene tubes. This component is an adjunct to ETN.

0408Serum-free Defined Medium Nutrient Supplements - Glutamine (GLUT). This essential amino acid was purchased from Sigma (Catalog No. G5763). It is "cell culture tested" according to the manufacturer. Addition of glutamine (FW 146.1) to the culture media is necessary because of its relatively short half-life (i.e. about 80% is lost in 20 days at 35°C). See the Sigma product information for the decay curves at different temperatures and pH. Purchased D-MEM/F- 12 stored in the refrigerator for about three weeks lost most of the original glutamine present. For serum-free applications, additional supplementation is required to sustain growth. For a preparation, 11.7 g was dissolved in 400 mL of water to give 200 mM glutamine. This solution was sterilized using 0.2 μm pore filter membranes. Aliquots are stored at -20°C polystyrene tubes. The final glutamine concentration added to serum-free defined medium is 2 mM. Glutamine is a major metabolite and energy source for cells growing in culture. Serum-free Defined Medium Nutrient Supplements - Reduced Glutathione (GSH).

0409Crystalline reduced glutathione (FW 307.3) was purchased from Sigma (Catalog No. G4251). A stock of 40 mg/mL was prepared by dissolving 400 mg in 10 mL of water. This stock was very quickly sterilized with a 0.2 μm pore filter unit. Aliquots were quickly stored at -20°C in polystyrene tubes. According to Sigma technical service, this sulfhydryl (-SH) compound is unstable in aqueous solutions, including tissue culture medium, and is rapidly converted to the oxidized GS-SG form by exposure to air. Addition every two to four days to the culture medium may be required for reducing agent requiring cells. Another reducing agent that also is effective is mercaptoethanol. It is more stable and often effective at lower concentrations than GSH. Reducing agents act as "scavengers" of free radicals generated by the oxygen atmosphere of cell culture.

0410Serum-free Defined Medium Nutrient Supplements - Selenium (Se). A powder of sodium selenite (100 mg/vial) is obtained from Collaborative Research or Sigma (Catalog No. S5261). It has been sterilized by irradiation. The contents of a single vial are dissolved in 100 mL of sterile water to give final stock of 1.0 mg/mL. This preparation should not be filter sterilized because Se binds to filters. The final volume was diluted to 100 mL with sterile saline. Aliquots are stored at - 20°C in polystyrene tubes. Selenium is an important cofactor for enzyme systems that protect the cells from oxidation effects.

0411Serum-free Defined Medium Nutrient Supplements - Diferric Transferrin (2FeTf). Lon Fe (JS) saturated (98%) human transferrin (diferric transferrin) was purchased from Collaborative Research (Catalog No. 40304) or Sigma (Catalog No. T3309) as bottles containing 1 gram of red colored powder. The contents of one bottle are dissolved in 100 mL of normal saline. This red colored solution is sterilized using 0.2 μM pore membrane filters. This stock is 10 mg/mL. Aliquots are stored at - 20°C in polystyrene tubes. All growing cells require diferric transferrin as a source of iron for a great many metabolic processes. Serum-free Defined Medium Growth Factor Supplements - Epidermal Growth

0412Factor (EGF). EGF prepared from mouse submaxillary gland (tissue culture grade) was purchased from Collaborative Research (Catalog No. 40001) as 100 μg in a sterile vial or from Sigma (Catalog No. E4127). The original vials are stored at 4°C according to the manufacturer's instructions. To prepare a stock solution, 5.0 mL of sterile saline was added to a vial to yield a 20 μg/mL EGF solution. Aliquots are stored frozen at -20°C polystyrene tubes. Repeated freeze- thaw must be avoided. This growth factor is useful because of its very broad cell specificity range. Serum-free Defined Medium Growth Factor Supplements - Acidic Fibroblast

0413Growth Factor (aFGF). Acidic FGF is purchased from Sigma (Catalog No. F5542). It is the human recombinant product from E. coli. This product has very specific handling requirements. , It is provided sterilized in 25 μg vials lyophilized from PBS containing 1.25 mg of BSA. The contents of each vial are reconstituted in 25 mL of sterile PBS containing 1.0 mg/mL of BSA and 10 μg/mL of heparin. Filtration of this product at this concentration must absolutely be avoided. This solution is stored at -20°C in polystyrene tubes. The solutions of aFGF definitely cannot be freeze-thawed more than twice. This growth factor is highly labile. Careless handling will result in problems. Keratinocyte growth factor (KGF) can substitute for aFGF. The fibroblast growth factor family is very important in growth of urogenitial tissues including prostate. Serum-free Defined Medium Growth Factor Supplements - Heparin. Heparin is used to stabilize FGF in cell culture (Gospodarowitz D and Cheng J (1986) J Cell Physiol 128, 475-484). Heparin is obtained from Sigma (Catalog No. H3149) as the sodium salt, Grade 1-A, from porcine intestinal mucosa. A solution of 1.0 mg/mL is made in saline and sterilized with 0.2 μm pore membrane filters. An aliquot of 250μL is added to the 25 mL of aFGF reconstitution solution used above. Sterile heparin is stored at 4°C.

0414Serum-free Defined Medium Adhesion Protein Supplement - Fibronectin (Fbn). Human plasma derived Fbn can be purchased from many commercial sources. Bovine Fbn is also available and is effective. Fbn is prepared from units of fresh human plasma (unfrozen) or fresh bovine (unfrozen) plasma by two methods (Retta SF et al. (1999) Methods in Molecular Biology 96, 119-124; Smith RL and Griffin CA (1985) Thrombosis Res 37, 91-101). Purity is evaluated by SDS-PAGE with Coomassie Brilliant Blue staining or silver staining (Pierce Chemicals ® kits). Adhesion activity is confirmed with cells in serum-free defined medium. Vitronectin can substitute for fibronectin.

0415Serum-free Defined Medium Iron (Fe (HI) Chelator Supplements - Deferoxamine mesylate (DFX). DFX (FW 656.8) is purchased from Sigma (Catalog No. D9533). The stock solution is made at 10 mM by adding 131 mg to 20 mL of highly purified water as described above. The solution is sterilized by filtration with 0.2μM pore membranes. Aliquots are stored at -20°C in polystyrene tubes.

0416Serum-free Defined Medium Iron (Fe (HI) Chelator Supplements - Apotransferrin (apoTf). Human serum ApoTf can be purchased from Sigma (Catalog No. T4382). It is minimum 98% iron-free. ApoTf is also prepared as described previously (Sirbasku DA et al. (1991) Biochemistry 30, 295-304; Sirbasku DA et al. (1991) Biochemistry 30, 7466-7477). ApoTf is prepared by dialysis against citrate buffer pH 5.0- 5.5 with 1 μg/mL DFX present to chelate > 98% of the iron. Handling and storage were as described for diferric transferrin but with great care to avoid contact with iron sources. Serum-free Defined Medium Nutrient Supplements - Bovine Insulin (INS). This hormone was purchased from either of two sources. From Gibco-BRL it is Insulin, Bovine Zinc Crystals for Cell Culture Applications (Catalog No. 18125-039). It was also obtained from Collaborative Research (Catalog No. 40305) and stored at 4°C, according to that manufacturer's recommendation. Gibco-BRL recommends solid insulin storage at -5°C to 20°C. A stock of 10 mg/mL in 0.01 N HC1 was prepared by adding 250 mg of insulin to 25 mL of the acid. The HC1 was made by adding 172 μL of concentrated (11.6 N) HC1 to 100 mL of water. The final stock solution of lOmg/mL of insulin is filter sterilized using 0.2 μm pore diameter membranes. Aliquots are stored at 4°C in polystyrene tubes. Care was taken not to freeze-thaw the aliquots of stock solution. Insulin is a very broad range cell growth-stimulating factor as well as a regulator of specific metabolic processes.

0417Serum-free Defined Medium Nutrient Supplements - Thyroid Hormones. The preferred thyroid hormone is T₃ (3', 5-Triiodothyronine, FW 673, purchased from Sigma as Catalog No. T2752). It is stored desiccated at -20 ° C. To prepare stocks, 0.5 N NaOH was made by addition of 20 grams of pellets to one liter of water. Then, 67.3 mg of T₃ was added. After dissolving the T₃ with stirring for a few minutes, 25 mL of this stock was diluted up to 250 mL with water, for a final concentration of 0.05 N NaOH. This dilution was sterilized using the 0.2 μm pore diameter filter. At this point, the final stock for storage was 10 μM T₃. Aliquots of this final stock are stored in polystyrene tubes at -20°C. The second thyroid hormone, thyroxin (T₄, sodium salt, pentahydrate FW 888.9), is prepared by the same procedure. For this stock solution, 88.9 mg of T₄ are used. T is purchased from Sigma (Catalog No. T2501). T is used at 10 to 20 times higher concentrations than T₃. Care is taken not to freeze-thaw these preparations. Thyroid hormones have a very broad range of biological effects on metabolism and growth. Many cells in culture require these for growth.

0418Compositions of Serum-free Defined Media. TABLE 7 presents the formulations of the preferred serum-free defined media developed for use in detecting high-level steroid hormone reversible inhibition by steroid hormone-stripped serum fractions and by purified inhibitors in serum-free cell growth assays. As indicated in the footnotes to the table, when a particular component is included in one of the formulations, the concentration that provides a suitable cell growth medium can fall within the indicated range. TABLE 7

0419Composition of Serum-free Defined Media Based on Standard Gibco-BRL D-MEM/F-12

0420<img file="EP1328293A2_D0007.tif" />

0421When a component is added, the following are the effective concentration ranges used:

0422^I INS range 100 ng/mL to 10 μg/mL ² EGF range 1 ng mL to 50 ng mL

0423³ aFGF range 0.2 ng/mL to 20 ng/mL T₃ range 0.3 nM to 10 nM ⁵2FeTf range 2 μg/mL to 50 μg/mL ⁶ ETN range 5 μM to 100 μM ⁷ PETN range 5 μM to 50 μM ⁸ BSA range 0.2 mg/mL to 5.0 mg/mL ⁹ Lin-Alb range 50 μg/mL to 500 μg/mL ^lβ Se range 5 ng/mL to 20 ng/mL

0424^II GSH range 1 μg/mL to 50 μg/mL ¹² Glut range 0.5 mM to 2.0 mM

0425¹³ Heparin range 1 μg/mL to 10 μg/mL ¹⁴ DFX range 2 μM to 20 μM

0426¹⁵ Fbn range 15 μg to 50 μg per 35-mm diameter dish

0427Serum-free Media Variations. Standard phenol red-containing Gibco-BRL D-MEM/F- 12 is a preferred basal medium to which the defined media components are added. It contains 0.6 mM to 1.0 M CaCl₂. D-MEM/F-12 can be purchased from Gibco-BRL in the liquid form or can be prepared from the powder formulation using only highly purified water. Alternatively, another suitable basal medium could be used as long as it provides at least the required minimum amounts of necessary nutrients, vitamins and minerals to maintain cell viability of the desired cell line. The calcium concentration range preferred is 0.6 to 10 mM. Calcium stabilizes the inhibitor in cell culture without impairing cell growth. The human breast cancer cell medium, DDM-2MF, was a modification of the original DDM-2 medium (Danielpour D et al. (1988) In Vitro Cell Dev Biol 24, 42-52) and MOM-1 (Ogasawara M and Sirbasku DA (1988) In Vitro Cell Dev Biol 24, 911-920) and contained modified hormone concentrations, deferoxamine (DFX) and fibronectin. Aqueous salt solutions such as tissue culture medium contain hydrolytic polymeric forms of Fe (JS) (Spiro TG et al. (1966) JAm Chem Soc 88, 2721-2726). DFX binds this form of Fe (JS) with very high affinity (Schubert J (1964) In; Iron Metabolism: The Chemical Basis ofChelation, Springer, Berlin, pp 466- 498). If not removed, Fe (JS) inhibits hormone-responsive growth in serum-free defined medium (Sirbasku DA et al. (1991) Mol Cell Endocrinol 11, C47-C55; Sato H et al. (1992) Mol Cell Endocrinol 83, 239-251; Eby JE et al. (1993) J Cell Physiol 156, 588-600; Eby JE et al. (1992) Anal Biochem 203, 317-325). Fibronectin was used with DDM-2MF to promote cell attachment. The 35- mm diameter assay dishes were pre-coated by incubation with the designated amount of fibronectin (TABLE 7) for 16 to 48 hours at 37°C in 2.0 mL of D-MEM F-12. CAPM human prostatic cancer cell medium was developed to support the growth of tumor cells from this tissue. The composition of CAPM is described in TABLE 7. CAPM also supports the growth of the H301 Syrian hamster kidney tumor cells. DDM-2A, which is a modified form of DDM-2 (Danielpour D et al. (1988) In Vitro Cell Dev Biol 24, 42-52), was preferred for growing MTW9/PL2 cells. PCM-9 defined medium was developed for growing the rat pituitary cell lines. This medium differs from previous PCM formulations (Sirbasku DA et al. (1991) Mol Cell Endocrinol 11, C47-C55; Sato H et al. (1992) Mol Cell Endocrinol 83, 239-251; Eby JE et al. (1993) J Cell Physiol 156, 588-600; Eby JE et al. (1992) Anal Biochem 203, 317-325) in that DFX was substituted for apotransferrin and the triiodothyronine concentration was increased to 1.0 nM. Although DFX and apotransferrin (2 to 50 μg/mL) are the preferred chelators based on their very high specificity and affinities for Fe (JS), EDTA at 1 to 10 μM or sodium citrate at 10 to 1000 μM also effectively neutralize the cytotoxic effects of Fe (JS) (Eby JE et al. (1993) J Cell Physiol 156, 588-600). Ascorbic acid (vitamin C) also chelates Fe (JS), but is used less often because it is unstable in cell culture medium at 37°C in an oxygen environment in the presence of salts and metals in the medium. Also, at concentrations of 50 to 100 μg mL, apo-ovofransferrin and apo-lactoferrin also were effective Fe ( T) chelators in serum- free defined medium (Eby JE et al. (1993) J Cell Physiol 156, 588-600). Although EGF, aFGF and insulin are the preferred growth factors, several other human recombinant proteins are effective. They have either been purchased or obtained as gifts from Gibco-BRL, Sigma or EVICERA Bioproducts. Lisulin-like growth factors I and II (IGF-I and IGF-II) can be used to replace insulin, fransforming growth factor α (TGFα) replaces EGF, TGFβ as an inhibitory supplement, and basic fibroblast growth factor (bFGF) partially replaces aFGF. Insulin can be used to replaced IGF-I and IGF-JJ. All of these protein growth factors are dissolved under sterile conditions according to manufacturers' instructions and stored as indicated.

0428Discussion of Example 10. The preferred serum-free media described above provide an ideal scenario for the study of growth responses of hormone responsive cancers without the myriad of potential interactions accompanying the presence of serum with its 5000+ proteins and other compounds. The formulations presented permit dissection of growth into its individual parts caused by different stimulators. When of interest, a combination of a few factors can be investigated to achieve an understanding of growth promoter/inhibitor interactions (i.e. cross-talk). This is exceptionally difficult to achieve in the presence of full serum. The serum-free medium described herein provided a tool for the assessment of growth inhibitor(s) isolated from CDE-horse serum, whose actions are reversed by sex-steroid hormones, as mentioned at the beginning of this Example and described in more detail in subsequent Examples. The preferred serum-free media of the present invention raise hope for the provision of new insight that could help to clarify the mechanisms involved in the control of breast, prostatic and other mucosal cancers under conditions not previously available.

0429Moreover, because of widespread concern today about possible contamination of commercial animal sera by disease causing agents such as bovine spongiform encephalopathy ("mad cow disease"), there is a great need for serum-free cell culture media that can support a variety of cell types. The new media compositions fill that need. The new serum-free media can be used not only for assays but also for large scale testing purposes and industrial uses such as cell culture production of a desirable protein. For example, an antigen for vaccine production, or a monoclonal antibody can be prepared without fear of contamination by a serum-derived infectious agent. They are also useful for producing a quantity of virus for vaccine manufacture or for producing recombinant viruses for gene therapy. Basic cell culture methods for producing quantities of proteins or viruses are well known in the art and have been described in the literature.

0430Example 11. Serum-free Defined Medium that Supports Hormone Sensitive and Autonomous Cancer Cell Growth In this Example, it is shown that media derived according to the present methods are effective for supporting hormone sensitive and autonomous cancer cell growth.

0431Selection of Models to Study Hormone Dependence and Autonomy in Serum-free Defined Culture Media. One goal was to develop serum-free defined media that can be used to directly compare negative serum factor regulation with steroid hormone responsive and steroid hormone autonomous cancers of the same tissue. That meant establishing a medium that supported the growth of both cell types. As models, human prostatic carcinoma and human breast carcinoma cells were chosen because responsive and autonomous (unresponsive) cell lines are currently available for both types of cancers. Furthermore, as discussed above, these cancers have many common characteristics including their tendency to pass from steroid hormone receptor positive to steroid hormone receptor negative in a process called tumor progression. During the course of development of such defined media, one observation was made consistently: breast cancer cells that were ER⁺ (i.e. estrogen sensitive) and prostate cancer cells that were AR⁺ (i.e. androgen sensitive) grew less well in defined medium based on standard D-MEM/F12 than in defined medium based on "low-Fe" D-MEM/F12. The results of an example with T47D cells in DDM-2MF are shown in Fig. 38. The example with LNCaP cells in CAPM is shown in Fig. 39. Another example is the thyroid hormone responsive MDCK kidney tubule epithelial cells in CAPM as shown in Fig. 40. Standard D-MEM/F-12 contains both ferric nitrate and ferrous sulfate as nutrient additions. When purchased without these salts, the medium was designated "low-Fe" D-MEM/F-12. The iron concentrations in standard and "low-Fe" D-MEM/F-12 were 1.0 μM and 0.15 μM, respectively (Eby JE et al (1992) Anal Biochem 203, 317-325). Even in "low-Fe" medium, iron is present as a contaminant in the chemicals used to make the formulation, the 2.2 g/L NaHC0₃ added as a metabolic requirement and buffer, and the 15 mM HEPES buffer necessary for stabilizing the pH under serum-free conditions (Eby JE et al (1992) Anal Biochem 203, 317-325). It is noteworthy that as low as 1.0 μM Fe (JS) inhibits epithelial cell growth completely within five to seven days. Li another test the thyroid hormone responsive human HT-29 colonic carcinoma cells in CAPM also grew better in "low-Fe" than standard D-MEM/F-12 (data not shown). This indicates that restriction of Fe (JS) in culture medium will have implications even beyond sex steroid hormone dependent cells.

0432Modifications of the Usual Growth Assays for Experiments in "low-Fe" Medium versus "Standard" Medium. Specific modifications of the customary cell growth assays were required for assays done under iron-restricted conditions. For example, the 35-mm assay dishes were incubated for 16 to 24 hours prior with 20 to 25 μg of fibronectin in 2 mL of "low-Fe" D-MEM-F12 medium. Serum-free components were added to "low-Fe" D-MEM/F-12 at double the concentrations needed (2X) or to "standard" D-MEM/F-12 at (2X) as the experiments dictated. Each assay dish received 1.0 mL of this solution. Next, the cells to be used in the assays were washed three times in either "low-Fe" medium or "standard" medium depending upon the experimental protocol. These washes were done with the same care as described above in General Materials and Methods. Each dish received 1.0 mL of cells in the appropriate medium. At this point, the components final concentrations were (IX) as summarized in TABLE 7. Also, TABLE 7 describes medium containing deferoxamine as the Fe (JS) chelator. Although less preferred, due in part to cost considerations, specificity, and affinity for Fe (JS), as noted above, apotransferrin is also effective, especially at the preferred apotransferrin concentration of 50 μg/mL. When apotransferrin binds Fe (JS), it is converted to one of three forms of ferric transferrin (Eby JE et al (1992) Anal Biochem 203, 317-325). These become additional support for cell growth in defined medium, thereby converting a toxic substance to a useable nutrient.

0433Growth in Serum-free Defined Medium versus D-MEM/F-12 with 10% (v/v) Fetal Bovine Serum. To demonstrate the usefulness of the formulations in TABLE 7, cell growth was compared in serum-free defined medium + steroid hormone versus growth supported by fetal bovine serum. It is generally accepted that fetal bovine serum represents one of the most effective sera for tissue culture. As an example, growth of the LNCaP cells was compared in CAPM ± DHT versus growth in 10% (v/v) fetal bovine serum (Fig. 41). CAPM plus 10 nM DHT supported growth at about 80-90% of the rate of fetal bovine serum. Growth promoted by 10% fetal bovine serum obtained from conventional commercial sources reached 6.57 (± 0.48) CPD or, a 96-fold increase on cell number in 12 days. By day 12, cell densities in CAPM nearly equaled those in serum. Growth promoted by the serum-free medium reached 6.22 (± 0.35) CPD or 84-fold increase. CAPM was able to support LNCaP growth even in the absence of sex-steroid hormones. Maximum growth obtained without sex-steroid hormones was of 5.35 (± 0.12) CPD or a 49-fold increase. The androgenic effect is therefore marginal, with differences of less than one CPD between the presence and absence of DHT. Also shown, the cells did not grow in D-MEM/F-12 without any additions (Fig. 41). Similar studies were done with other cell lines to determine growth rates versus serum and to establish the periods for single time assays (e.g. 1, 10, 12 or 14 days). Fig. 42 shows the same analysis with DU145 and PC3 cells in CAPM and in D-MEM/F-12 with 10% fetal bovine serum. As the cell number data show, growth was logarithmic. After 12 days, growth in the serum-free medium was identical to that in 10% fetal bovine serum for both cell lines. Growth of PC3 in 10% serum reached 6.98 (± 0.71) CPD or a 112-fold increase in cell number versus 6.97 (± 0.44) CPD or the same fold increase for cell numbers in serum-free medium. Growth of DU145 in 10% fetal bovine serum was 6.71(± 0.58) CPD versus 6.73 (± 0.18) CPD in serum-free conditions. The results in Figs. 41 and 42 demonstrate by example that the serum-free defined media in TABLE 7 are effective with both hormone sensitive and hormone autonomous cells.

0434Determination of Component Concentrations and the Requirement for a Fe (HI) Chelator. The optimum concentration of each single component was determined by dose-response analysis in the presence of other components. The technology used to establish early forms of serum-free defined media has been described (Danielpour D et al. (1988) In Vitro Cell Dev Biol 24, 42-52; Ogasawara M and Sirbasku DA (1988) In Vitro Cell Dev Biol 24, 911-920). An example of this process is shown in Fig. 43 with LNCaP cells. Dose-response effects of bovine serum albumin, apotransferrin, T₃, ethanolamine, selenium, and EGF are shown. The results show clearly that the addition of the iron chelator apotransferrin was required for cell growth. After deteπnining optimum concentrations for each component, the contribution of each to the total was assessed by another assay. Individual components were deleted one at a time. As an example, the three most widely used prostatic carcinoma cell lines were compared (i.e. LNCaP, PC3 and DU145) in CAPM that contained deferoxamine in place of apotransferrin (Fig. 44). The deletions were done ± DHT. The first and most striking result was the major differences between the growth requirements of the DHT sensitive LNCaP cells and those of the autonomous DU145 and PC3. Only the deletion of diferric transferrin substantially prevented the growth of autonomous cells. Also, it was clear that deletion of deferoxamine had only a small (i.e.&lt; 20%) effect on growth of the DU145 and PC3 cells. The DU145 and PC3 cell lines also were T₃, insulin, EGF, fibronectin and deferoxamine independent. As expected ± DHT had no significant effect on DU145 or PC3. By contrast, LNCaP growth was significantly (p &lt; 0.01) reduced or arrested completely by deletion of fibronectin, T3, diferric - transferrin or deferoxamine. LNCaP growth also was inhibited by deletion of EGF or insulin, but these effects were pronounced only in the absence of DHT.

0435Discussion of Example 11. The media described in TABLE 7 were optimized for the specific cell types designated. Additionally, they were optimized to permit direct comparison of the growth properties of ER⁺ and AR⁺ steroid hormone sensitive tumor cell lines to their ER^" and AR^" steroid hormone insensitive (also called autonomous) counterparts. This careful optimization was done originally to study rat mammary tumor cells of both types in DDM-2A defined media. The appropriate cell lines for this approach have been developed from the MTW9/PL2 population and described (Danielpour D and Sirbasku DA (1984) In Vitro 20, 975-980). The medium DDM-2MF has been developed for the same purpose only for comparisons of ER⁺ and ER^" forms of these cancers. TABLE 1 lists the most important ER⁺ human breast cancer cell lines in use today. Li addition a number of other ER^" human breast cancer cells lines have been evaluated. They are the MDA-MB-231 (Cailleau R et al. (1974) JNatl Cancer Inst 53, 661-674), BT-20 (Lasfargues EY and Ozzello L (1958) J Natl Cancer Inst 21, 1131-1147), Hs0578T (Hackett AJ et al. (1977) J Natl Cancer Inst 58, 1795-1806), MDA-MD-330 (Cailleau R et al. (1978) In Vitro 14, 911-915), and the myoepithehal HBL-100 (Gaffiiey EV (1982) Cell Tissue Res 227, 563-568). The demonstration of ER^" status of these lines has been described (Reddel RR et al. (1985) Cancer Res 45, 1525-1531). With regard to human prostatic cancer, the only reliable androgen responsive cell line available today is the LNCaP (TABLE 1). Another, the ALNA-41, has been described as androgen growth responsive (Νakhla AM and Rosner W (1994) Steroids 59, 586-589). However, as shown in subsequent Examples, this line is autonomous by the criterion of lack of DHT effects in CDE-horse serum. Two other human prostate cancer cell lines are commonly used as autonomous examples. These lines are the DU145 (Stone KR et al. (1978) Int J Cancer 21, 274-281) and the PC3 (Kaighn ME et al. (1979) Invest Urol 17, 16-23). Previously, there was a defined medium established for PC3 cells (Kaighn ME et al. (1981) Proc Natl Acad Sci USA 78, 5673-5676). This medium was evaluated and did not support LΝCaP cell growth. However, others have reported "serum-free" media that was stated to be effective with LΝCaP, DU145, PC3 and ALVA-31 cells (Hedlund TE and Miller GJ (1994) The Prostate 24, 221-228). The problem was this medium was not serum-free nor was it defined. The experiments began with cells plated into 5% serum and then preceded to use a serum fraction called fetuin to support growth. Fetuin is a complex undefined mixture of > 4% of the proteins in serum. Under those conditions, an accurate analysis of hormonal and growth factor effects cannot be done satisfactorily. The completely serum-free CAPM in TABLE 7 supports the growth of all of these prostate cell lines. Li addition, CAPM has been applied to the ER⁺ estrogen growth stimulated H301 Syrian hamster kidney cells (Sirbasku DA and Moreno JE (2000) In Vitro Cell Dev Biol 36, 428-446) and its autonomous derivative cell line A195. As has been reviewed (Evans RM (1988) Science (Wash DC) 240, 889-895), steroid hormones and thyroid hormones belong to the same superfamily of receptors. Both are important in growth. Therefore, it was expected that some tissues might be thyroid hormone positive regulated, while others might be positive regulated by steroid hormones. CAPM has also been applied to the study of thyroid hormone reversal of purified inhibitors with the human colonic carcinoma cell line HT-29. Similar use has been made of CAPM with the MDCK dog kidney tubule cell line (Leighton J et al. Science (Wash DC) 158, 472-473). CAPM replaces a different defined medium prepared for MDCK cells (Taub M et al. (1979) Proc Natl Acad Sci USA 16, 3338-3342). It is likely that the prostaglandins in that earlier medium interfere with the action of the thyroid hormones. Li any case, that medium was not useful for demonstration of thyroid hormone reversal of purified MDCK cell growth inhibitors. All of these observations support the view that a series of uniquely optimized media have been formulated to define the growth requirements of epithelial cells from several of the very prominent cancers of humans. Furthermore, the technology developed promises application to the optimization of growth of other types of epithelial cells from a variety of target tissues.

0436Example 12. Differential Effects of Fe (DJ) on the Growth of Hormone Responsive and Autonomous Human Breast and Human Prostate Cancer Cells

0437This Example demonstrates that iron has an inhibiting effect on steroid responsive cell growth, independent of the above-described immunoglobulin effects, and which is distinguishable from its effect on autonomous cells.

0438Approaches to Demonstration of Iron Toxicity. The fact that standard D-MEM/F-12 contains sufficient Fe (JS) to inhibit cell growth, led to the next series of studies. Other approaches were used to further demonstrate the deleterious effects of Fe (JS) on hormone responsive tumor cell growth. To add Fe (JS) to culture medium, it must be in a soluble form. Ferric ammonium citrate was selected for use. However, ferric ammonium sulfate is also effective. Ferric ammonium citrate is a mixture that contains 16.6% of ferric iron by weight. The amount of mixture added to each dish was adjusted to achieve the desired Fe (JS) concentrations. Due to the light sensitivity of the mixture, the solutions were prepared fresh daily and the experiments carried out under restricted light conditions. Also, the mixture was prepared in water. Buffers without phosphate may be used, but they are generally less effective. The ferric mixtures and the iron chelators EDTA, deferoxamine mesylate and sodium citrate were purchased from Sigma.

0439Iron Toxicity with Human ER⁺ Breast Cancer Cells. Li the first experiments, two ER⁺ cell lines were evaluated for Fe (JS) sensitivity in DDM-2MF defined medium prepared with 10 μg/mL apotransferrin in place of the deferoxamine shown in TABLE 7. The effect of addition of ferric ammonium citrate on MCF-7A growth ± E. at 10 days is shown in Fig. 45. Either with or without the steroid, Fe (JS) was completely inhibitory at 10 μM. There were no viable cells in the dishes at > 10 μM. The EI₅₀ of Fe (JS) with MCF-7A cells was 5 to 7 μM. A similar analysis with T47D cells in DDM-2MF with 10 μg/mL apotransferrin instead of deferoxamine showed complete inhibition at 10 days with 2 μM Fe (JS) (Fig.46). At > 2 μM there were no viable cells in the dishes either with or without E>. The EI₅₀ of FE (JS) with T47D cells was 1 μM.

0440Iron Toxicity with AR⁺ and AR^" Human Prostate Cancer Cell Lines. The effect of Fe (JS) on AR⁺ LNCaP cell growth was assessed in CAPM defined medium in which apotransferrin (500 nM) was substituted for deferoxamine, and the-results are shown in Fig.47. Clearly, 10 μ M Fe (JS) arrested growth to seed density levels (t.e. 12,000 cells per dish) in a 12-day assay. The EI₅₀ for LNCaP cells was 5 μM. In another experiment in CAPM, the effects of ferric ammonium citrate were evaluated with AR⁺ LNCaP cells and AR^" PC3 and DU145 cells (Fig. 48). Again, Fe (JS) inhibited LNCaP cells to seed densities levels by 8 to 10 μM. However, effects on the androgen autonomous PC3 and DU145 cells were markedly less (Fig. 48). Reductions of 10 to 30% in cell number for PC3 and DU145, respectively, were observed in 10 μM Fe (JS). The inhibitory effects of Fe (JS) on the androgen independent PC3, DU145 and ALVA-41 cells were variable, and never as marked as with the steroid hormone responsive LNCaP cells. The insert in Fig. 48 shows a correlation between hormone responsiveness and Fe (JS) effects. The results show a correlation between iron effects and thyroid hormone responsiveness. LNCaP cells are T₃ responsive whereas PC3 andDU145 arenot.

0441Reversal of Fe (HI) Inhibition by Iron Chelators. The inhibitory/cytotoxic effects of Fe (JS) were reversible by the addition of iron chelators. Those studied were selected based on data showing their relative affinities and specificities for Fe (JS) (Schubert J (1963) In: Iron Metabolism, Gross F, ed, Springer-Verlag, Berlin, pp 466-496). Deferoxamine is most specific and has the highest affinity for Fe (III). Citrate is next most effective. EDTA is not as effective nor is it as specific as the first two chelators. Li experiments with T47D cells, the deferoxamine usually present in the DDM-2MF medium was removed and an additional 1.5 μM Fe (JS) added to ensure complete inhibition of the cells. Fig.49 shows the relative effects of addition of these three chelators to T47D serum-free defined medium cultures. The order of effectiveness was as expected from the affinities and specificities of these chelators. Clearly, addition of Fe (JS) chelators restored growth. Fig. 50 shows a similar study with LNCaP cells in CAPM defined medium from which the deferoxamine also was removed and 1.5 μM Fe (JS) added. It was clear that chelation of the Fe (JS) restored growth. It should be noted that this conclusion is reasonable based on the fact that deferoxamine has near absolute specificity for Fe (JS). Concentrations as low as 0.5 μM of deferoxamine were sufficient to induce 3.5 CPD with LNCaP cells. Maximum growth with this chelator (5.81 CPD) was obtained at 10 μM. Citrate and EDTA were also effective growth stimulators of LNCaP cells incubated at high iron concentrations. Maximum growth was obtained with the addition of 500 μM and 10 μM respectively. The growth induction achieved with EDTA is lower than with citrate or deferoxamine. This probably could be explained by the fact that EDTA is a less discriminatory chelator, and essential metals other than iron were affected. Concentrations of the chelators higher than the ones showed in the Figs. 49 and 50 were associated with cell damage and death. In particular, chelation of calcium by citrate and EDTA will cause cell death in culture. As controls, stimulation by chelators was prevented by resupply of Fe (JS) (data not shown).

0442Correlation Between Hormone Autonomy and Lack of Iron Effects. Li the next series of studies, data was sought supporting the concept that loss of steroid hormone dependence correlates positively with loss of Fe (in) effects. As shown in Fig. 39, LNCaP cells grew better in 'low-Fe" serum-free defined medium than in defined medium based on "standard" D-MEM/F-12. This difference was also evaluated with the androgen insensitive DU145 (Fig. 51) and PC3 (Fig. 52) cells. The results were clear. The autonomous lines grew equally well in CAPM based on both types of D-MEM/F-12. The presence of the higher Fe (JS) level in CAPM based on standard D- MEM/F-12 had no effect. To confirm that these cell lines were androgen autonomous as defined by the loss of steroid and inhibitor growth regulation in CDE-serum, the next studies were done. DU145 cells showed no inhibition of growth in 50% CDE-serum (Fig. 53). There was no androgenic effect whatsoever. A similar assay with PC3 cells showed essentially the same results (Fig. 54). There was no inhibition even in 50% CDE-horse serum, and no androgenic effect. Additionally, ALVA-41 cells are not iron sensitive (results not shown), and also are not sensitive to the serum-bome inhibitor (Fig. 55).

0443Discussion of Example 12. Together with the studies presented above, it appears that AR⁺ cells are sensitive to the serum-borne inhibitor, sensitive to the positive effects of steroid hormone and sensitive to Fe (JS) inhibition. Li contrast, the DU145 and PC3 cells are insensitive to the serum-borne inhibitor, insensitive to the positive effects of androgen, and insensitive to Fe (JS). The results presented in this example continue to demonstrate the requirement for the action of a serum- borne mediator to demonstrate steroid hormone responsive cell growth in culture. The use of CDE- serum was essential for the demonstration of androgen and other steroid hormone responsiveness in culture, but its use limits the understanding of stimulatory or inhibitory roles of hormones or factors on prostate and other cancer cells because of the inclusion of an undeteπriined amount of undefined components. A serum-free medium circumvents this problem, as shown in subsequent Examples. In addition, autonomy may be the loss of the receptor for the serum factor and/or the loss of the intracellular steroid hormone receptor. If this hypothesis is correct it should be possible to identify cells that possess steroid receptors but still have lost "sensitivity" to the hormone by virtue of the lack of the effect of the inhibitor. Most notably, this is the case with DU145 and ALVA-41 cells. As defined by immunohistochemistry, the DU145 cells are definitely AR⁺ (Brolin J et al. (1992) The Prostate 20, 281-295). As defined by a number of criteria, the ALVA-41cells are AR⁺ (Nakhla AM and Rosner W (1994) Steroids 59, 586-589). A new concept explaining the progression of normal tissue cells to hormone autonomous cancers is discussed in more detail in an Example below.

0444Exposure of androgen responsive prostate cancer cells to Fe (JS) results in cell death. Compounds containing available Fe (JS) offer the possibility of new therapies for prostate cancer localized to the tissue. It is proposed that deprivation of iron will be a highly effective means of eliminating the most dangerous hormone autonomous forms of prostate cancer. The measurement of thyroid hormone receptors in prostate cancer should be initiated as a diagnostic tool to determine iron sensitivity. Moveover, new therapy mode for tumors containing mixtures of both hormone responsive and autonomous cells is suggested, based on the observation that deprivation of iron can equally kill both types of cancer. This suggests that systemic Fe (JS) therapy for disseminated prostate cancer may be efficacious.

0445It is definitely possible that iron in the Fe (JS) form and compounds containing it will be effective anti-prostate cancer treatments, and that direct injection (or painting) of localized prostate tumors or metastasis at other sites (e.g. bone) might effectively kill these cancers without concomitant systemic effects. This therapy potentially could replace such protocols as systemic chemotherapy (physically damaging), radiotherapy (damage to collateral tissues) or the use of locally acting radioactive gold chips that are complex to handle in the surgical environment and must be implanted and removed surgically. Furthermore, iron therapies can be repeated frequently by application via transrectal or fransurethral access, using conventional techniques. This approach is unique and has not been discussed or suggested anywhere else in the literature. Such iron treatments may be a useful therapy for benign prostatic hypertrophy (BPH). As discussed above, this condition is very common in older men and is treated usually by surgery. Application of iron compounds is a new approach to treatment of BPH. Similarly, a Fe (JS) solution could be applied to breast cancer lumpectomy or mastectomy sites at the time of surgery, and/or applied by injection to the sites subsequent to surgery.

0446Example 13. Growth in Serum-free Defined Medium versus Growth in CDE-Serum + E₂

0447The defined media described in Example 10 were used to verify the presence of a serum- borne inhibitor. The growth of six different ER⁺ cell lines was compared in serum-free defined media (TABLE 7) to the effects seen in cultures supplemented with CDE-horse serum. These studies are shown in Figs. 56 and 57. Estrogenic effects are recorded for each set of conditions with each cell line.

0448MCF-7K Cells in Serum-free and Serum Containing Medium ± E₂. The first studies were done with steroid hormone responsive human cancer cell lines. Fig. 56A shows MCF-7K cell growth in serum-free DDM-2MF ± 10 nM E^. The population replicated logarithmically for 12 days. E₂ had no effect on growth rate or saturation density. These results were in contrast to assays done in D-MEM/F-12 supplemented with CDE horse serum (Fig. 56B). Above 10% (v/v) serum, growth was progressively inhibited. The inhibition caused by any serum concentration was reversed by E_. Measured on assay day 10, a 3 CPD estrogenic effect was observed which was a 2³ or 8-fold cell number increase. The experiments were also done with MCF-7A cells with similar results (data not shown). This effect in CDE-serum was as great as that reported for a special response clone of the MCF-7 cell line (Wiese TE etal (1992) In Vitro Cell Dev Biol 28A, 595-602).

0449T47D Cells in Serum-free and Serum Containing Medium ± E₂. Fig. 56C shows the growth of T47D cells in serum-free defined DDM-2MF + 10 nM E^ Although a small effect of estrogen was observed on growth rate, the most significant effect was an increase in stationary densities by 0.5 to 1.0 CPD. Li contrast, the effect of E₂ was much greater in medium containing CDE horse serum (Fig. 56D). At 50% (v/v) CDE-serum, growth was completely inhibited. The estrogenic effect under these conditions was > 5 CPD. This was more than a 2⁵ or 32-fold hormone effect on cell number. Comparison of these results with those of others (Chalbos D et al (1982) J Gin Endocrinol Metab 55, 276-283; Schatz RW et al. (1985) J Cell Physiol 124, 386-390); Soto AM et al. (1986) Cancer Res 46, 2271-2275; Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44- 52;Reese CC et al. (1988) Ann NY Acad Sci 538, 112-121) confirmed that the conditions in Fig. 56D were substantially more effective. Comparable experiments with the ZR-75-1 line gave results intermediate between MCF-7 and T47D cells (data not shown). ZR-75-1 cells showed no effect of Ε₂ in serum-free defined DDM-2MF. This line grows more slowly than MCF-7 or T47D cells in defined medium and in serum-supplemented cultures (Ogasawara M and Sirbasku DA (1988) In Vitro Cell Dev Biol 24, 911-920). The maximum estrogenic effects of the preferred embodiment recorded with ZR-75-1 cells in D-MEM/F-12 with 50% (v/v) CDE-horse serum ranged between 3 and 4 CPD after 14 days. This was greater than reported by others in serum containing (Darbre P et al. (1983) Cancer Res 43, 349-355; Kenney m et al. (1993) J Cell Physiol 156, 497-514) or "serum- free" medium (Allegra JC and Lippman ME (1978) Cancer Res 38, 3823-3829; Darbre PD et al. (1984) Cancer Res 44, 2790-2793).

0450LNCaP Cells in Serum-free and Serum Containing Medium + E₂. In another study, the effects of J^ on the growth of the LNCaP human prostatic carcinoma cell lines in defined medium and in serum-supplemented culture were compared. This cell line bears a point mutation in the AR that permits high affinity binding of estrogens to the altered receptor (Veldscholte J et al. (1990) Biochem Biophys Res Commun 173, 534-540; Veldscholte J et al. (1990) Biochim Biophys Ada 1052, 187-194). In addition, it is possible that estrogens cause LNCaP growth via a separate functional ER (Castagnetta LA and Carruba G (1995) Ciba Found Symp 191, 269-286). Irrespective of mechanism, estrogens are known to promote LNCaP growth (Belanger C et al. (1990) Ann NY Acad Sci 595, 399-402; Veldscholte J et al. (1990) Biochem Biophys Res Commun 173, 534-540; Veldscholte J et al. (1990) Biochim Biophys Ada 1052, 187-194; Castagnetta LA and Carruba G (1995) Ciba Found Symp 191, 269-286). As presented herein (Fig. 56E), this cell line in serum-free defined CAPM showed essentially no E₂ effect on growth rate and < 1.0 CPD on saturation density. When LNCaP growth assays were done in medium with CDE-horse serum, the mitogenic effect of E₂ was > 5 CPD (Fig. 56F). Estrogenic effects herein were larger than reported by others with LNCaP cells in serum containing culture (Belanger C et al. (1990) Ann NY Acad Sci 595, 399-402; Castagnetta LA and Carruba G (1995) Ciba Found Symp 191, 269-286).

0451LNCaP Cell Growth in CAPM Defined Medium with CDE-Horse Serum and + DHT or E₂. To confirm that the serum-bome inhibitor can be assessed even in the presence of all of the components of serum-free defined medium, an example experiment is shown in Fig. 58. The LNCaP cells were grown in serum-free CAPM supplemented with increasing concentrations of CDE-horse serum without steroids and in assay dishes with the CDE-serum plus 10 nM E^ or 10 nM DHT. Without steroid, the CDE-horse serum showed the expected progressive inhibition. Both the estrogen and androgen reversed this inhibition completely at every serum concentration. Clearly, the inhibitor in serum possesses a very special quality that blocks the action of the many mitogenic agents present in defined media.

0452GHjC.Cells in Serum-free and Serum Containing Medium + E₂. Li the next studies, shown in Fig. 57A, the growth of rodent ER⁺ cell lines in defined medium and CDE serum- containing medium with and without E₂ were compared. The study was with the GILA rat pituitary tumor cell line. Li serum-free PCM-9, E₂ had no effect on growth rate or saturation density (Fig. 57A). Li contrast, the cells were highly estrogen responsive in CDE-horse serum (Fig. 57B). Li > 30% (v/v) CDE-serum, the estrogenic effect was > 4.5 CPD (i.e. > 22-fold cell number increase). The GH C₁ response obtained was substantially greater than that previously reported in cultures containing serum from a gelded horse (Amara JF and Dannies PS (1983) Endocrinology 112, 1141- 1143). Replicate studies with the GHi and GH₃ rat pituitary tumor cells gave results equivalent to those shown in Figs.57A and 57B (results not shown).

0453MTW9 PL2 Cells in Serum-free and Serum Containing Medium + E₂. Fig. 57C shows the effect of E₂ on growth of the MTW9/PL2 rat mammary tumor cells in serum-free DDM- 2A. There was a small effect on growth rate and a < 1.0 CPD effect on saturation density. When the same cells were assayed in D-MEM/F-12 containing CDE horse serum, the effect of E₂ was remarkable (Fig. 57D). Cell number differences of 2⁶ (i.e. 64-fold) were recorded in 50% (v/v) serum in a seven-day assay. This result agrees with those presented above in this disclosure. Furthermore, comparison of MTW9/PL2 responses (Fig. 57D) to those of the human breast cancer cells (Figs.56B and 56D) confirms that the ER⁺ rat cells are the most estrogen responsive mammary origin line yet developed. H301Cells in Serum-free and Serum Containing Medium + E₂. Li the final studies, the effect of E₂ on the growth of the H301 hamster kidney tumor cells in serum-free medium was compared to that in CDE horse serum containing medium. Estrogen had no effect on H301 cell growth in serum-free defined CAPM (Fig. 57E). In contrast, F^ induced H-301 cell number increases of > 2⁴ (i.e. > 16-fold) were recorded in D-MEM/F-12 containing > 30% (v/v) CDE serum (Fig. 57F). The H301 response was similar to the MCF-7 cells in that 50% (v/v) CDE-serum did not fully inhibit. The magnitude of the estrogenic effect with H301 cells was equal to that reported by others studying this line in cultures supplemented with CDE serum prepared by different methods (Soto AM et al. (1988) Cancer Res 48, 3676-3680).

0454Discussion of Example 13. The new serum-free defined medium serves as part of a model system for identifying physiologically relevant new molecules. When completely serum-free defined conditions were employed in the past, the effects of estrogens were either marginal or insignificant as has been discussed above. The earlier observations in completely serum-free defined culture medium have been extended in the present investigation. Direct comparisons were made between estrogenic effects in serum-free defined culture and estrogenic effects in medium containing CDE serum. The results were unequivocal. With every cell line tested, CDE serum was required to demonstrate significant estrogenic effects on logarithotnic cell growth rates. A major advance provided was the clear demonstration that high concentrations of serum are required to observe large magnitude estrogenic effects. Furthermore, the inhibitory effects of serum are dose dependent even in the presence of the components used to formulate serum-free medium. This indicates that growth is progressively negatively regulated. This observation has physiological implications. Changes in the serum concentration of the inhibitor, or changes in availability to target tissues, will have direct effects on the rate of cell replication. The results in Figs. 56-58 point to serum as the best source yet identified to obtain the component that regulates sex steroid responsive growth. The tissue origin of the serum regulator remains to be investigated.

0455Example 14. Action of DES on Human AR⁺LNCaP Prostate Cancer Cells.

0456In this Example, it is demonstrated that DES des not inhibit steroidogenic cell growth and may be suitable for use in cancer therapies, including but not limited to other therapies disclosed herein. LNCaP Cells and DES Action. Diethylstilbestrol (DES) is now used as one of the primary treatments for prostatic cancer (Seidenfeld J et al. (2000) Ann Intern Med 132, 566-577). Its action is likely mediated through the hypothalamus-pituitary axis (Seidenfeld J et al. (2000) Ann Intern Med 132, 566-577). DES causes suppression of anterior pituitary gonadotrophins and therefore suppresses testicular output of androgens. Although it is thought that DES has no direct effects on prostate cancer cells, the development of the assay methodology set out herein permitted a direct assessment of this issue. The AR⁺ LNCaP cells were used as a model for these tests (Fig. 59). As shown in Fig. 59 A, 10 nM DHT effectively reversed the inhibition caused by higher concentrations of CDE-horse serum in D-MEM/F-12. Likewise, 10 nM E_ also reversed the CDE- serum caused inhibition completely (Fig. 59B). However, the same concentration of DES was entirely ineffective (Fig. 59C). DES did not reverse the serum caused inhibition. The synthetic estrogen had no direct positive effect on LNCaP cell growth. Li the final study of this series, DES addition to medium containing DHT or -E^ did not affect the reversal caused by these two natural steroids (Fig. 59D). Therefore, DES is not a direct inhibitor of androgen or estrogen promoted LNCaP cell growth. The view that DES acts indirectly to cause chemical castration is consistent with the present results. These results are supported by other studies indicating that DES does not bind to the AR of LNCaP cells (Montgomery BT etal. (1992) The Prostate 21, 63-73).

0457Discussion of Example 14. The fact that DES is a major treatment for prostate cancer but does not act directly on the tissue has therapeutic implications. For prostate cancer localized to the organ, or specific metastases in other locations (e.g. bone, liver or lung), direct application of Fe (JS) offers a therapy with a different mode of action. It is also possible that local Fe (JS) therapy (as described in Example 12) can be used in conjunction with conventional systemic DES treatment to increase effectiveness above that with either treatment alone. There is another potential advantage of local Fe (JS) treatment over systemic DES treatment. DES has many side-effects in males. Some present considerable discomfort or medical problems. Locally applied Fe (JS) is absorbed by the body to form non-toxic mono ferric and diferric transferrin by chelation with the large pool of available apotransferrin. The iron containing proteins formed are no problem for the body because they are the natural physiological forms of iron delivered to all tissues.

0458Example 15. Preparation of Inhibitor Depleted Serum for Control Studies and Stability Properties of the Inhibitor This Example, lists several acceptable techniques for useful inactivated immunoglobulin inhibitors, and distinguishes the inhibitors from the classical "estrocolyone. "

0459Effect of Dialysis on Estrogenic Effects. CDE-horse serum was dialyzed at 4°C against

04600.05M Tris-HCl, pH 7.4, for up to 72 hours with buffer changes every 24 hours using a Spectropor dialysis membrane. The resulting serum was tested for estrogenic effects with MWT9/PL2 cells as shown in Fig. 60. There was near a total loss of estrogen reversible inhibitory activity accompanying this treatment. It was found consistently (N=14) that this treatment resulted in the appearance of an estrogen irreversible inhibitor at serum concentrations above 10% (v/v). It was possible that the estrogen reversible inhibitor was low molecular weight and had passed through the dialysis membrane.

0461Ultrafiltration of CDE-Serum and Estrogenic Effects. CDE-horse serum was submitted to nitrogen gas pressure ultrafiltration using an Amicon unit and an YM-30 membrane (i.e. a 30,000 molecular weight cut-off). The filtrate was assayed with MTW9/PL2 cells directly whereas the retentate was diluted to the original volume with normal saline before assay. The filtrate (Fig. 61B) supported growth but without any estrogenic effect. The retentate (Fig. 61A) demonstrated the usual high estrogenic effect (i.e. 6 CPD) seen in the other MTW9/PL2 cell assays presented in this Example. It is unlikely the Tris dialysis results described above came from passage of the inhibitor through the membrane. The ultrafiltration results confirm a molecular weight > 30,000 daltons. The combined results of dialysis and ultrafiltration suggest a lower molecular weight cofactor that might help stabilize the estrogen reversible inhibitor.

0462Heat Treatment and Estrogenic Effects Inhibitor Content of CDE-serum. The heat stability of the estrogen reversible inhibitor of CDE-horse serum was investigated at 50°C and 60°C with the MTW9/PL cells. Heating at 50°C for 30 minutes reduced the estrogen effect to 4.6 CPD (Fig. 62) instead of the usual 5 to 6 CPD. The effect of heating at 50°C for up to 30 hours is shown in Fig. 63. By 20 hours, the estrogenic effect with MTW9/PL2 cells was reduced to &lt; 1.0 CPD. Nonetheless, this serum still supported full growth of the MTW9/PL2 cells (Fig. 64). Another effective method requiring less time is shown in (Fig. 65). Heating at 60°C for 90 minutes yielded serum that supported high levels of growth (i.e. ≥ 6 CPD) but had lost all inhibitor activity. This easy treatment, which is especially fast and inexpensive to perform, provides a control serum that has applications in assay of test substances.

0463Affi-Gel Blue Extraction of CDE-Serum. An aliquot of CDE-horse serum was passed through a 5 mL Affi-Gel Blue™ affinity chromatography column (Bio-Rad, Lie), prepared according to the manufacturer's instructions. The flow through fraction was tested in the assay for estrogen mitogenic activity at 0 to 50% (v/v). The results are shown in Fig. 66. The 5 mL Affi-Gel Blue™ column removed more than 80% of the inhibitory activity in the serum. Increasing the column bed to 10 mL resulted in removal of more than 90% of the inhibitory activity in the serum.

0464Acid Treatment of CDE-serum. CDE-horse serum was adjusted to pH 4.5 with HC1 and incubated for 16 hrs at 4°C. The resulting serum readjusted to pH 7.4 and tested as previously done with MTW9/PL2 cells. Acid treated CDE-serum promoted only limited MTW9/PL2 cell growth and an estrogenic effect of &lt; 0.5 CPD. Not only was the inhibitor acid labile, but the serum components that support growth at &lt; 10% (v/v) were also adversely affected (data not shown).

0465Urea Treatment of CDE-serum. CDE-horse serum was dialyzed against 0.05 M Tris- HC1, pH 7.4, with 50 mM calcium chloride and 6 M urea for 16 hours at 4°C. The urea was removed by dialysis against the buffer without urea. The addition of CaCl₂ to the Tris buffer protects the activity (see results below). The resulting serum was tested as previously described. As shown in Fig. 67, the inhibitory activity was inactivated. Also, the growth promoting activity of &lt; 10% (v/v) was also adversely affected.

0466Discussion of Example 15. The preparation of inhibitor-depleted serum has applications with regard to testing compounds that might possess cytotoxic activity independent of any steroid hormone-like cell growth stimulating ("steroidogenic") effects or other hormone-like properties. The methods outlined will permit assays of commercial, environmental, industrial and medical compounds, substances and mixtures for inhibitor-like activity and/or cytotoxic activity in the same preparations. There is another very important application of this technology. Development of compounds with estrogen reversible and estrogen irreversible inhibitor-like activity, including peptides, recombinant DNA products, or synthetic organic or inorganic compounds can be sought using inhibitor-depleted serum as the assay base. The new agents can be compared directly to the purified serum inhibitor to determine their efficacy and potency. It is anticipated that this technology will yield compounds that mimic the serum inhibitor and can be used to treat various forms of mucosal cancers including breast and prostate and colon. This method is expected to allow rapid examination of many compounds. The preferred preparation method for control serum is heating at 50°C for about 20 to 30 hours or 60°C for about 90 minutes. Affi-Gel Blue treatment is effective, but only with small volumes of serum (e.g. 1 to 2 liters). Affi-Gel Blue is more expensive and time consuming than the heating methods. Tris dialysis, acid pH treatment and urea treatment are not as satisfactory but can be applied as required under special circumstances.

0467The results presented herein distinguish the estrogen reversible inhibitor sought here from estrocolyone 1 (Soto AM et al. (1992) J Steroid Biochem Mol Biol 43, 703-712). Estrocolyone 1 is stable to treatment with 6 M urea, stable at 60°C for 2 hours, and stable in 2 M acetic acid. Furthermore, estrocolyone does not bind to Affi-Gel Blue. The serum-bome inhibitor described herein does not share any of these properties.

0468Example 16. Effects of Conventional Purification Methods on the Properties of the Estrogen Reversible Serum-borne Inhibitor This Example demonstrates the mainly adverse effects of conventional purification techniques on the desired properties of the present inhibitors. This Example also illustrates that the conventional purification techniques can be used to produce certain desired effects on the inhibitors.

0469CDE-horse Serum Effects Used to Calculate the ED₅₀ Required for Purification

0470Quantification. Conduct of purifications requires measurement of specific activity (i.e. ED₅₀) and definition of units of activity. The results in Fig. 68 present examples of how estimates of the ED₅₀ concentrations protein required for half-maximum estrogenic effects were determined. Two representative commonly studied estrogen sensitive human cell lines and two established rodent lines were selected for presentation. Figs. 68A, 68B, 68C and 68D show assay results with the MCF-7K and ZR-75-1 human breast cancer cells, the MTW9/PL2 rat mammary tumor cells, and the GH₄ rat pituitary tumor cells, respectively. With all four lines, the maximum estrogen reversible inhibition was observed at 50% (v/v) CDE-serum. Under these conditions, estrogen reversed cell number increases (i.e. estrogenic effects) ranged from 2³"¹ to 2⁵'⁵ (i.e. 2^CPD) or 8-fold with MCF-7K cells to 45-fold with MTW9/PL2 cells. Serum concentrations of 8 to 16% (v/v) supported ED₅₀ effects. This corresponded to protein concentrations of 2.4 to 4.8 mg/mL (TABLE 8). One unit of activity is the amount that achieves ED₅₀. To achieve maximum inhibition, 15 ± 2.5 mg/mL of protein were required (i.e. 50% serum). The experiments presented here support the previous conclusion that serum contains an estrogen reversible inhibitor.

0471Table 8 Inhibitor Purification by General Methods

0472<img file="EP1328293A2_D0008.tif" /> Ammonium Sulfate Precipitation. Ammonium sulfate precipitation was studied both to increase specific activity and to decease the volume of the serum. The activity precipitated over a broad concentration range. The precipitate from 0 to 40% saturated ammonium sulfate contained &lt; 30% of the protein and activity of whole CDE-serum. Replicate 40 to 75% saturation precipitates contained approximately 80% of the total activity units and 60% of the serum protein. The ED₅₀ of this fraction ranged between 3.0 to 6.7 mg/ml (Fig. 69A). The ammonium sulfate results are summarized in TABLE 8. The ammonium sulfate fractions were not stable to freezing and thawing nor were they stable during storage at 4 C or 23 C. To best preserve activity, the precipitated material had to be used immediately in the next isolation step. Our results differ from those in another report (Soto AM et al. (1992) J Steroid Biochem Mol Biol 43, 703-712) indicating that the estrogen reversible inhibitory activity was stable in ammonium sulfate.

0473Proteinase Inhibitors. L replicate studies, the addition of the proteinase inhibitors 4- amidinophenylmethanesulfonyl fluoride (0.01 mg/ml), phenylmethylsulfonyl flouride (0.10 mg/ml), N-tosyl-L-phenylalanine chloromethyl ketone (0.1 mg/ml), and leupeptin (0.01 mg/ml) to the serum before precipitation, or to the 40 to 75% precipitated fraction, was not beneficial. Although effective as a metaloproteinase inhibitor, EDTA was not used because it was expected to remove stabilizing calcium.

0474Affi-Gel Blue Chromatography - General Considerations. Affi-Gel Blue fractionation of CDE-horse serum was done as described by the manufacturer for the isolation of albumin (Bio- Rad Affi-Gel® BlueGel, 50 to 100 mesh, Instruction Manual, Catalog Numbers 153-7301 and 153- 7302). This study addressed two issues. First, it was important to establish that the inhibitor localized as a single protein, or at most only a few proteins. This issue has not been addressed directly before. Affi-Gel BlueGel is a mild method that effectively separates functionally active serum proteins (Sirbasku DA et al. (1991) Biochemistry 30, 295-304; Travis J et al. (1976) Biochem J 157, 301-306; Gianazza E and Arnaud P (1982) Biochem J 201, 129-136; Iqbal MJ and Johnson MW et al. (1977) J Steroid Biochem 8, 977-983). Second, Affi-Gel BlueGel offers a reliable and convenient means of isolating relatively pure native serum albumin (Travis J et al. (1976) Biochem J 157, 301-306; Gianazza E and Arnaud P (1982) Biochem J 201, 129-136). This is especially significant because reports from other laboratories have cited albumin as "the" serum inhibitor (Laursen I et al. (1990) Anticancer Res 10, 703-712; Sonnenschein C et al. (1996) J Steroid Biochem Mol Biol 59, 147-154).

0475Affi-Gel Blue Chromatography - Technical Applications. A one-liter column (5 cm x 51 cm) was equilibrated with 0.05 M Tris-HCl, pH 7.4. The 40 to 75% ammonium sulfate precipitated material (438 mL, with 23.5 grams of protein) was dialyzed against this buffer and applied to the column. After washing with equilibration buffer, elution was done with increasing step concentrations of 0.15, 1.0 and 3.0 M NaCl in the buffer. The four pools contained 3.7, 3.2, 27.3 contain albumin, as expected (Travis J et al. (1976) Biochem J 157, 301-306; Gianazza E and Arnaud P (1982) Biochem J 201, 129-136). The 1.0 and 3.0 M NaCl pools contained 70% albumin (Travis J et al. (1976) Biochem J 157, 301-306; Gianazza E and Arnaud P (1982) Biochem J 201, 129-136). SDS-PAGE and Coomassie Blue staining confirmed albumin in these pools (results not shown). Assay of the four pools showed no inhibitory activity (TABLE 8). Affi-Gel BlueGel either retained the activity even with a 3.0 M NaCl wash, or it caused inactivation. Fig. 69B shows an example assay of an albumin rich pool. The same results were obtained with whole CDE-serum applied to the same column (results not presented). The same four pools shown in TABLE 8 also showed no recovery of the estrogen reversible inhibitory activity. The Affi-Gel BlueGel results in TABLE 8 suggested another use for this matrix. Passage of CDE-horse serum through Affi-Gel Blue removed the majority of the estrogen reversible inhibitory activity for MTW9/PL2 cells (Fig. 70). This was effective even though the volume of serum applied was more than five times the volume of the resin. The maximum estrogenic effect seen with Affi-Gel BlueGel treated serum as 1.5 CPD (Fig. 70) whereas the maximum effect in control CDE-serum was 5.5 CPD (Fig. 68C) with MTW9/PL2 cells. The residual activity in the by-pass fraction is likely due to IgA IgM. Small amounts of these immunoglobulins are usually in the by-pass of this column (Gianazza E and Arnaud P (1982) Biochem J 201, 129-136).

0476Human Serum Albumin as Inhibitor. Li studies not presented, three preparations of human serum albumin were assayed for estrogen reversible inhibitory activity with ER⁺ human and rodent cell lines. Globulin containing (96 to 99% albumin), crystalline, and Cohn's fraction V human serum albumin (all from Sigma) were not inhibitory at concentrations up to 12 mg/mL. The assays showed the same pattern as in FIG. 69B. These results further support our earlier conclusion that albumin was not the estrogen reversible inhibitor.

0477DEAE Sepharose Chromatography - General Considerations. We next applied DEAE Sepharose chromatography. Because ammonium sulfate precipitation provided no benefit beyond sample concentration (TABLE 8), we instead used whole CDE-serum. The DEAE Sepharose column was eluted with both step increases in NaCl concentration and with linear gradients of NaCl. Eight permutations of pH, NaCl elution concentrations, and gradient protocols were analyzed. The results presented in Fig. 71 were the optimum conditions identified. DEAE Sepharose Chromatography - Technical Considerations. DEAE Sepharose used at pH 8.6 provided the best separations. Seven pools were obtained (Fig. 71). Activity assays of each pool are summarized in TABLE 8. The flow-through fraction (pool I) and the wash with equilibration buffer (pool H) together contained 389% of the applied activity and 5.6% of the protein applied. Pool II was particularly active, with the lowest ED₅₀ concentration of any from general chromatography (i.e. 30 μg/mL). Pools UI, JN and V were less active although they still contained

0478126%o of the units and 37.7%> of the protein. Pool VI contained 1042% of the activity applied and

047932% of the protein. Pool VI alone contained 10 times more activity than applied to the column (TABLE 8). Pool VII was markedly less active with only 19% of the activity and 4.5% of the protein. Although 79.8% of the applied protein was recovered, more than 15 times the expected units were recovered. Although the reason(s) for the apparent increase is not clear, other investigators (Dell'Aquila ML and Gaffiiey EV (1984) J Natl Cancer Inst 73, 397-403) have reported identification of estrogen irreversible inhibitors of the ER⁺ MCF-7, T47D and ZR-75-1 human breast cancer cell lines in fractions from DEAE Sepharose. It is possible that the higher pH conditions have inactivated these inhibitors and thereby allow greater effect of the estrogen reversible form(s). It is also possible that exposure to high pH alters the inhibitor to yield a more active form. DEAE Sepharose Chromatography - Stability Considerations and Evidence of more than one Activity. Activity was found in pools I, II and VI. This suggested more than one estrogen reversible inhibitor. In any case, there was a stability problem. An example of this is shown in replicate assays of pool VI over a period of 21 days. The first assay of activity immediately after isolation is shown in Fig. 69C. The effects on ER⁺ cell growth were biphasic. At lower protein concentrations, estrogen reversible inhibition was observed. At higher concentrations, the pool material became irreversibly inhibitory. With all active DEAE pools, this biphasic pattern was consistent. Additional sequential assays initiated after 14 days storage at 4°C or 23°C showed another consistent finding. The maximum estrogenic effect caused by Ε₂ decreased > 90% (Fig. 69D). As estrogen reversible inhibition decayed, only the estrogen irreversible inhibition remained (Fig. 69D). From these results, decay most likely resulted in formation of an altered inhibitor that was estrogen irreversible.

0480Phenyl Sepharose Chromatography. Phenyl Sepharose chromatography has been previously reported to effectively enrich human serum-derived estrocolyone 1 (Soto AM et al. (1992) J Steroid Biochem Mol Biol 43, 703-712). When samples of that estrogen reversible activity were applied under high salt conditions, and the elution done with decreasing salt concentrations, a single inhibitory fraction was separated from the bulk of the proteins (Soto AM et al. (1992) J Steroid Biochem Mol Biol 43, 703-712). In the present studies, further purification of DEAE Sepharose pool VI was investigated using phenyl Sepharose. Fig. 72 presents the results of the optimum of four elution protocols investigated. Activity was located in fractions 100 through 130 that contained 12% of the applied protein. The specific activity increased 2.3-fold compared to DEAE Sepharose pool VI (TABLE 8). The initial assay results (data not shown) were similar to those in Fig. 69C. Sequential assays again confirmed a rapid inactivation ending with estrogen irreversible inhibition similar to that shown in Fig. 69D. As seen before, an estrogen irreversible inhibitor was generated upon standing in buffer. HTP Bio-Gel Chromatography and other Methods. Further purification of DEAE

0481Sepharose pool VI was also attempted using HTP Bio-Gel (hydroxylapatite). Fig. 73 shows the results of the most effective of three HTP Bio-Gel elution protocols attempted. The protein and total units of activity recovered from this column were 69.6% and 51.6% respectively. As summarized in TABLE 8, the specific activities (i.e. ED₅₀) of pools I, III and V were not improved compared to DEAE pool VI. The specific activities of pools II and V were significantly improvement. Sequential assays of these showed responses similar to those in Figs. 69C and 69D. The initial estrogen reversible activity decayed within 14 days to irreversible inhibition. Li studies not presented, Concanavalin A Sepharose and metal (Zn²⁺) chelate affinity chromatography were also attempted with whole serum, the active fractions from DEAE Sepharose and with the 40 to 75% ammonium sulfate precipitate. These methods were not effective. When activity was obtained, it decayed within two to three weeks in the same pattern as shown in Figs. 69C and 69D. Analysis of the report (Soto AM et al. (1992) J Steroid Biochem Mol Biol 43, 703-712) attempting estrocolyone isolation confirmed substantially the same instability and low yield problems. In the studies described in this Example, it is very clear that purification of the inhibitor had not yet been achieved using conventional purification methods, although useful ways of producing irreversible inhibitor compositions were revealed. Discussion of Example 16. All of the methods described in this Example are conventional protein purification methods in general use today. They commonly yield high specific activity or high purity protein preparations. Because they were carried out under what are considered non- denaturating conditions and non-reducing conditions, the expected outcome was isolation of an active estrogen reversible inhibitor. It is clear that a spectrum of the usual methods will not yield the estrogen reversible inhibitor in an active, stable form.

0482It should be noted that the ammonium sulfate experiments alone clearly differentiate the present serum described inhibitor from that described in U.S. Patent Nos. 4,859,585 (Sonnenschein) and 5,135,849 (Soto). Those patents teach the use of a stable inhibitor obtained by ammonium sulfate fractionation. Li the present case, however, an ammonium sulfate fraction yields unstable activity.

0483Affi-Gel BlueGel has usefulness as a method of preparation of inhibitor depleted serum. It may be preferred under circumstances where heating (the other very effective method) might destroy some component in serum needed for a specialized mucosal or other origin cancer cells. The Affi- Gel Blue results, and others presented herein, support the view that serum albumin is not the serum- bome estrogen reversible inhibitor activity sought herein.

0484The results of the DEAE chromatography indicate that there is more than one inhibitor. The number cannot be established by that method, but elution localization suggests at least two and possibly more. Further, in light of the results shown in Example 20, the DEAE elution pattern shown in Fig. 69D is consistent with the two or more inactivated inhibitors being denatured forms of immunoglobulins IgA IgM/IgGl/IgG2 may be very potent antitumor agents and worthy of consideration as new treatment modalities. It should be noted that these fractions were very freeze- thaw sensitive, especially in the absence of calcium. Li every case where activity was localized to a chromatographic pool, the activity decayed within 21 to 28 days to an estrogen irreversible agent. This suggests that a denaturation process is ongoing in buffers mostly without calcium, and that a potentially important product is formed that may have cancer therapeutic value. This may be a means of generating very high potency irreversible inhibitors of mammary cancers and other types of mucosal cancers.

0485Example 17. Calcium Stabilization and Correlation with ³H-DHT Binding and Immunoprecipitation by Antibodies Raised to Human SHBG

0486General Protein Isolation Principals. Before continuing the isolation attempts described above, principals common to all protein isolation attempts were applied. Information was sought about the conditions that would best stabilize the activity, and in doing so an understanding of new/ appropriate methods of purification was gained. Effect of Calcium on Tris-HCl Dialysis Retention of Estrogenic Activity. CDE-horse serum was dialyzed against 0.05M Tris-HCl, pH 7.4, with 50 mM CaCl₂ for 72 hrs, buffer changes every 24 hrs at 4°C using a 6000-8000 molecular weight cut-off dialysis membrane. The resulting serum was assayed with MTW9/PL2 cells + 10 nM E₂. As shown in Fig. 74, the usual large magnitude estrogenic effects were identified. The presence of calcium in the buffer completely prevented the inactivation found when dialysis was done in buffer without calcium (compare to Fig. 60).

0487Chelex™ Treatment and Protective Effects of Calcium Ions Against Heat Inactivation. CDE-horse serum was treated with Chelex™ resin beads to remove free metal ions including calcium. This was done to continue the evaluation of calcium as a stabilizer. The serum was incubated with 10% (w/v) prewashed Chelex™ 100 resin (100-200 mesh, sodium form) (Bio- Rad) for 2 hrs at room temperature. At the end of the incubation, the serum was separated from the Chelex beads by 0.2 μm pore filtration. Calcium concentrations were determined with a calcium- detecting probe. They were < 10 nM. Next the Chelex treated serum was incubated at 50°C either without added calcium or in the presence of 1.0, 10 and 50 mM CaCl₂. The serum was assayed with MTW9/PL2 cells at the designated times shown in Fig. 75 (30% Chelex treated serum + 10 nM E>) to determine estrogenic effects. Without calcium, total activity was lost within 3 hours. In the presence of increasing calcium, the activity was progressively stabilized. At 50 mM CaCl₂, < 15 % of the activity was lost even after 30 hours at 50°C. A control with CDE-horse serum is shown. CDE-serum alone lost complete activity by 20 hours at 50°C as expected from the results in Fig. 64. Clearly, the addition of calcium stabilized the activity. Chelex™ Treatment and Protective Effects of Metal Ions Against Heat Inactivation.

0488CDE-horse serum was treated with Chelex™ resin beads to remove free metal ions including calcium. This was a continuation of the study above but with other metal ions to determine if they substituted for calcium. The experiment was conducted as described in the paragraph above but with the change that the incubation temperature was lowered to 37°C to permit more accurate estimation of the early kinetics of inactivation. Chelex treated serum lost 80% activity in 30 hours at this temperature (Fig.76). With 10 mM CaCl₂, protection was nearly complete at this temperature for 30 hours. However, zinc, magnesium and manganese ions offered no protection. Because these are expected to be common substitutes for calcium, it is likely that stabilization by calcium is quite specific.

0489Labeled Steroid Hormone Binding in CDE-Serum and Scatchard Analysis. One of the basic tenets of the estrocolyone hypothesis is that there are serum proteins that bind sex steroid hormones at affinities (i.e. Ka) in the picomolar range (Soto AM et al. (1986) Cancer Res 46, 2271- 2275). However, what was found instead with CDE-horse serum is specific binding of ³H-DHT in the of K_d range 1 to 5 nM as determined by Scatchard analysis (Scatchard G (1949) Ann NY Acad Sci 51, 660-672) (Fig. 77). The binding methods are presented in the General Materials and Methods section. With CDE-rat serum the binding affinities were even higher (results not shown). Routinely, the affinities for specific binding of ³H-DHT with CDE-rat serum were in the K_d range 15 to 40 nM. With both CDE-horse and CDE-rat serum, binding K_d for ^-E^ was about two to five times higher concentrations (results not presented). CDE-serum shows specific binding of sex steroid hormones, but the affinity was not sufficiently high to support the conclusions of others (Soto AM et al. (1986) Cancer Res 46, 2271-2275) concerning picomolar affinities and the estrocolyone hypothesis.

0490Correlation between Calcium Stabilization of Inhibitor Activity and Calcium Stabilization of ³H-DHT Binding to CDE-horse Serum. As shown in Fig. 75 with CDE-serum that had been Chelex treated, calcium protected the inhibitor activity from heat inactivation. This study was repeated with another batch of Chelex treated CDE-horse serum (Fig. 78). This study showed essentially the same results as presented in Fig. 75. However, in parallel, the Chelex treated serum was also assayed for ³H-DHT binding (Fig.78B). Clearly, as the calcium concentration was increased in the serum, there was protection of ³H-DHT binding that paralleled the protection of the estrogenic effect shown in Fig. 78A. These results implied a relationship between the estrogen reversible inhibitor activity being sought and a sex steroid hormone binding protein.

0491Immunoprecipitation of the ³H-DHT Binding Activity and Estrogenic Activity in CDE-horse Serum. Li the final studies of this series, rabbit antibodies against human SHBG (Accurate Chemicals) were assayed for immunoprecipitation of the ³H-DHT binding activity of CDE-horse serum. After incubation of the serum with the designated dilutions antiserum, Protein G- Sepharose (Pierce kit) was added to absorb the immune complexes, and the resulting serum assayed for binding under conditions described in General Materials and Methods. The results of this study are shown in Fig. 79 A. Increasing antibody decreased the steroid binding activity. Li parallel, the same samples were used to assess estrogenic effects with MTW9/PL2 cells (Fig. 79B). Increasing anti-SHBG decreased the estrogenic effect by decreasing the concentration of the inhibitor in the serum. There appeared to be some type of cross-reaction, but it was still not clear that this proved SHBG-like properties for the estrogen reversible inhibitor. In both Figs. 79A and 79B, addition of control rabbit serum had no effect.

0492Discussion of Example 17. The effect of calcium on both the estrogenic activity and the binding of ³H- DHT to CDE-serum was remarkably similar to data presented by others concerning the stability of human SHBG (Rosner W et al. (1974) Biochim Biophys Ada 351, 92-98). Other investigators have raised the issue of classical SHBG as the sex hormone reversible inhibitor of target cell growth. This seems highly unlikely, however, in light of the results presented above. Both CDE horse serum and CDE rat serum contain concentrations of inhibitor about equal to any of the other serum types investigated. Furthermore, it is accepted knowledge that horse and adult rat serum do not contain SHBG (Corvol P and Bardin CW (1973) Biol Reprod 8, 277-282; Renior J-M et al. (1980) Proc Natl Acad Sci USA 11, 4578-4582; Wenn RV et al. (1977) Endokrinologie 69, 151- 156). Nevertheless, anti-human SHBG purchased from Accurate Chemicals not only immunoprecipitated the activity in serum, but also the ³H-DHT binding activity. This data initially suggested that the inhibitor was a SHBG like activity (Sirbasku DA et al. "Serum factor regulation of estrogen responsive mammary tumor cell growth." Proceedings of the 1997 Meeting of the "Department of Defense Breast Cancer Research Program: An Era of Hope", (Abstract) pp. 739- 740, Washington, D.C., Oct. 31- Nov. 4, 1997). However, there were enough physical differences to indicate that the activity was not actually SHBG and that the cross-reaction with anti-SHBG was possibly misleading.

0493Despite the ambiguity in its identity at that point, it was clear that the estrogen reversible inhibitor sought by herein had different properties than that described in U.S. Patent Nos. 4,859,585 (Sonnenschein) and 5,135,849 (Soto). Li those patents, the activity was not shown to cross-react with anti-human SHBG nor was it stated to share SHBG-like properties. The putative kinship to SHBG provided impetus to use a method that had already been applied to the purification of SHBG in order to identify the inhibitor. Li Example 18 the purification of the estrogen reversible inhibitor activity is described, performing the first, and third through sixth isolation.

0494Example 18. Cortisol Affinity and Phenyl Sepharose Isolation of the "SHBG-like" Estrogen Reversible Inhibitor from CDE-Horse Serum

0495Outcome of the Search for the Estrogen Reversible Inhibitors. As cited above, neither horse or rat serum contains SHBG. Therefore, these were the preferred sera to begin isolation. Partial purification of the inhibitor from serum has been achieved initially by a two-step procedure. The partially purified inhibitor fractions are different than the serum derived inhibitor described in U.S. Pat. No. 4,859,585 (issued to Sonnenschein and Soto), which has been more recently identified as a subtype domain of albumin. By contrast, it has been discovered that IgA and IgM, preferably in dimeric/polymeric form, are steroid hormone reversible inhibitors of cell growth.

0496Two-step Cortisol-agarose and phenyl Sepharose Isolation Method. Based on the perceived SHBG-like properties described above, a new approach to the purification was taken. This method used a two-step cortisol-agarose affinity and phenyl-Sepharose chromatography protocol. It had been employed by others to simultaneously yield purified human cord serum CBG and SHBG (Fernlund P and Laurell C-B (1981) J Steroid Biochem 14, 545-552). The method first required the synthesis of the cortisol affinity matrix. The cortisol-agarose affinity matrix was synthesized and the initial purifications done as described (Fernlund P and Larell C -B (1981) J Steroid Biochem 14, 545-552). An 80 mL bed volume cortisol-agarose column (2.5 cm x 17.8 cm) was equilibrated with a buffer containing 0.05 M piperazine, pH 5.5, with 0.2 M NaCl. Two liters of horse serum were charcoal-dextran extracted at 34°C as described above. For two of the six preparations used in these studies, the serum was depleted of steroid hormones by the Amberlite ™ XAD-4 resin method. There was no resulting difference in the purifications. After removing a 30 mL sample for pre-column activity assay, the remaining volume was adjusted to pH 5.5 with 1.0 N HC1. This was applied to the column at a flow rate of 30 to 40 mL per hour. Throughout the purification, the flow rates were maintained with a peristaltic pump. The effluent was collected and a sample and adjusted to pH 7.2 for post-column assessment of estrogen reversible inhibitory activity. After all of the serum had been applied, the column was washed for 7 days at the same flow rate with the equilibration buffer until the A_280nm of the effluent was &lt; 0.06 versus water.

0497To recover the activity, the cortisol-agarose column was eluted with a 500 mL linear gradient formed with 250 mL of the piperazine/NaCl buffer and 250 mL of the buffer with 1.0 mg/mL cortisol and 10% (v/v) methanol. After completion of the gradient, the column was washed with one volume of the cortisol/methanol buffer. A total volume of 600 mL was collected as 10 mL fractions. As reported by Fernlund &amp; Laurell (Fernlund P and Laurell C-B (1981) J Steroid Biochem 14, 545-552), two separate A_280nm or protein concentration ranges could be recognized, but their separation and individual chromatography on phenyl-Sepharose was no more effective than pooling the entire 600 mL gradient elution and using it for the next step. The total volume from the cortisol gradient was reduced 5 to 8-fold by nitrogen gas pressure Amicon ultrafiltration (YM-10 membrane) and applied directly to the next column without dialysis or pH adjustment.

0498A 28 mL bed volume phenyl-Sepharose (1.5 cm x 16 cm) was equilibrated with 0.05 M Tris-HCl, pH 7.5, containing 0.5 M NaCl. The concentrated cortisol gradient volume was applied at a flow rate of 60 mL/hour (10 mL fractions). The first A_280nm peak observed was a mixture of cortisol and CBG (Fernlund P and Laurell C -B (1981) J Steroid Biochem 14, 545-552). These fractions were combined as cortisol affinity-phenyl Sepharose pool I (CA-PS-pool I). The column was then washed with equilibration buffer until the A₂₈o_ιπn was reduced to 0.002 versus water. The next buffer applied was 0.05 M Tris-HCl, pH 7.5 (60%, v/v) containing 40% (v/v) ethylene glycol. The A_280nm pe k observed with this wash was combined to form CA-PS-pool π that corresponded to SHBG from human serum (Fernlund P and Laurell C -B (1981) J Steroid Biochem 14, 545-552). The two pools were separately concentrated to approximately 40 mL each and dialyzed separately against storage buffer which was 0.05 M Tris-HCl, pH 7.5, containing 0.15 NaCl, 0.05 M CaCl₂ and 60% (v/v) glycerol. The dialysis further concentrated each sample. As last additions, 0.1 mM cortisol was added to CA-PS-pool I and 0.1 mM DHT to CS-PS-pool π. The pools were stored unfrozen at - 20 C. Six replicate isolations were done. The protein yields ranged from 22.8 to 37.7 for CA-PS- pool I and 5.82 to 12.2 mg for Ca-PS-pool π. Based on an average of 60 grams of protein per two liters of CDE-horse serum (i.e. 30 mg/mL), CA-PS-pool II represented about 0.013% of the total protein in serum.

0499Cortisol affinity and phenyl Sepharose Isolation Results and SDS-PAGE Molecular Weight Estimation. The chromatography profiles from the two-step cortisol affinity and phenyl Sepharose isolation of the inhibitors) activity from CDE-horse serum are shown in Fig. 80. The elution from phenyl Sepharose gave the CA-PS-pools I and π. CA-PS-pool I contained predominantly 58 kDa CBG (Rosner W and Bradlow HL (1971) J Gin Endocrinol Metab 33, 193- 198) as confirmed by SDS-PAGE and Western immunoblotting with rabbit anti-horse CBG as well as by partial N" amino acid sequencing of the first 10 to 20 residues (results not presented). SDS- PAGE analyses of three example preparations of CA-PS-pool U are shown in Fig. 81A. Components of 67, 58, 54, and 29 kDa were identified. These were compared to the 48 and 46 kDa units identified for purified human SHBG (Khan MS et al. (1985) Steroids 45, 463-472) (Fig. 81A). Native Molecular Weight Estimation. Analyzes done under non-reducing and non- denaturing conditions using Superdex molecular sieve FPLC at neutral pH in buffers identified components CA-PS-pool I in the exclusion volume at > 900 kDa, and components approximately 350 and 168 kDa (Sirbasku DA et al. "Serum factor regulation of estrogen responsive mammary tumor cell growth." Proceedings of the 1997 Meeting of the "Department of Defense Breast Cancer Research Program: An Era of Hope", (Abstract) pp. 739-740, Washington, D.C., Oct. 31- Nov. 4, 1997). Comparison of the results from denaturing and non-denaturing conditions confirmed that the CA-PS-pool U was still heterogeneous and that the activity was most likely a subunit containing high molecular weight protein(s).

0500Removal of Storage Solution Components before Bioassay. Before conducting bioassays of the inhibitory activity in the phenyl-Sepharose pools, the glycerol and steroid hormones in the storage buffers were removed. If DHT is not removed completely from CA-PS-pool II, the inhibitory activity was substantially diminished or eliminated entirely. Samples (0.5 to 15 mL) were introduced into Slide-A-Lyzer® (Pierce) cassettes of molecular weight cutoff 10,000. The cassettes were incubate twice with stirring in two liters of Tris-HCl, pH 7.4, containing 10 mM CaCl₂ for four hours at 34°C to remove excess free steroids and glycerol. Next, the cassettes were transferred to the same buffer containing 20% (v/v) of a charcoal-dextran mixture prepared as described above. After 18 hours at 37°C, the cassettes were transferred to another two-liter volume of the same buffer containing 10% (v/v) of the charcoal-dextran mixture and dialysis continued with stirring for another 6 to 8 hours. Finally, the cassettes were rinsed lightly with water and the dialyzed material recovered according to manufacturers instructions. The contents were sterilized by 0.2-μm-pore membrane filtration and stored at 4°C. These preparations were usually used within a few weeks.

0501Assay of CA-PS-pool I Estrogen Reversible Inhibitory Activity with MTW9/PL2 Cells. When assayed with MTW9 PL2 cells, CA-PS-pool I contained 20 to 25% of the units of estrogen reversible inhibitory activity recovered from the phenyl Sepharose column (data not shown). With two preparations not presented, the cortisol gradient pool shown in Fig. 80 was made 1.5 M NaCl before application to the phenyl Sepharose column equilibrated at the same higher salt concentration. Under these conditions, the CA-PS-pool I contained > 90% CBG, as estimated by SDS-PAGE, but showed either no estrogen reversible activity or only traces (results not presented). Irrespective of the ionic strength or pH of the cortisol affinity pool applied to phenyl Sepharose, ethylene glycol was required to elute the majority of the activity.

0502Assay of CA-PS-pool H Estrogen Reversible Inhibitory Activity with Several ER⁺ Cell Lines. Despite method variations with phenyl Sepharose, CA-PS-pool π always contained > 75% of the activity recovered. Li a crucial test of significance, CA-PS-pool π was assayed to determine if it replaced the effects of CDE-serum with eight different ER⁺ cell lines. The results are shown in Fig. 82. The estrogen reversible inhibitory effects of CA-PS-pool π were investigated with five rodent tumor cell lines derived from three different estrogen target tissue tumors, and three separate estrogen sensitive human breast cancer cell lines. The cells were added to medium with 2.5% (v/v) CDE-horse serum plus increasing concentrations of CA-PS-pool II + 10 nM E^ The first lines evaluated were the GHi, GH₃, and GH₄C₁ rat pituitary tumor cells (Fig. 82A, 82B and 82C, respectively). They were chosen first because these lines are well known for hormone regulation of differentiated tissue specific functions in culture and exceptional sensitivity to a variety of hormones including estrogens (Tashjian AH Jr (1979) Methods Enzymol 58, 527-535; Haug E and Gautvik KM (1976) Endocrinology 99, 1482-1489; Haug E (1979) Endocrinology 104, 429-437; Amara JF and Dannies PS (1983) Endocrinology 112, 1141-1143). At lOμg/mL, CA-PS-pool π was fully inhibitory with all three GH lines. Growth was reduced to near seed density levels (i.e. < 0.5 CPD). By this measure, > 1, 700-fold increase in potency had been achieved versus full CDE-serum. The ED₅₀ with the GH cells was 6 to 8 μg/mL which was a 300 to 800-fold specific activity increase compared to full serum. E₂ reversed the effects of the CA-PS-pool II at every inhibitory concentration. CA-PS-pool π replaced the effects of full CDE-serum with these cells. Figs. 82D and 82E show similar experiments with the estrogen sensitive H301 hamster kidney tumor cells and the MTW9/PL2 rat mammary cells, respectively. CA-PS-pool II was most inhibitory at 15 μg/mL with both lines. The ED₅₀ were in the range of 5 to 10 μg mL. As with the GH lines, F^ completely reversed the effects of the inhibitor. Again, CA-PS-pool II replaced the effects of full CDE-serum with these cells. With human breast cancer cell lines MCF-7K, ZR-75-1 and T47D, the results were similar (Figs. 82F, 82G, and 82H, respectively). Addition of 10 to 15 μg/mL of CA-PS-pool JJ caused maximum inhibition. The ED₅₀ concentrations were 6 to 9 μg mL. As with ER⁺ rodent cell lines, E₂ completely reversed the inhibition caused by CA-PS-pool II. Again, CA-PS-pool II replaced the effects of full CDE-serum with these cells. Cortisol-agarose Affinity Removal of the Inhibitor from CDE-serum. Next it was determined if the cortisol affinity chromatography had not removed the majority of the activity from serum. To test this, three cell lines were analyzed with pre- and post cortisol column samples. Figs. 83A and 83B show the effect of a single column passage on the inhibitory activity for T47D human breast cells. The ED₅₀ of the pre-column CDE-serum was 7% (v/v). Post column, even 50% (v/v) serum did not achieve ED₅₀. Figs. 83C and 83D show the same studies with the GH₃ rat pituitary cells. Li this case, a single column passage completely depleted the activity. Complete depletion was also observed with the H-301 hamster kidney cell line (Figs. 83E and 83F).

0503Storage Conditions and SHBG Related Properties. At completion of the two-step isolation, the pools were stored in the presence of sufficient glycerol to prevent freezing at -20°C. Li experiments not shown, the estrogen reversible inhibitor was progressively less stable without addition of glycerol, calcium and/or steroid hormone. Dialysis against buffers without calcium is most definitely to be avoided. Freeze/thaw is very harmful, even with calcium and DHT present. Assays of -20°C glycerol stored CA-PS-pool II over a two year period indicated no decay in activity. Clearly, the storage conditions known to stabilize functional SHBG (Fernlund P and Laurell C-B (1981) J Steroid Biochem 14, 545-552; Rosner W et al. (1974) Biochim Biophys Ada 351, 92-98) also favored retention of estrogen reversible inhibitor activity in CA-PS-pool II.

0504Labeled Steroid Hormone Binding to CA-PS-pool I. CA-PS-pool I was determined to contain CBG by criteria cited above. Additionally, this pool was examined by Scatchard analysis for binding of tritium labeled steroid hormones. The results are summarized in TABLE 9. The association constants (K_a) of the labeled hormones showed the order cortisol > progesterone >» sex steroid hormones. The K_a of cortisol binding at 34°C was 1.41 x 10⁹ M^"1 that was equal to that of native rat CBG when analyzed at 4°C (Rosner W (1990) Endocr Rev 11, 80-91). However, it was higher than the K_a of 5.2 x 10⁷ M¹ for human CBG measured at 23°C (Rosner W and Bradlow HL (1971) J Gin Endocrinol Metab 33, 193-198). The binding characteristics of steroids to CBG from several species have been studied (Rosner W (1972) J Steroid Biochem 3, 531-542). The similarity of the results further supports the conclusion that CA-PS-pool I contains predominantly CBG. Labeled Steroid Hormone Binding to CA-PS-pool H. The estrogen reversible inhibitor activity in CDE-serum correlated with the binding of tritium labeled sex steroid hormones. This suggested a relationship between the estrogen reversible inhibitor and SHBG. However, the K_a for ³H-DHT binding to CDE-serum at 34°C was 3.90 x 10⁷ M^"1. However, it is important to note that this was at least 20 times lower than that of purified human SHBG at 0.99 x 10⁹ M^"1 for DHT or 2.2 x 10⁸ M^"1 for E₂ at 37°C (Rosner W and Smith RN (1975) Biochemistry 14, 4813-4820). To determine if CA-PS-pool II possessed the same sex hormone binding properties as whole CDE-serum, and/or human SHBG, the next study was conducted. Scatchard analysis of ³H-DHT binding to CA-PS-pool II was done at 34°C. The estimated K_a was 5.88 x 10⁷ M^"1. Replicates (N = 3) gave a K_a range 4.5- 10 x 10⁷ M^"1. Computer analysis indicated a single class of binding sites although correlation coefficients were approximately 0.7. Similar analyses were done with ³H-E₂, ³H-progesterone and ³H-cortisol. The results with all four labeled steroids are summarized in TABLE 9. The K_a order was DHT > Ε₂ >»cortisol > progesterone. The K_a for sex steroid hormone binding to the CA-PS- pool II was similar to whole CDE-serum but 20 to 50-fold lower than human SHBG.

TABLE 9

0506Summary of the Scatchard Analysis of phenyl-Sepharose pools I and H with four labeled steroid hormones

0507<img file="EP1328293A2_D0009.tif" />

0508Western immunoblotting with anti-human SHBG. The above shows that the estrogen reversible inhibitor shared immunological properties with human SHBG. To investigate further, Western immunoblotting of CA-PS-pool U was done with anti-human SHBG. The results are presented in Fig. 81B. Western analysis with the anti-SHBG recognized the same four components seen with Coomassie Blue staining in Fig. 81A. These same four components have also been identified with whole CDE-serum using Western analysis with anti-human SHBG (data not shown). Li Western immunoblotting studies not presented, anti-human SHBG did not identify horse serum albumin. This confirmed that the 67 kDa Coomassie Blue stained component present in the CA-PS- pool II was not 68kDa horse serum albumin. These results provided additional support for the conclusion that albumin is not the estrogen reversible inhibitor activity of serum. These results also very clearly demonstrated that the SHBG used to raise antibodies in rabbit had not been purified to homogeneity, but rather had been used at a more crude state. (It was also confirmed by the manufacturer of the anti-SHBG antibody that the SHBG fraction used for antibody production was not highly purified and had not been size fractionated.)

0509Discussion of Example 18. There has been one very critical problem with the estrocolyone hypothesis. Estrocolyone has never been purified and shown to act as described (Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52). The active pool isolated from the two-step procedure (i.e. CA-PS-pool S) certainly does not bind steroid hormones with sufficient affinity to act as estrocolyones (TABLE 9). Growth is activated at picomolar concentrations while the affinity (Kd) of E₂ with CA-Pool II is about 10^"8 M. This discrepancy is simply far too large to accept the role of estrogens in growth as binding the inhibitor and thereby preventing its action on target cells (Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52). The fact that proteins in CA-PS-pool II bind steroids is not germane to the mechanism of action of these hormones in growth regulation under physiological conditions.

0510The results of steroid hormone binding may however be germane to the use of high dose treatments of breast cancer. Care must be taken when considering that high doses of estrogen, androgen, progesterone and cortisol all have the potential for binding the active agent in CA-PS-pool II and therefore may reduce the effective concentration of inhibitor. The assays described in this Example can be applied to biological fluids and plasma to determine if steroid concentrations are excessive and to evaluate proper levels with changes in treatment regimes.

0511The results presented herein indicate that the proposed new model of cell growth is a favored mechanism. Steroid hormones appear to act as positive agents via internal high affinity receptors (e.g. ERγ) whereas serum-bome inhibitors act at the surface to block growth. The combination of the two signals dictates cell proliferation rates. This data further supports the assertion that the ERγ can be used for diagnostic purposes in ER⁺ cancers in the same way that conventional ER receptor screening is now performed. A highly enriched fraction of serum protein was prepared whose estrogen reversible inhibitory activity is stable and whose effects replicate those seen with full serum with a variety of sex steroid hormone target tumor cell types in culture. Because early studies mistakenly indicated that the inhibitor shared various properties with SHBG, a two-step cortisol-agarose affinity and phenyl-Sepharose chromatography protocol was applied. A highly enriched "SHBG-like" preparation was obtained. At 10 to 15 μg/mL, it replicated the E₂ reversible inhibition caused by 30 to 50% (v/v) serum with steroid responsive human breast cancer cells, and responsive rat mammary, rat pituitary and Syrian hamster kidney tumor cells in culture. The inhibitor retained full activity for more than one year when stored unfrozen at -20 °C in the presence of calcium, dihydrotestosterone and glycerol. This study demonstrated that the longstanding problem of inhibitor stability has been overcome and that a high specific activity preparation was now available to further probe molecular identity. These results clearly differentiate this inhibitor preparation from any previously described type of estrogen reversible inhibitor (i.e. estrocolyone). Moreover, no previous inhibitor composition, at a concentration < 15 μg/mL, can supplant the effects of full serum to give estrogenic effects > 3 CPD with several ER⁺ cell lines from different tissues and different species. The most active inhibitor preparation obtained in this study appeared to have multiple components present. The separation and identification of these components, as discussed in Example 20, would yield additional assays and preferred reagents and methodologies for testing new hormone-like and anti-hormone like substances. The active serum-derived inhibitor fraction can be used directly in tests of new compounds, substances, mixtures and preparations from natural and synthetic sources to estimate both estrogenic and androgenic activity in culture. Large-scale preparation of this purified serum fraction is possible by using larger affinity columns and proportionately increased serum volumes, similar to existing technology employed for purifying other biological products. It is advantageous that only small quantities of the purified serum fraction are needed for cell growth assays.

0512Example 19: Serum-free Assay Systems for Measuring Large Magnitude Steroid Hormone Mitogenic Responses with the Two-Step Purified Inhibitor.

0513The above-described studies with several different sex steroid sensitive cell lines demonstrated that the effects of a partially purified estrogen reversible inhibitor could readily be assayed in the presence of a low concentration (i.e. 2.5%) of CDE-serum. The next step was to eliminate the serum completely and to show estrogen responsiveness under far more defined conditions.

0514Second Analysis of Serum-free Growth ± E2. Experiments were conducted using completely serum-free medium, and the magnitude of the estrogenic effects observed in defined medium was again compared to those seen in medium containing CDE-serum. ER⁺ tumor cell growth was measured first in serum-free defined culture + 10 nM E₂. Similar experiments have been reported in Figs. 56 and 57. The reassays are included because the first experiments were done two years earlier. The results show the stability of the cell lines used and the fact that serum-free defined medium is highly reproducible. More recent results are shown with the MCF-7K human breast cancer cells (Fig. 84A), the T47D human breast cancer cells (Fig. 84B), the GH₄C₁ rat pituitary tumor cells (Fig. 84C), and the H301 Syrian hamster kidney tumor cells (Fig. 84D). All four-cell lines grew logarithmically for several days in defined and reached densities of 0.5 to 1.0 x 10⁶ cells per 35-mm dish. The media formulations were based on standard D-MEM/F-12 as described in TABLE 7. Growth rates were optimized to 70% or more of D-MEM/F-12 containing 10% (v/v) fetal bovine serum. The results presented in Fig.84 show little or no E₂ effect on growth in defined medium. Barnes and Sato (Barnes D and Sato G (1980) Nature (Lond) 281, 388-389) have reported similar negative results with another strain of MCF-7 cells in a different formulation of defined medium. Considering the variety of cell types assayed herein, the present results and the results of others, the lack of estrogenic effects in serum-free defined medium is not related to chemical composition of any one medium nor is there a major problem with time dependent variation of cell line properties.

0515Effects of CDE-Serum on ER⁺ Cells in Different Formulations of Serum-free Defined Medium. The experiments in Fig. 85 were done to show that serum could be added to different formulations of defined medium and still cause estrogen reversible inhibition. Effects are shown with CDE-horse serum + 10 nM - and T47D cells DDM-2MF (Fig. 85A), MTW9/PL2 cells in DDM-2A (Fig. 85B) and GKL cells in PCM-9 (Fig. 85C). Definitely, the serum-bome inhibitor(s) was fully effective in three different formulations of defined medium and with three different estrogen target tissue cell types.

0516Effects of CA-PS-pool H on ER⁺ Cell Growth in Serum-free Defined Medium. The estrogen reversible inhibitory effects of CA-PS-pool JJ were examined with eight ER⁺ cell lines growing in different serum-free defined media (Fig. 86). The cell lines were the MCF-7K cells (Fig. 86A), the T47D cells (Fig. 86B), the ZR-75-1 human breast cancer cells (ATCC) (Fig. 86C), the GHj (ATCC) (Fig. 86D), GH₃ (ATCC) (Fig. 86E), and GH (Fig. 86F) rat pituitary tumor cells, the MTW9/PL2 rat mammary tumor cells (Fig. 86G), and the H301 Syrian hamster kidney tumor cells (Fig. 86H). At 20 to 30 μg/mL, this fraction completely inhibited growth. The inhibition was totally reversed by 10 nM E₂. The E^ effects on cell number were in the range from 33 to 72-fold (i.e. CPD = 2 ' to 2 ' ). The activity was not replaced by serum albumin at 5 mg/mL (data not shown). The estrogen mitogenic effects seen in defined medium containing only a few μg/mL of protein were equal to or greater than those seen in medium containing 30 to 50% (v/v) CDE-horse serum with every ER⁺ cell line tested (TABLE 10). Plainly, the serum-free conditions established herein are the most defined model assay systems yet established to demonstrate estrogen responsiveness in vitro.

TABLE 10

0518Summary of the Maximum Estrogenic Effects in D-MEM/F-12 plus CDE-horse Serum

0519+ 10 nM E2 versus those in Serum-free Defined Medium Supplemented with CA-PS-pool H

0520<img file="EP1328293A2_D0010.tif" /><img file="EP1328293A2_D0011.tif" />

0521CPD i?X" = Fold Cell Number Increases Above Controls Without Estrogen)

0522Discussion of Example 19. The studies presented in Fig. 86 and TABLE 10 summarized unequivocally, and for the very first time, that large magnitude estrogen mitogenic responses can be observed in completely serum-free defined media containing 2 mg/mL total protein. Furthermore, the responses shown in Fig. 86 either equal or exceed others previously observed in partially serum- free media with ZR-75-1 human breast cancer cells (Allegra JC and Lippman ME (1978) Cancer Res 38, 3823-3829; Darbe PD et al. (1984) Cancer Res 44, 2790-2793) or with a variety of other estrogen sensitive (ER⁴) human and rodent cell lines in medium with hormone depleted or deficient serum (Amara JF and Dannies PS (1983) Endocrinology 112, 1141-1143; Natoli C et al. (1983) Breast Cancer Res Treat 3, 23-32; Soto AM et al. (1986) Cancer Res 46, 2271-2275; Wiese TE et al. (1992) In Vitro Cell Dev Biol 28 A, 595-602).

0523These results have a number of important implications. First, they support the aspect of the estrocolyone hypothesis (Soto AM and Sonnenschein C (1987) Endocr Rev 8, 44-52) that relates to the presence in serum of a meaningful inhibitor(s). Also, there is no doubt that the inbibitor(s) is completely estrogen reversible. However, the present experiments do not confirm that the steroid hormones interact with sufficient affinity with the inhibitor to support that aspect of the estrocolyone hypothesis. The results in TABLE 9 indicate that this aspect of the estrocolyone hypothesis is highly unlikely. The estrogen reversibility of the inhibitor with every target cell type studied under the rigorous conditions of serum-free defined culture suggests physiologic relevance. The large magnitude of the effects is a strong statement in favor of significance. This is especially clear when considering the fact that the first experiments with 30 to 50% (v/v) serum contained 15 to 25 mg/mL of protein, whereas the later tests using serum-free medium required only 20 μg/mL of isolated protein.

0524The active fraction isolated from horse serum represented only 0.01 to 0.04% (w/w) of the total protein. Nonetheless, it effectively regulated eight ER⁺ cell lines derived from three species and three different target tissues. These observations are evidence that a broadly applicable serum fraction has been identified. Furthermore, the serum-free medium results suggest that a common agent(s) may coordinately regulate estrogen responsive tissue growth in vivo and that the concept of estrogen reversible negative control may be far-reaching. The results support the conclusion that in vitro studies can be used to identify important new aspects of in vivo endocrine physiology.

0525The results in defined medium have practical applications. Cells in serum-free medium grow in response to nutrients, growth factors, metal delivery proteins, adhesion proteins, and various classes of hormones. All of these components are mitogenic in the sense that they contribute to cell replication. Nonetheless, the addition of only 20 μg/mL of inhibitor to block growth completely bears directly on the question of the progression of normal steroid target cells to fully hormone autonomous cancers.

0526The inhibitor preparation used herein has the properties of a family of tissue regulators first named "chalones". These proposed cell regulators are water-soluble and tissue specific (but not species specific) proliferation inhibitors that are reversible by physiologic stimuli including hormones (Bullough WS (1975) Life Sci 16, 323-330; Finkler N and Acker P (1978) Mt Sinai JMed 45, 258-264). The studies presented here support this classic concept as it applies to sex steroid hormone target tissues. As further demonstrated in subsequent Examples, the molecular identification of the serum inhibitors) promises not only to further support the role of estrogens as "necessary", but also to establish that "chalone-like" entities likely are the missing "sufficient" components that account for estrogen regulation of tissue growth. The application of serum-free defined medium conditions along with the use of a high specific activity fraction to demonstrate estrogen responsiveness in culture is unique.

0527Example 20: Chemical and Immunological Properties of the Partially Purified CA-PS-Pool H

0528Inhibitors and Identification as IgA and IgM

0529This Example describes chemical and physical confirmation that the sought-after serum-bome cancer cell growth inhibitors) include at least IgA and IgM.

0530Proteinase and Chemical Fragmentation followed by HPLC and Amino Acid Sequencing. Although it was clear from SDS-PAGE (Fig. 81A) that the CA-PS-pool π preparation was not homogeneous, chemical fragmentation with cyanogen bromide and proteolytic enzymes was used for protein/peptide sequencing in an attempt to identify at least some of the proteins present, employing standard techniques (Work TS and Burton RH, eds (1981) Laboratory Techniques in Biochemistry and Molecular Biology, Volume 9, G. Allen, Sequencing of Proteins and Peptides, Elsevier/North-Holland, Amsterdam, pp 43-71). This set of protocols yielded many peptides. These were applied to reverse phase HPLC columns in trifluoroacetic acid and eluted with organic solvents as described (Ogasawara M et al. (1989) Biochemistry 28, 2710-2721). The separated peptides were sequenced. The analyses of several peptides are shown in Fig. 87. Attempts were made to align these sequences with human SHBG (hm SHBG) (Walsh KA et al. (1986) Biochemistry 25, 7584- 7590), rabbit SHBG (rb SHBG) (Griffin PR et al. (1989) J Biol Chem 264, 19066-19075), rat androgen binding protein (rt ABP) (GENBANK registration) and a partial sequence of hamster androgen binding protein (hs ABP) (GENBANK registration). A BLAST SEARCH match showed &lt; 30%) homology for the peptides sequenced. If the material in CA-PS-pool II had significant shared primary structure with SHBG, it could not be confirmed. Antibodies Against the CA-PS-Pool H Components. Preparative SDS-PAGE was done on the CA-PS-pool π fraction, and after localization of the 54kDa band, the 54kDa band was eluted and prepared for rabbit antibody production by HTI (Ramona, CA). The antibodies raised were very potent and reacted with CA-PS-pool π (Fig. 88). They did not cross react with CBG (CA-PS-pool I). However, despite great care, it was evident that the anti-54 kDa was raised against a mixture of 67, 58 and 54 kDa subunits (Fig. 88). The reaction was definitely strongest with the 54 kDa component, but clearly identifiable with the 67 kDa and 58 kDa bands as well. This apparent problem turned out to be an advantage, and allowed positive identification of the active agents in CA-PS-pool π. It was investigated whether the activity in CA-PS-pool II might have been isolated because of affinity for the agarose matrix rather than as a consequence of the steroid hormone ligand attached to agarose, noting from interpretation of unrelated studies, that agarose alone can bind immunoglobulins and give SDS-PAGE bands at 67, 58 and 54 kDa. Therefore, it was thought possible that IgG was the estrogen reversible inhibitor.

0531Antibodies Against the 54 kDa Component of CA-PS-Pool H and Blocking of the Estrogen Reversible Inhibitor Activity. Based on the results in Fig. 88, it was apparent that the 54 kDa antiserum might be used to determine if the biological activity resided in any of the 67, 58 or 54 kDa bands. The next study was done to resolve this important issue. The results were pivotal. Fig. 89 shows that the purified material in CA-PS-pool II was completely inhibitory at 20 to 40 μg/mL. Addition of even a 1:5000 dilution of anti-54kDa blocked the effect of the inhibitor. Li control studies, rabbit pre-immune serum had no effect even at 1:100 a dilution (data not shown). It was evident that anti-54 kDa serum contained the antibody to the activity.

0532Anti-54 kDa Serum Recognizes Authentic Horse IgA, IgM and IgG. Next, authentic horse IgA was obtained from Accurate Chemicals, and horse IgM was obtained from Accurate Chemicals and Custom Monoclonal Litemational. The material from Custom Monoclonals was custom purified by an affinity method with a monoclonal antibody against horse IgM Fc and further purified by molecular sieve chromatography to be sure of elimination of other immunoglobulins (a common problem). IgGs were obtained from Zymed (San Francisco, CA), Sigma (St. Louis, MO) or The Binding Site (San Diego, CA). The Western analysis shown in Fig. 90 demonstrates these results. The results show clear cross-reaction with 67kDa IgM heavy chain, 58 kDa IgA heavy chain and 54 kDa IgG heavy chain but no reaction with horse albumin. Assay of Estrogenic Effects Controlled by Commercially Purchased Horse IgG, IgA and IgM in 2.5% CDE-horse Serum with MTW9/PL2 Cells. Fig. 91 demonstrates that at concentrations up to 59 μg/mL, horse IgG did not cause inhibition of MTW9/PL2 cell growth in 2.5% CDE-horse serum. There was no significant estrogenic effect caused by IgG. Fig. 92 shows very clearly that commercially prepared horse serum derived IgM (Custom Monoclonals), was very active. At concentrations of 20 to 50 μg/mL, IgM completely inhibited the growth of the MTW9/PL2 cells (i.e. &lt; 1.0 CPD). Addition of 10 nM Ea reversed the inhibition nearly completely. Estrogenic effects of 4 to 5 CPD were seen (Fig. 92). Fig. 93 shows the same general results with commercially prepared horse serum derived IgA (Accurate). The only apparent difference was that IgA was slightly more effective than IgM. These results clearly proved that the active components in CA-PS-pool II were IgA and IgM. That these Igs would prove to be the inhibitor was completely unexpected. Although these two active classes of immunoglobulins (IgA and IgM) are well- established secretory products of normal breast cells, there was no previous suggestion in the prior art that they play a role in the negative regulation of estrogen-dependent cell growth. These immunoglobulins are major proteins in milk whose hormone-related local production in breast tissue is well documented, and their function in the body's secretory immune system is well known.

0533Alternate Methods of Obtaining Horse Serum IgG, IgM and IgA. IgG can be purified using a Hytrap matrix, which is a mixture of immobilized Protein A and Protein G, employing a technique described by others (Lindmark R et al. (1983) J. Immunol Meth 62, 1-13; Kronvall G et al. (1969) J Immunol 103, 828-833; Akerstrom B et al. (1986) JBiol Chem 261, 10240-10247). IgM can be obtained using a mannan binding protein isolation method normally applied with human serum (Nevens JR et al. (1992) J Chrom 597, 247-256). However, yields are low. Another method based on anti-IgM immunoglobulins linked covalently to Sepharose is far more effective. This same procedure with immobilized anti-IgA immunoglobulins can be used to isolate IgA (Tharakan J Li: Antibody Techniques, Malik VS &amp; Lillehoj EP, Eds, 1994, Academic Press, San Diego, CA, Chapter 15). Horse IgA can also be purified using an immobilized Jacalin lectin method usually reserved for human samples (Roque-Barreira MC et al. (1986) Braz J Med Biol Res 19, 149-157). However, it can be modified for non-human species. The buffers are modified to contain 10 to 50 mM CaCl₂ to bind IgA from other species. Even then, yields are not high. The preferred methods for horse IgA and IgM use immobilized antibodies. , Purification of Rat Serum Immunoglobulins. Three isolations of the estrogen reversible inhibitor from separate one-liter batches of adult rat serum were conducted. This was done for two important reasons. First, the estrogen reversible activity in all types of adult serum, including rat, were assayed with a highly estrogen sensitive MTW9/PL2 rat mammary tumor cell line. It was useful to corrfirm the horse serum purification results with a homologous experimental system. Second, the confirmation that rat IgA and IgM regulated rat mammary tumor cell growth would open the possibility of combined testing of new therapeutic substances both in vitro and in vivo. To summarize, the same "CBG" and "SHBG" fractions were obtained from rat serum by the methods of Fernlund &amp; Laurell as had been obtained from horse serum. The chromatography profiles of the rat separations (not presented) were very similar to those presented in Fig. 80. The only major difference was that no rat CBG was obtained. At pH 5.5, rat CBG did not significantly bind to the affinity matrix. Rat serum CA-PS-pool I and CA-PS-pool π both contained only two Coomassie Blue stained bands when analyzed by SDS-PAGE (Fig. 94A). These were approximately 55kDa and 54 kDa. They were somewhat lower molecular weights than found with horse, and there were fewer bands. To test if either rat band was IgG, a Western analysis was performed with rabbit anti-rat IgG (Fig. 94B). The antibody did not recognize the Coomassie stained bands but did react with control IgG. However, when examined with very specific heavy chain monoclonal antibodies raised to rat IgGl, IgA, and IgM (purchased from Zymed), the Western analysis was clear (Fig. 95). Both the commercially purified rat immunoglobulins (purchased from Zymed) and the two-step purified pools showed cross-reaction with anti-IgA (weakly), anti-IgGl subtype (strong reaction) and anti-IgM (moderate reaction) (Figs.95A, 95B and 95C, respectively). Rat and Horse Serum Active Pools Isolated by the Two-Step Procedure of Fernlund and Laurell have the same Classes of Immunoglobulins. The same classes of immunoglobulins obtained by the two-step procedure of Fernlund and Laurell (Femlund P and Laurell C -B (1981) J Steroid Biochem 14, 545-552) with horse serum were found when rat serum was the starting material. This was considered to be further confirmation that binding to the agarose matrix was more important than to the immobilized cortisol. It should be noted that in the original Femlund and Laurell report using human cord serum does not address possible immunoglobulin contamination, however (Femlund P and Laurell C -B (1981) J Steroid Biochem 14, 545-552). This is particularly curious because human immunoglobulins bind to agarose (Smith RL and Griffin CA (1985) Thombosis Res 37, 91-101). Amino Acid Sequencing Of Rat "SHBG-like" Proteins. Protein fragmentation and amino acid sequencing of rat "SHBG-like" proteins were done as described above for horse CA-PS- pool U. The analyses of several peptides are shown in Fig. 96. Attempts were made to align these sequences with human SHBG (hm SHBG) (Walsh KA et al. (1986) Biochemistry 25, 7584-7590), rabbit SHBG (rb SHBG) (Griffin PR et al. (1989) J Biol Chem 264, 19066-19075), rat androgen binding protein (rt ABP) (GENBANK registration) and a partial sequence of hamster androgen binding protein (hs ABP) (GENBANK registration). A BLAST SEARCH match showed ≤ 30% homology for the peptides sequenced. If the rat "SHBG-like" pools have significant shared primary structure with SHBG or rat androgen binding protein, it could not be confirmed by these studies.

0534Labeled Steroid Hormone Binding to The "SHBG-like" Pools from Rat Serum. As described in TABLE 9, CA-PS-pool π from horse serum binds sex steroids with an affinity of about

053510^"8 M. This same Scatchard analysis was done with an active fraction from rat serum. TABLE 11 shows the results of these studies with four labeled steroid hormones. It is clear that sex steroid hormones bind with a higher affinity than progesterone or cortisol. The binding affinities of rat and horse preparations were very similar. Li both cases, the affinities tend to rule out the estrocolyone hypothesis because it requires E₂ binding in the picomolar range.

TABLE 11

0537Summary of the Scatchard Analysis of the "SHBG-like" Pools from Rat Serum with Labeled Steroid Hormones

0538<img file="EP1328293A2_D0012.tif" />

0539Evaluation of the Rabbit Anti-SHBG Cross-Reaction with the Active Pools from the Two-Step Isolation of Fernlund and Laurell. As shown above in Fig. 81B, Western analysis with the anti-SHBG detected horse IgA, IgM and IgG. Additionally, anti-SHBG immunoprecipitated the estrogenic activity of horse serum (Fig. 79B). To extend these results, it was established that rabbit anti-human SHBG recognized a number of the major classes and subclasses of rat immunoglobulins. SDS-PAGE with Coomassie blue staining (Fig. 97A) was compared to identification of the same proteins by Western analysis with anti-SHBG (Fig. 97B). These results leave very little doubt that the SHBG used to raise antibodies in rabbits was not homogeneous but in fact was a "crude" preparation contaminated with several immunoglobulins.

0540Test of Rat IgG, IgA and IgM for Estrogen Reversible Inhibitory Activity with MTW9/PL2 Rat Mammary Tumor Cells. All of the rat immunoglobulins described in this section were purchased from Zymed as the highest quality available. Their activity was assessed with MTW9/PL2 cells in 2.5% (v/v) CDE-rat serum, as described above. The activity of rat IgG (all subclasses combined) was assessed (Fig. 98). There was no inhibitory effect at up to 50μg/mL. Rat IgA was a potent estrogen reversible inhibitor (Fig. 99). At 20 to 50 μg/mL, it completely inhibited growth. Addition of 10 nM E₂ completely reversed the inhibition. The estrogenic effects recorded were > 5 CPD. The results with rat IgM were very similar (Fig. 100). At 20 to 50 μg/mL, it completely inhibited growth. Addition of 10 nM E₂ reversed the inhibition. The estrogenic effects recorded were > 5 CPD. It is essential to note that IgA or IgM replaced the effect of full CDE-rat serum with MTW9/PL2 cells. With a completely homologous system (i.e. cell line, basal 2.5% CDE- serum, and immunoglobulins), the results were clear. IgA and IgM were the sought after serum- bome inhibitors from rat. Discussion of Example 20. The identification of IgA and IgM as serum-bome inhibitors fully separates these inhibitors from the teachings of U.S. Patent Nos. 4,859,585 (Sonnenschein) and 5,135,849 (Soto), which arrived at no molecular identification of the inhibitor. The series of investigations presented above demonstrate that a very longstanding problem has been solved. While the solution is significant, an even more an important consequence of this knowledge is the fact that for the very first time, mucosal cell hormone dependent growth has been linked to a natural immune regulation. Moreover, this information has direct application to the diagnosis, genetic screening, prevention and therapy of breast and prostate cancer and a high likelihood of applications to other mucosal cancers, as described in more detail in U.S. Patent App. No. (Atty. Dkt. No. 1944-00800)/PCT/US2001/ (Atty. Dkt. No. 1944-00801) entitled "Compositions and

0541Methods for the Diagnosis, Treatment and Prevention of Steroid Hormone Responsive Cancers," which is hereby incorporated herein by reference..

0542During the purification of both the horse serum and the rat serum estrogen reversible activity, SUPERDEX™ (Pharmacia) molecular sieve chromatography of the final mixtures indicated the presence of < 20% 160 kDa monomeric immunoglobulins. The majority of the material was of much larger mass. Because IgA exists naturally as monomer, dimer and polymers, there was a question concerning which of these is/are irihibitory form(s). The SUPERDEX™ results strongly favor the dimer/polymer form. This was confirmed also with commercially prepared IgA that was obtained from hybridoma and myeloma cell lines. The IgA from these was > 80% dimer/polymer. It was very active as an inhibitor. Li light of these results, it is suggested that these forms are the "good" type of IgA in the body, and that direct measurement of their concentration in plasma and body fluids has diagnostic and prognostic applications.

0543The introduction of test methods done with minimum serum plus purified immunoglobulin inhibitor ("spiked serum") provides a new approach to substances, mixtures and compounds that might be influenced by serum components. For example, a serum composition might contain steroid hormone free serum, such as a standard, commercially available fetal bovine serum preparation, and a predetermined amount of an immunoglobulin inhibitor, i.e., one or more of IgA, IgM or IgG. Testing under these conditions, with a known amount of inhibitor in the serum, may be desirable or required when the substance has potential for inactivation/activation by a serum component or when it has lipophilic properties that require a minimum protein concentration in the medium to prevent loss.

0544Another valuable application of the immunoglobulin inhibitors will be in identifying substances that may have direct effects on the action of the immunoglobulins to cause inactivation. An assay of this nature is unique in the sense that incubation of substances with the immunoglobulin can be done before the assay to determine effects on natural immune responses. Changes in environmental/chemical factors that affect the body's immune system are of major medical concern. They also are of great concern to veterinary medicine. Chemicals/nutritional supplements may affect immune function of domestic animals and thereby affect human food supplies.

0545This series of investigations demonstrate at least two immunoglobulin inhibitors in serum. More than one inhibitor was suggested by the conventional purification data in a preceding Example, and was proved true in succeeding examples. There may still be other useful estrogen reversible immunoglobulin inhibitors in serum that are yet to be identified from serum or tissue sources. The methods described in this Example have direct application to the search for new compounds that mimic the effect of the immunoglobulins as estrogen reversible inhibitors. Such application opens a new avenue of search for anticancer drugs.

0546Example 21: Regulation of Steroid Hormone-responsive and Thyroid Hormone-responsive Cancer Cell Growth in Serum-free Defined Medium by Secretory and Plasma Forms of IgA and Plasma and Cell Culture Derived IgM

0547This Example demonstrates that the determination of whether purified IgA and IgM from several species mimic the sex steroid hormone reversible inhibitors isolated from horse in serum was sought. These studies included ER⁺ tumor cells derived from rodents as well ER⁺ and AR⁺ cells from human cancers. Completely serum-free defined culture conditions were used to perform cell growth assays using the purified inhibitors. The total protein concentration in the media was &lt; 2 mg/mL. The estrogenic and androgenic effects observed in these assays are unique, as like effects have not been achieved previously in completely serum-free defined medium.

0548Sources of Purified IgA and IgM. Human IgM was purified from human plasma as described using immobilized mannan-binding protein (Nevens JR et al. (1992) J Chromatography 597, 247-256). As an example of the effectiveness of this isolation, Fig. 101 shows SDS-PAGE and Coomassie Blue Staining with two preparations of human plasma IgM prepared. Human IgAl and IgA2 were purified using immobilized Jacalin (Roque-Barreira MC and Campos-Neto A (1985) J Immunol 134, 1740-1743; Kondoh H et al. (1986) J Immunol Methods 88, 171-173; Pack TD (1999) American Biotechnology Laboratory 17, 16-19; Loomes LM et al. (1991) J Immunol Methods 141, 209-218). Rat IgA and IgM were purchased from Zymed. The effectiveness of the Jacalin method with human plasma is shown in Fig. 102. Horse IgA and IgM were purchased from Accurate, Sigma and Custom Monoclonals. IgA and IgM from other species or as products from cell culture are purchased from Sigma or Accurate. Human IgA and IgM were bought also from Sigma and Accurate. Human secretory (milk) IgA (slgA) was purchased from Sigma or Accurate.

0549A. MTW9/PL2 rat mammary tumor cells. For this series of experiments the serum- free defined medium was the preferred formulation of DDM-2A described in TABLE 7. The cell growth assays with this cell line in DDM-2A testing increasing concentrations of human plasma IgM is shown in Fig. 103. Human plasma IgM completely inhibited growth by 20 to 60 μg/mL. _.

0550The ED₅₀ was about 12 μg/mL. Based on an IgM M_r of 950,000, the ED₅₀ concentration was 1.3 x 10^"8 M. Complete inhibition was at 2.2 x 10^"8 M. These concentrations are certainly within the physiological range of IgM in the plasma and body fluids such as breast milk. Based on these studies, a comparison was done in completely serum-free defined DDM-2A medium of the effects of 40 μg/mL of rat plasma IgA + E₂, rat plasma IgM ± E₂, and horse plasma IgM + E₂ (Fig. 104, expressed as (A) cell numbers and (B) CPD). From the CPD calculations it was clear that no matter the species source, IgA and IgM were very potent estrogen reversible inhibitors of MTW9/PL2 cell growth.

0551One problem occurred with the MTW9/PL2 cell assays that initially caused concern. Human IgA was purchased from Sigma as the milk derived immunoglobulin. It was far less expensive than plasma IgA. For reasons that at first were not clear, this material was at best only partially inhibitory and often not inhibitory. As will be discussed below with GHj cells, this turned out to be a significant clue to the mechanism of action of the immunoglobulins. Nonetheless, it is known that the heavy chains of IgM and IgA from different species share primary structure homology. This is not true of the variable regions of the light chains. The results presented support the possibility of Fc-like receptor mediation of the IgA and IgM effects on MTW9/PL2 cells.

0552B. GHi, GH₃ and GH₄C₁ rat pituitary tumor cells. For this series of experiments the serum-free defined medium was the preferred formulation of PCM-9 described in TABLE 7. The next serum-free defined medium studies were done with GH] cells. Example assays are shown. This cell line was highly estrogen responsive in the presence of homologous rat myeloma derived IgA (Fig. 105). Maximum estrogenic effect was > 5 CPD or more than a 32-fold estrogen-induced increase in cell number in 10 days. A similar assay with human plasma derived IgA showed nearly the same results (Fig. 106). Indeed, human IgA showed greater inhibition at 10 μg/mL. Another study with human IgM demonstrated that it was also an estrogen reversible inhibitor of GHi cell growth (Fig. 107). It was not as inhibitory as IgA with this cell line, but certainly still effective. As discussed above, in the Background of the Invention, during the secretion process a fragment of about 80% of the poly-Ig receptor (including the five extracellular domains) becomes attached to the dimeric/polymeric form of IgA to form secretory IgA or slgA. The receptor fragment is called the "secretory component". After secretion, slgA can be readily isolated from human milk. The effect of milk derived secretory IgA (slgA) was evaluated with the GH_! cells in PCM-9, and the results of a representative study are shown in Fig. 108. These results were strikingly different than those obtained with plasma derived IgA (plgA) (Fig. 106). SlgA was not inhibitory even at 20 μg/mL. Considering why the two different forms of IgA behaved so differently, the poly-Ig receptor was recognized as a potential candidate for the mediator of the action of IgA IgM. As discussed in the Introduction of this Detailed Description, this receptor has not been previously associated with any growth related function. The poly-Ig receptor is concerned with process of transcytosis of IgA/IgM, as conceptually illustrated in Fig. 109. SlgA already has the receptor bound in the sense of the secretory piece in association with the Fc domains of the dimer. Fig. 110 illustrates schematically the structures of inactive monomeric IgA, the connecting or joining "J" chain, the structure of the active dimer with "J" chain, the secretory piece or secretory component, and the dimeric IgA structure plus secretory component attached, as generally understood. The illustration shows that the Fc domains of dimeric IgA are blocked by the secretory piece/component. Access to the Fc domains is required for binding to the poly-Ig receptor.

0553The present series of cell growth assays above were continued with the related GH₃ cells, again in serum-free defined the preferred formulation of PCM-9 medium. Rat myeloma derived IgA was an effective estrogen reversible inhibitor of these cells in a 9 day growth assay (Fig. 111). The maximum estrogenic effect exceeded 5 CPD. A similar assay with rat IgM was conducted (Fig. 112). It showed even greater inhibition at 10 μg/mL than with IgA. The estrogenic effect recorded in 10 days was nearly 6 CPD. These same assays were next repeated with the human immunoglobulins. Human plgA was an estrogen reversible inhibitor of GH₃ cell growth (Fig. 113). It was not as effective as its rat counterpart, but the estrogenic effect with the human immunoglobulin was still 4 CPD. Also, human IgM was effective with GH₃ cells (Fig. 114). Again the estrogenic effect was about 4 CPD. In the final study with GH₃ cells, it was again apparent that human milk derived slgA was not inhibitory (Fig. 115). The studies above with GHi and GH₃ cells were continued with the related GBLA line, again in serum-free defined PCM-9 medium. Rat myeloma derived IgA was an effective estrogen reversible inhibitor of these cells in a 9 day growth assay (Fig. 116). The maximum estrogenic effect approached 5 CPD. A similar assay with rat plasma IgM was conducted (Fig. 117). It showed slightly less inhibition than IgA. The estrogenic effect recorded in 10 days was nearly 4 CPD. These same assays were next repeated with the human immunoglobulins. Human plgA was an estrogen reversible inhibitor of GELA cell growth (Fig. 118). It was not as effective as its rat counterpart, but the estrogenic effect with the human immunoglobulin was still almost 4 CPD. Also, human plgM was effective with GHA cells (Fig. 119). The estrogenic effect was about 5 CPD. In the final study with H cells it was again apparent that human milk derived slgA was not inhibitory (Fig. 120).

0554C. H301 Syrian hamster kidney tumor cells. The studies with this cell line were done in the preferred formulation of CAPM defined medium described in TABLE 7. Because hamster IgA and IgM were not available, these experiments began with plasma IgA from mouse (Fig. 121). Mouse IgA was very effective with hamster H301 cells. The estrogenic effect was > 5 CPD. Human plasma IgA was also effective (Fig. 122A). The maximum estrogenic effect reached 4

0555CPD. Secretory IgA was inactive (Fig. 122B). With this cell line, human IgM also was an estrogen reversible inhibitor. As shown in Fig. 123, a dose-response study demonstrated that in serum-free defined medium with 40 μg/mL of human plasma IgM, concentrations of 0.1 to 1.0 picomolar E₂ caused significant growth (p &lt; 0.01). This data demonstrate the extraordinary sensitivity of the serum-free defined cell growth assays in the presence of immunoglobulin. The data in Fig. 123 provide strong support for the view that the H301 cells can be used to characterize the new ERγ and characterized in preceding Examples.

0556D. MCF-7A and MCF-7K human breast cancer cells. For this series of experiments the serum-free defined medium was the preferred formulation of DDM-2MF described in TABLE 7. Two highly applied MCF-7 human breast cancer cell strains were applicable to this series of investigations. As shown with MCF-7A cells in DDM-2MF serum-free defined medium, plasma IgA was highly effective as an estrogen reversible inhibitor. The estrogenic effect exceeded 4 CPD in 10 days (Fig. 124A). Li contrast, slgA was inactive (Fig. 124B). With the MCF-7K strain, the results were nearly identical. Plasma IgA was effective (Fig. 125A) and slgA was inactive (Fig. 125B). The estrogenic effects caused by plgA were replicated by substitution of plasma IgM. With MCF-7A and MCF-7K, plgM was an effective estrogen reversible sustaining estrogenic effects of > 4 CPD (Figs. 126 and 127, respectively). Li a final study of this series, an E₂ dose-response experiment was conducted with MCF-7K cells in DDM-2MF plus 40 μg/mL of plasma IgM. The results were remarkable. Estrogen at as low as 0.1 picomolar caused more than one-half maximum growth response (Fig. 128). The extraordinary sensitivity of this assay methodology is clearly established. These results add more evidence that a very high affinity estrogen receptor (t.e. ERγ) regulates growth and is yet to be defined in human breast cancer cells.

0557E. T47D human breast cancer cells. The T47D cell line was assayed for immunoglobulin effects in the preferred formulation of serum-free defined medium DDM-2MF described in TABLE 7. As shown in Fig. 129A, human plasma IgA was a very effective estrogen reversible inhibitor with T47D cells. The maximum estrogenic effect was 6 CPD or a 72-fold cell number increase in 12 days. In contrast, slgA was inactive at up to 20 μg/mL (Fig. 129B). Likewise, human plasma IgM is effective (Fig. 130), demonstrating complete inhibition of cell growth by 20 μg/mL IgM. The estrogenic effect was 5 CPD in 12 days. In experiments not shown, the effects of plasma derived IgM were compared to myeloma derived IgM. This study yielded the same estrogenic effects with both sources of IgM. Again, the antigenic determinant appears to be unimportant. The results support the view that the heavy chains dictate the activity. In other studies with T47D cells in defined medium containing 40 μg/mL, the dose-response effects with E₂ showed more than one-half maximum growth at 0.1 picomolar (Fig. 131). These results continue to fortify the theme that the methods described in this Example allow investigation of potential estrogenic compounds and substances that might be present in samples of industrial or biological materials at very low concentrations. It is also apparent that the data supports the view that a high affinity ERγ regulates growth. F. ZR-75-1 human breast cancer cells. For these experiments the serum-free medium was the preferred formulation of DDM-2MF described in TABLE 7. Plasma IgA was an estrogen reversible inhibitor with ZR-75-1 cells (Fig. 132A). The estrogenic effect was recorded at 5 CPD in 14 days. As seen before with the other ER⁺ cell lines above, slgA was not an inhibitor with ZR- 75-1 cells (Fig. 132B). Plasma IgM was also assayed with the ZR-75-1 cell's (Fig. 133). It was a potent estrogen reversible inhibitor under these completely serum-free defined conditions. As discussed above, this line had been thought to be estrogen responsive in serum-free culture. However, the former methods were not serum-free. As disclosed herein, it has now been established in entirely different culture conditions and shown that this line is truly estrogen growth responsive in culture.

0558G. HT-29 human colon cancer cells. For this series of experiments the serum-free defined medium was the preferred formulation of CAPM described in TABLE 7. As expected from endocrine physiology, colon is not a sex steroid hormone growth regulated tissue as are others such as breast, uterus, ovary and pituitary. However, it was discovered that this tissue is thyroid hormone growth responsive. As shown in Fig. 134, HT-29 human colonic carcinoma cells grow in CAPM independently of the presence of thyroid hormone. This growth is promoted by the other factors present in CAPM minus T₃. However addition of plasma IgM at 40 μg/mL had a dramatic effect. Li the absence of T₃ HT-29 cell growth was inhibited to &lt; 1.0 CPD in 10 days. Addition of increasing concentrations of T₃ restored growth (Fig. 134). This demonstrates that colonic cancer cells respond to thyroid hormones in the same manner that ER⁺ cells respond to E₂. Estrogens and thyroid hormones belong to the same superfamily of receptors and both are required for normal physiologic growth and development (Williams GR and Franklyn JA (1994) Baillieres Gin Endocinol Metab 8, 241-266; Tsai MJ and O'Malley BW (1994) Annu Rev Biochem 63, 451- 486). This is the first demonstration of a secretory immunoglobulin acting directly as a thyroid hormone reversible growth inhibitor of a human origin colon cancer cell line.

0559H. LNCaP human prostatic carcinoma cells. For this series of experiments the serum- free defined medium was the preferred formulation of CAPM described in TABLE 7. LNCaP cells were negatively regulated by plasma IgA (Fig. 135A). The immunoglobulin was a DHT reversible inhibitor that was completely effective at 10 μg/mL. The androgenic effect was > 5 CPD in 12 days. As seen with the ER⁺ cell lines above, slgA was not inhibitory with LNCaP cells (Fig. 135B). Two different types of human IgM were also compared with LNCaP cells (Fig. 136). They were plasma derived and myeloma derived IgM. Despite the differences in antigen binding domains, both forms were equally inhibitory and both forms were reversed by 10 nM DHT. These results indicate that the Fc/heavy chain of IgM is the functional activator of the inhibition. Summary of the estrogenic effects of IgM on ER⁺ cell growth. Fig. 137 presents a summary of the effects of IgM derived from different species with a variety of ER⁺ cell lines. This summary presents the maximum estrogenic effects recorded under conditions described above in serum-free defined medium with each cell line + 10 nM E₂. Estrogenic effects ranged from 4 to > 7 CPD. Comparison of the results in Fig. 137 with those in TABLE 10 show in general that the results achieved in completely defined medium are equal to or greater than those seen in CDE- serum cultures.

0560Discussion of Example 21. These methods will permit evaluation of industrial, environmental, biological, medical, veterinary medicine and other potential sources of estrogenic or androgenic activity under the most sensitive conditions yet developed. Estrogenic activity is measurable at < 1.0 picomolar concentrations. Two cell lines, MTW9/PL2 and H301, are preferred potential sources of identification of the new growth regulatory ERγ. The evidence presented with MCF-7 and T47D human breast cancer cells support the presence of a new growth regulatory ERγ. Because the purified IgA and IgM described herein are as effective as serum- bome inhibitors, the serum-free methods described herein provide unique tools to search for ERγ. Assays conducted under these conditions permit estimation of estrogen sensitivities in ranges not approachable by other technology. These methods can also be adapted to measurement of the inhibitor in biological fluids available in only small supply. For example, coupled with use of XAD-4 resin extraction to remove steroids, bodily fluids and other source materials can be assayed on small scale to determine the concentration of effective inhibitor. This is of particular interest because IgA in plasma is > 90% inactive monomer and < 10% active dimer/polymer. Measurement of IgA by conventional methods gives total concentrations, and does not determine the concentration/presence of active inhibitor. The present biological activity method has distinct features and advantages, and can serve as an adjunct measurement.

0561Serum-free defined medium assays can be used to search for new compounds that mimic the action of immunoglobulins to block cancer cell growth in its early stages. This screening can be done under conditions in which serum proteins might interfere. Compounds so-identified can next be evaluated by addition of CDE-serum or XAD-4 treated serum to determine if serum proteins interfere and to determine drug efficacy in vitro under both serum-free defined medium conditions and serum supplemented conditions. Serum-free defined medium method can be used for screening of compounds that may either enhance or inhibit immune function at the epithelial cell level. Compounds with these activities may have utility as immune enhancers to help reduce the risk of cancer development. These assay methods offer a screening tool for such compounds that has not been available before. Larger magnitude effects permit greater accuracy with the new assay methods when estimating effects of substances that are less potent than natural estrogens. Example 22: Effect of Tamoxifen Antiestrogen in Serum-free Defined Medium

0562This Example illustrates the use of one of the present assays to detect the estrogenic or anti-estrogenic effect of a substance. In particular, the classical antiestrogen was assayed as a demonstration of the usefulness of the present assay system. Background of Tamoxifen Effects and Clinical Applications. The antiestrogenic effects of tamoxifen are well documented. Most evidence suggests this compound and its active metabolite 4-hydroxyl-tamoxifen prevent growth of ERα positive cells via interaction with the receptor (Coezy E et al. (1982) Cancer Res 42, 317-323; Bardon S et al. (1984) Mol Cell Endocrinol 35, 89-96; Reddel RR et al. (1985) Cancer Res 45, 1525-1531). However, it has also been suggested that tamoxifen blocks growth factor promoted MCF-7 breast cancer cell growth (Vignon F et al. (1987) Biochem Biophys Res Commun 146, 1502-1508). Also, tamoxifen has high affinity binding sites and actions distinct from the estrogen receptor (Sutherland RL et al. (1980) Nature (Lond) 288, 273-275; Phaneuf S et al. (1995) J Reprod Fertil 103, 121-126). Despite its complex actions, tamoxifen has widespread use as a treatment for breast cancer (Fisher B et al. (1998) JNatl Cancer Inst 90, 1371- 1388; Jaiyesimi IA et al (1995) J Gin Oncol 13, 513-529; Clinical Trial Report (1997) J Gin Oncol 15, 1385-1394; Clinical Trial Report (1987) Lancet 2(8552), 171-175; Forrest AP et al. (1996) Lancet 348(9029), 708-713; Tormey DC et al. (1996) JNatl Cancer Inst 88, 1828-1833; Gundersen S et al. (1995) Breast Cancer Res Treat 36, 49-53; Gelber RD et al. (1996) Lancet 347(9008), 1066- 1071; Raabe NK et al. (1997) Ada Oncol 36, 2550260). Serum-free Medium Effects of Tamoxifen. The effects of tamoxifen (TAM) were reexamined under completely serum-free defined conditions. It is very important to note that throughout the Examples herein, data is presented showing that estrogens have either had no effect on growth in defined medium or at most a 1.0 CPD effect that was related to saturation density. This was true no matter if phenol red was present or absent from the medium, as shown in Example 8 and reported (Moreno-Cuevas JE and Sirbasku DA (2000) In Vitro Cell Dev Biol 36, 447-464). Li similar assays, 1.0 x 10^"7 M tamoxifen was completely inhibitory with T47D cells in culture, as shown in Fig. 138. The study shown in Fig. 138 examined the concentrations of tamoxifen needed to fully inhibit T47D cell growth in the preferred formulation of DDM-2MF serum-free defined medium without any source of estrogens. The expected outcome was no tamoxifen inhibition. As shown, estrogen alone had only a 1.0 CPD effect in serum-free defined medium. However, tamoxifen had unexpected effects revealed by the use of serum-free defined medium. Tamoxifen effectively arrested growth at 1.0 x 10^"7M. Higher concentrations were cytotoxic. It should be noted that tamoxifen had the same effect as immunoglobulins IgA and IgM. To demonstrate this fact another way, the experiment in Fig. 139 shows that estrogens completely reverse the effect of 1.0 x 10^"7 M tamoxifen. This sequence of experiments shows the same results as that shown above with plasma IgA and IgM and ER⁺ cell lines. Discussion of Example 22. The observation of inhibition of cell growth by a classical antiestrogen demonstrates the usefulness of this technology to search for other antiestrogenic compounds. Furthermore, because of the current intense focus on the search for SERMs (i.e. Selective Estrogen Receptor Modulators) the serum-free technology disclosed herein has particularly useful applications. Specific types of SERMS can be sought for different cell types. Those SERMs that do not cause breast cancer cell growth can be readily identified by this technology. Those SERMs with multiple activities can be identified before conducting expensive animal testing.

0563The technology presented permits a clear definition of antiestrogens with "mixed" functions (e.g. tamoxifen-like, that act at several sites) versus those with a "pure" function mediated only by the estrogen receptor. To date, no similar easily applied in vitro method based on serum-free defined medium and secretory immunoglobulins is available that produces growth as an endpoint of the assay. An entirely new function for the drug tamoxifen is proposed, in which the tamoxifen mimics the immune system effects on ER⁺ cancers thereby inhibiting growth. Estrogen reverses these effects, not as a consequence of interaction with the classical ERα, but as a consequence of the ERγ. Tamoxifen may also be an antagonist of ERγ, and this possible use for tamoxifen is now proposed.

0564The serum-free defined medium technology presented herein has direct application to the assay of a great variety of drugs now in use by women either before the onset of breast cancer or after the onset. Drugs or preparations such as antidepressants, herbal extracts, soy products, other food, plant or microorganism extracts, estrogenic creams and cosmetic preparations can be assessed for anti-estrogenic or estrogenic activity.

0565Serum-free assay methods are also applicable to exploration of additional antiandrogenic compounds. Furthermore, in view of the possible role of estrogens as well as androgens in prostate growth, this technology can be used to search for compounds with both activities.

0566Example 23: IgGl and IgG2 as Immunoglobulin Regulators of Estrogen and Androgen Responsive Cancer Cell Growth

0567This Example investigates and discusses the relative effectiveness of certain IgGs as inhibitors of steroid hormone responsive cancer cell growth. The IgG Subclasses and the Importance of Assessing Each for Activity. As reviewed above, in the Background of the Invention, three classes of immunoglobulins are secreted by mucosal tissues. The IgG class is lowest in concentration in secretions, but still physiologically important because of its capacity to neutralize pathogens by different mechanisms. Additionally, the studies above showed that bulk purified mixtures of all subclasses of horse and rat IgG were not estrogen reversible inhibitors for MTW9/PL2 rat mammary tumor cells. The human clinical importance of understanding and measuring IgG subclasses has been growing steadily. From a few clinical reports per year in 1970, the literature now exceeds four hundred reports a year. These assays are significant for several reasons, (i) They provide a clearer picture of an individual's susceptibility to disease, (ii) An awareness that treatment for subclass deficiencies is important, (iii) The subclasses can be used to assess the state of a number of diseases, (iv) The importance of IgG subclass difference between ethnic groups and different races is a potential area for expanded control of disease.

0568Test of Rat IgG Subclasses as Estrogen Reversible Inhibitors of MTW9/PL2 Rat Mammary Tumor Cell Growth. The IgG subclasses of rat are IgGl, IgG2A, IgG2B and IgG2C. These IgGs, obtained from commercial sources previously identified herein, were tested at 15 μg/mL with MTW9/PL2 cells in DDM-2A serum-free defined medium (Fig. 140). All four IgG subclasses were compared to rat plgA and rat plgM. The latter two were estrogen reversible inhibitors, as expected (Fig. 140). However, the four IgG subclasses were not inhibitors at a concentration that was effective with IgA or IgM. The estrogenic effects recorded in cultures with them were no larger than seen in serum-free defined medium alone (Fig. 140). Clearly, IgG are not effective steroid hormone modulators in rat. Test of Human IgG Subclasses as Estrogen Reversible Inhibitors of Breast and

0569Prostate Cancer Cell Growth. The subclasses of human IgG are IgGl, IgG2, IgG 3 and IgG4. They are formed with both λ and K light chains. A series of studies was performed, and it was found that with human breast cancer cells, only IgGlκ was a significant estrogen reversible inhibitor. Fig. 141 shows a comparison of its activity to human plgA and plgM. At 40 μg/mL, it was 37% as effective as plgM. A similar study with LNCaP cells showed that only IgGlκ had activity greater than the estrogenic effect seen in CAPM serum-free defined medium only (Fig. 142). L some experiments with prostate cells, IgG2κ also showed androgen reversible inhibitory activity (data not shown).

0570Discussion of Example 23. The effect of IgGlκ raises an issue not encountered with IgA or IgM. The preference for the K light chain implies that a different receptor mediates the effects of this immunoglobulin. This immunoglobulin may have greater effect on normal breast or prostate cells as an inhibitor. It is also believed that part of the transformation/progression process leading to hormone responsive cancers is an attenuation of the effectiveness of IgGlκ as an inhibitor. The present IgGl observations have other applications, as well, including the measurement of the IgGlκ subclass in different populations such as black American, Asian, white, Native American and Hispanic with contrasting susceptibilities to breast and prostate cancer, or individuals within any one ethnic group, may provide additional information and confirmation of the usefulness of such measurements. These measurements can be made in bodily fluids or plasma. Measurement in milk and breast fluid may provide an indication of susceptibility to the development of breast cancer. Irrespective of the receptor that mediates the growth response of IgGlκ, this receptor will be a candidate for the missing transcytosis receptor for IgG. Its molecular identification has utility in diagnostic specimens of breast, prostate and other cancers and can be used to determine new uses of the immune system for therapeutic applications. Once it is completely identified, the receptor that mediates the IgGl growth inhibition effect will provide another target for development of compounds that mimic the immune system inhibition of cancer cell growth.

0571Example 24: Mediation of IgA/IgM Effects by the Poly-Ig Receptor

0572In this Example, it was determined that a poly-Ig receptor or a poly-Ig like receptor mediates the inhibition of cell growth by IgA and IgM. The negative response to IgA and IgM is mediated by the mucosal poly-Ig receptor or a very similar structure with the same immunoglobulins specificity as well as the same immunological and M_r properties. The known poly-Ig receptor is a M_r 100,000 transmembrane protein with several properties that place it in the Ig superfamily of receptors (Krajci P et al. (1992) Eur J Immunol 11, 2309-2315; Williams AF and Barclay AN (1988) Annu Rev Immunol 6, 381-405).

0573Genetic Properties of the Poly-Ig Receptor. The complete genomic and cDNA sequences of the poly-Ig receptor have been determined (Krajci P et al. (1991) Hum Genet 87, 642-648; Krajci P et al. (1992) Eur J Immunol 22, 2309-2315). Poly-Ig receptor gene has been localized to chromosome 1 at Iq31-q42 locus Krajci P et al. (1991) Hum Genet 87, 642-648; Krajci P et al. (1992) Eur J Immunol 11, 2309-2315; Krajci P et al. (1995) Adv ExpMed Biol 371A, 617-623). The long arm of chromosome 1 had initially been described as the location of the most frequent ctyogenetic abnormalities found in human breast carcinoma (Bieche I et al. (1995), Gin Cancer Res 1, 123-127). More recently this conclusion was modified to state that distal alterations of the short arm of chromosome 1 are the most frequent cytogenetic abnormalities in human breast carcinoma (Bieche I et al. (1999) Genes Chromosomes Cancer 24, 255-263). The gene encoding the poly-Ig receptor is linked to D1S58 on the long arm of chromosome 1 (Krajci P et al. (1992) Hum Genet 90, 215-219). This locus (i.e. D1S58) is a known site for "allelic imbalances" in a remarkable 75% of all breast cancers (Loupart M-L et al. (1995) Genes Chromosomes Cancer 12, 16-23). Allelic imbalances include "Allelic Loss, Allelic Gain, and Imbalances". Loss of heterozygosity (LOH) is consistently high along the length of the long arm of chromosome 1 at D1S58 (i.e. 46%) in breast cancers (Loupart M-L et al. (1995) Genes, Chromosomes &amp; Cancer 12, 16-23). LOH is strongly associated with development of cancer. Viewed in light of the present invention, these published published reports gain new meaning and significance. The report describing changes in D1S58 did not specify what gene or type of gene or function might be impaired by damage to this locus (Loupart M-L et al. (1995) Genes, Chromosomes &amp; Cancer 12, 16-23). The present results indicate that this "hot spot" is either the authentic poly-Ig receptor acting in its new capacity as a growth regulator, or a very closely related receptor with similar molecular weight, ligand binding and immunological properties. However, it must be recognized that the functional form of the growth regulatory receptor may arise from alternate splicing of the poly-Ig receptor gene. Alternate splicing of the poly-Ig receptor gene is known in rabbit (Deitcher DL and Mostov KE (1986) Mol Cell Biol 6, 2712-2715; Frutiger S (1987) JBiol Chem 262, 1712-1715) and bovine tissue (Kulseth MA et al. (1995) DNA Cell Biol 14, 251-256). It has yet to be proven (or disproved) in humans. Certainly this possibility is still open with hormone responsive cancer cells. Alternately the Iq31-q41 region of chromosome 1 contains several other genes of immunological interest (Krajci P et al. (1991) Hum Genet 87, 642- 648; Krajci P et al. (1992) Eur J Immunol 22, 2309-2315; Bruns GAP and Sherman SL (1989) Cytogenet Cell Genet 51, 67-77). There can be little doubt that the discovery of immune negative regulation of growth mediated by the poly-Ig receptor, or one very related, is an advance. It was arrived at not by the genetic approach described above which screens genes without regard for function, but instead by a functional approach based on the biochemical, endocrine and cell biology studies described above.

0574Structural Properties of the Poly-Ig Receptor. A very detailed structural analysis of the human poly-Ig receptor has been presented by others (Krajci P et al. (1992) Eur J Immunol 22, 2309-2315). Altogether, eleven exons cover the entire coding sequence. The five extracellular domains designated Dl, D2, D3, D4 and D5 were coded for by exons E3, E4, E5 (D3 &amp; D4), E5 and E6, respectively. The five extracellular domains are repeating disulfide stabilized Ig-like domains with homology to the Ig superfamily of receptors. The functions of D2, D3 and D4 are not well defined. The functions of Dl and D5 are well studied. Dl is the binding site for the Fc domains of IgA and IgM (Frutiger S et al. (1986) JBiol Chem 262, 1712-1715; Bakos M-A et al. (1993) J Immunol 151, 1346-1352; Røe M et al. (1999) J Immunol 162, 6046-6052). The presence of a "J" chain in the immunoglobulins appears essential for receptor binding and secretion (Vaerman J-P et al. (1998) Eur J Immunol 28, 171-182). Dl is highly conserved among species. The amino acid sequence of the Dl loop responsible for IgA/IgM binding has been established as residues 15→37 (Bakos M-A et al. (1991) J Immunol 147, 3419-3426; Bakos M-A et al. (1993) J Immunol 151, 1346-1352; Bakos M-A et al. (1994) Mol Immunol 31, 165-168). A monoclonal antibody recognizing this sequence blocks the binding of the immunoglobulins. Also, anti-J chain blocks binding (Vaerman J-P et al. (1998) Eur J Immunol 28, 171-182). The D5 domain forms a covalent disulfide bond with IgA to form secretory slgA. Exon 8 codes for the membrane spanning sequence/domain. Exons 9, 10 and 11 code for the three cytoplasmic domains that regulate various aspects of transcytosis (Breitfeld PP et al. (1990) JBiol Chem 265, 13750-13757; Reich V et al. (1996) J Cell Sci 109, 2133-2139; Krajci P et al. (1992) Eur J Immunol 22, 2309- 2315). These domains are highly conserved (Banting G et al. (1989) FEBS Lett 254, 177-183). One serine residue is particularly important for transcytosis (Hirt RP et al. (1993) Cell 74, 245- 255).

0575Clinical Studies of Secretory Component (Poly-Ig Receptor) Expression in Breast and Colon Cancer. Others have performed a study of the protein and mRNA expression of the poly-Ig receptor with a sample of human colon cancers (Krajci P et al. (1996) Br J Cancer 73, 1503-1510). Expression of secretory component was found in 33 colorectal adenomas (31 patients) and in 19 colorectal carcinomas from 19 patients. Although that study provides evidence that colon adenomas (i.e. a predisposition to colon cancer) and confirmed cancers express poly-Ig receptor, the investigators did not attempt to translate the observations further than to propose a role in "cellular dysplasia".

0576Likewise, the levels of secretory component have been measured by others in breast tumors from 95 patients with primary or metastatic disease (Stern JE et al. (1985) Cancer Immunol Immunother 19, 226-230). These authors proposed that low levels of secretory component (SC) were found in metastatic lesions and that this "could indicate a potential for SC involvement in immune regulation of tumor growth,"referring to conventional antigen-antibody recognition immune effects. However, it was not undertaken to identify growth effects related to either the immunoglobulins IgA/IgM or to identify a role of the poly-Ig receptor directly. Furthermore, this study was incomplete. There was no attempt made to determine the estrogen receptor status of the primary or metastatic disease. Therefore, there was no correlation to growth state based on the most accepted criterion of steroid hormone receptor status. This line of study appears to have stopped with 1985 observation. The present invention has directly addressed the problem by demonstrating growth regulation by the secretory immune system using several different ER⁺ cancers. Lines of Evidence Supporting Poly-Ig Receptor or a Poly-Ig-like Receptor in

0577Negative Growth Regulation. The present series of studies and observations indicate that the IgA/IgM mediating receptor has the properties of the poly-Ig receptor. Supporting facts include the following: (1) The source of the active IgA is not the deciding factor. Plasma or myeloma derived IgA are equally effective. Also, species makes little or no apparent difference in activity. IgA isolated from various species has major sequence homology in the heavy chains but differences in the variable chains. This is consistent with mediation by an Fc superfamily receptor. (2) IgA purchased commercially from myeloma cell sources contains predominantly dimeric and polymeric immunoglobulin. It is highly active as an inhibitor. This is consistent with mediation by the poly-Ig receptor. (3) Cultures containing the active CA-PS-pool π material (see Examples 18 and 20) are predominantly dimeric/polymeric forms of immunoglobulins. These preparations are active in serum-supplemented and serum-free defined medium. This is consistent with the expected binding to the poly-Ig receptor. (4) IgM is at least as active, or more active than IgA on a molar basis. The source of the IgM can be from plasma or myeloma cells. They are equally effective. This is also expected of the poly-Ig receptor. (5) Anti-secretory component antibodies completely blocked the inhibitory effects of IgA and IgM. This not only indicates poly-Ig receptor mediation, it supports the view that IgA and IgM act via the same receptor. The poly-Ig receptor is known to conduct transcytosis of both of these immunoglobulins. Secretory IgA is invariably inactive as an inhibitor. It has the five extracellular domains of poly-Ig receptor attached. By contrast, plasma derived IgA is fully active. To prove that plgA does not have the secretory component whereas slgA contains the 80kDa receptor fragment, the Western analysis in Fig. 143 was done. Secretory IgA shows an 80kDa cross- reaction band with anti-secretory component whereas plgA shows no reaction. This was the expected result and provides additional support for the view that the poly-Ig receptor is the mediator. Because secretory component is isolated from milk slgA, these results show that the secretory component used for immunization of the rabbits was free of the other subunits in IgA. This was a good control for the next experiments. Li the next experiments, anti-human secretory component antiserum was used to block the inhibiting effects of IgA and IgM. Fig. 144 shows the results with the T47D cells in serum- free defined medium DDM-2MF with human plasma IgM alone and with a series of dilutions of the antiserum. As shown, 10 nME₂ completely reversed the IgM inhibition. Dilutions of 1:500 to 1:5000 also blocked the inhibition. In the insert in Fig. 144, a control study with pre-immune rabbit serum demonstrated it had no inhibitor blocking activity. A similar study was done with LNCaP cells in serum-free defined CAPM with human plgA (Fig. 145). As shown, 10 nM E₂ completely reversed the plgA inhibition. Anti-serum dilutions of 1:00 and 1:1000 also reversed the inhibition. Differences between the effective dilutions with T47D and LNCaP cells is due to changes in lots of commercially prepared antiserum. To determine if IgA/IgM responsive cells expressed lOOkDa poly-Ig receptor, the

0578Western analysis shown in Fig. 146 was done. Amounts of extracts of the designated cell types were analyzed with a 1:1000 dilution of rabbit anti-human secretory component. As expected MDCK cells were positive. This cell line has been studied for several years as a model of poly-Ig receptor sorting and function. LNCaP cells showed the same receptor (Fig. 146). Cell lines that were negative were ALVA-41, DU145, human fibroblasts, and PC3 cells (Fig. 146). As shown in multiple experiments in preceding Examples, LNCaP cells are IgA/IgM inhibited. The results of the Western analyses show that they express the poly-Ig receptor.

0579In the final experiments of this series, plgA was tested with two of the cell lines that were poly-Ig receptor negative by the Western analysis shown in Fig. 146. The results with DU145 cells are shown in Fig. 147. Plasma IgA was not an inhibitor. A similar study with PC3 cells is shown in Fig. 148. Again, plgA was not an inhibitor even at 50μg/mL. These results demonstrate cells that lack the poly-Ig receptor are also insensitive to plgA. The HT-29 colon cancer cells are known to express only the authentic form of the poly-Ig receptor. They are also negatively growth regulated by IgM. Discussion of Example 24. For the first time a relationship between immunoglobulin growth regulation and the poly-Ig receptor is demonstrated. This receptor has in the past been studied only from the perspective of a transcytosis receptor. In view of all these results, the poly-Ig receptor very likely has more function than transcytosis only. Ongoing investigations are directed to identifying gene changes in the authentic poly-Ig receptor gene, which may include point mutations, deletions, insertions, and premature teπnination. The receptor mediating the effects of IgA IgM may be an alternate splicing form of the original transcytosis receptor. Alternatively, changes in the regulation of expression may influence the presence or absence of this receptor. The positive correlation between the presence of ER and AR, and expression of the growth regulating poly-Ig receptor indicates regulation or positive influence by steroid hormones.

0580One of the primary themes of cancer research has been that loss of "tumor suppressor genes" causes the release of cells from negative regulation and thereby contributes to the progression to cancer. The present invention indicates that the poly-Ig receptor has a "tumor suppressor" function. It is present in cells that are regulated by IgA/IgM and absent in cells that are insensitive to immune inhibitors. This is a new aspect of cancer immunology not recognized before the present invention. For the first time, the poly-Ig receptor is connected to the D1S58 linked locus that is a "hot spot" for genetic changes in breast cancer. It is now proposed that this locus or near neighbors contain the growth regulating form of authentic transcytosis poly-Ig receptor or a very similar immunoglobulin superfamily receptor. Alternately the Iq31-q41 region of chromosome 1 contains several other genes of immunological interest that include the receptor for IgA/IgM. Those genes can be employed as screens for breast and other mucosal cell cancers. They are expected to indicate susceptibility and may be useful in prognosis and other diagnostic applications with human tissue and cancer samples. Analyses of allelic imbalances in the receptor gene are also foreseen as a new tool to determine susceptibility and prognosis for development of breast and other mucosal cancers, as will be the detection of mutations in the growth regulating intracellular domains of the receptor. The known amino acid sequence of the poly-Ig receptor does not contain the immunoreceptor tyrosine-based inhibitory motif (ITIM) common to a new family of inhibitory motif receptors (Cambier JC (1997) Proc Natl Acad Sci USA 94, 5993-5995). Other amino acid sequences may serve this same function.

0581Example 25: Mediation of IgGlκ Effects by a Fc-like Receptor

0582In this Example, it is shown that the inhibiting effects of IgGlκ were mediated by an Fc receptor or Fcγ-type receptor. It is highly unlikely that IgGl acts via the poly-Ig receptor. The poly- Ig receptor has a requirement for "J" chain for binding (hence its specificity for dimeric/polymeric IgA or pentameric IgM each of which have one J chain. Also, as shown in TABLE 12, Fcγ receptors are localized in leukocyte series or bone marrow origin cells. There is no convincing evidence of their presence in epithelial cells or in secretory cells of the mucosa. It is now proposed that the receptor being sought is one analogous to the Fcγ in at least two significant properties. First, it binds monomeric IgGl via the Fc domain of the immunoglobulin with some participation of the K light chain. Second, that the receptor has inhibitory activity akin to a new family of Fc receptors. The amino acid sequence of the new IgGlκ receptor is expected to have an immunoreceptor tyrosine-based inhibitory motif (ΓTIM) (VxYxxL) common to the new family of inhibitory motif receptors (Cambier JC (1997) Proc Natl Acad Sci USA 94, 5993-5995). Other amino acid sequences may serve this same function.

0583It is proposed that the Fcγ family of receptors contains members that possess the very special property of mediating cell growth inhibition. The methods of identification are outlined below.

0584TABLE 12 Properties of the Fc γ Family of Receptors

0585<img file="EP1328293A2_D0013.tif" />

0586Discussion of Example 25. The amino acid sequence of a new Fc family receptor may include immunoreceptor tyrosine-based inhibitory motif (ITEM) common to a new family of inhibitory motif receptors (Cambier JC (1997) Proc Natl Acad Sci USA 94, 5993-5995). The results obtained in the present studies support the involvement of Fc receptors of mucosal cells that include one of the known members of the family of ITIMs or other amino acid sequences that serve this same function. Ongoing work includes genetic mapping to a specific chromosome number and locus. The genomic DNA sequence of the new receptor (or existing receptor if already known), introns and exons, will be obtained. This receptor may be used for diagnostic and clinical proposes, and as a screen for genetic susceptibility to breast and prostate and other mucosal cancers, as described in more detail in U.S. Patent App. No. (Atty. Dkt. No.

05871944-00800)/PCT/US2001/ (Atty. Dkt. No. 1944-00801) entitled "Compositions and

0588Methods for the Diagnosis, Treatment and Prevention of Steroid Hormone Responsive Cancers," which is hereby incorporated herein by reference. Identification of mutations and changes associated with progression from normal cells to autonomous cancer cells is also expected, and, along with detection of changes in regulation of expression and allelic imbalances in the receptor gene, will have very useful research, diagnostic and clinical applications. Example 26. Immunoglobulin Inhibitors as Tools for Identifying the Receptors that Mediate the IgA/IgM IgG Cell Growth Regulating Effects

0589As shown by Examples 24 and 25, the present Immunoglobulin Inhibitors can be used as reagents for identifying their mediating receptors.

0590The Mediating Receptors - Inhibitory Function. As discussed above, serum contains a great variety of mitogenic agents. On this point the present results in 50% (v/v) serum were especially relevant. This concentration of serum is a rich source of mitogens including insulin and the insulin-like growth factors. Nutrients and other serum components also have growth-promoting effects. Examples include diferric transferrin, unsaturated fatty acids bound to albumin, complex lipids and ethanolamine. Clearly, the inhibitor(s) also blocks their growth effects, which lends support to the conclusion that the mediating receptor for the serum-bome agent must have special properties. For instance, the immunoreceptor tyrosine-based inhibitory motif (ITIM) class of Fc receptors is of particular interest with respect to identifying the mediator(s) of immunoglobulin inhibition of cancer cell growth, because the hallmark of the ITIM receptors is that they have an intracellular amino acid sequence motif I VxYxxL that signals cell growth shutdown after ligand binding, and therefore shuts off growth factor dependent growth. Li the preceding Examples it is demonstrated that steroid hormones are selectively capable of reversing the effects of the serum inhibitors). Plainly, as predicted by the estrocolyone hypothesis, serum contains an inhibitor(s) that has a dominant role in the regulation of proliferation of steroid hormone target cells. The isolated IgA and IgM blocked growth factor dependent growth in serum-free defined medium. Because of its "master switch" character, the newly identified immunoglobulin inhibitors have many important and useful applications. Moreover, the results of the present investigations support the view that the inhibitors) will have biological implications extending well beyond estrogen and androgen target tissues.

0591The Receptor Mediating IgA/IgM/IgG Inhibitory Effects. The results shown in the foregoing Examples strongly indicate that the IgA/IgM growth inhibition is mediated either by the poly-Ig receptor or a very closely related receptor. Establishing a growth regulating function for this "transcytosis" receptor will open new directions in medical diagnosis, treatment and prevention of cancers of mucosal epithelial tissues. It will be determined whether the poly-Ig receptor, or a poly- Ig like receptor mediates the growth regulating effects of IgA on human breast and prostate cancer cells in culture. For this study, the poly-Ig receptor in these cancer cells will be identified using well-known PCR cloning technology, ¹²⁵I-labeled IgA chemical cross-linking and Western and immunohistochemistry methods described in the literature. Next, blocking polyclonal antibodies or blocking monoclonal antibodies will be employed to show that the poly-Ig receptor mediates the growth response. The antibodies will be raised against the poly-Ig receptor using known techniques. Reversal of the inhibitory effect of IgA and IgM by blocking the poly-Ig receptor will suggest that the poly-Ig receptor is not just a simple transport receptor, but that it has a central role in breast and prostate cancer cell growth regulation. There is no existing paradigm for breast or prostate cell growth regulation that involves the poly-Ig receptor or for that matter any receptor specific for the IgA class of immunoglobulins including Fcα receptors (Fridman WH (1991) FASEB J 5, 2684-2690).

0592The different forms and domains of IgG, IgA and IgM that act as inhibitors of normal prostate and breast and other mucosal epithelial cell growth and the hormone responsive and hormone autonomous forms of these cancers in serum-free defined culture medium will be determined and used as tools to evidence or confirm the identity of the receptor(s) responsible for mediating the growth regulatory effect. The properties of the ligand that elicits a response will be evidence supporting the identity of the receptor. Poly-Ig receptor is activated by Fc-domains as are Fcγ receptors. Normal cells are likely to be most inhibited by IgG, IgA and IgM, whereas the ER⁺ and AR⁺ cells will likely be inhibited primarily by IgA/IgM, and ER^" and AR^" cells will likely not be inhibited by any of the three classes of immunoglobulins, as predicted by the conceptual model described below. The methods employed will include direct tests of the activity of IgG, IgA and IgM on cell growth as well as assessment of the activity of specific size forms and Fc versus Fab fragments. Antibodies such as anti-J chain and anti-Fc will be used to extend these studies to demonstrate that the Fc is the active domain and that Fc binding receptors are involved.

0593More specifically, AR⁺ LNCaP cells, the AR^" PC3 and DU145 cells, and the AR⁺ ALVA- 41 cells will be studied. Normal human prostate and breast epithelial cells will be obtained from Clonetics. Growth assays will be done in completely serum-free CAPM (prostate) and DDM-2MF (breast), as described above. IgAl and IgA2 will be purified from human serum and colostrum, using techniques that are well known and have been described in the literature. Initial small samples will be obtained from a commercial supplier such as The Binding Site (San Diego, CA). The monomeric, dimeric and polymeric forms of IgA will be separated using techniques that are well known and have been described in the literature. If only IgA2 has activity, it will be further separated into the A2(m)l and A2(m)2 allotypes, using well-known techniques that have been described in the literature. Because the initial IgA IgM inhibitor preparations evaluated were mostly dimeric and monomeric, those forms are expected to be the most active in these series of tests. Confirmation that the most active forms are dimeric/polymeric IgA/IgM will be strong evidence for poly-Ig receptor mediation. Should the monomers be revealed as the only active inhibitor forms, however, it would favor Fc or Fc superfamily receptors, in which case the Fcα will be investigated as a possible mediator. IgA will be fragmented with a specific protease to yield Fc and Fab fragments from IgA, using techniques that are well known and have been described in the literature. The Fab and Fc fragments of IgM will be obtained using a Pierce Chemicals kit based on immobilized trypsin. Fab and Fc fragments of IgGl will be obtained using another Pierce kit. If only Fc fragments of IgA and IgM are active, mediation by the poly-Ig receptor is likely. If the Fc of IgGl is active, it will indicate an Fc receptor as the mediator.

0594The immunoglobulin inhibitors will also be used as tools or biological reagents to confirm whether IgG acts via a receptor different than IgA/IgM. Based on the results reported above, identification of Fcγ like receptors and the poly-Ig receptor (or related receptor) with normal cells, ER⁺ cells and AR⁺ cells is expected, and no functional receptors are expected in ER^" cells or AR^" cells. ¹²⁵I-labeled IgGl, IgA and IgM will be prepared using chloramine T or Iodogen beads or coated tube (Pierce Chemicals kits). Binding parameters, binding constants, analyses of the effects of reciprocal additions of labeled and unlabeled immunoglobulins to identify separate or similar binding sites, and determination of the effects of addition of purified secretory component on IgA and IgM binding will be performed as previously described or using well known published techniques. Specific binding will be as total binding minus binding in a 100-fold excess of unlabeled protein. For each form with activity, time, concentration and temperature dependence of binding will be assessed. Scatchard analysis will be used to estimate the number of sites per cell and the association constants (K_a). Reciprocal competitions with unlabeled and labeled immunoglobulins will be used to define interaction with the same or different receptors. This latter point is important because binding of both IgA and IgM to the same site strongly favors the poly-Ig receptor and plainly contra-indicates Fcα (IgA) or Fcμ (IgM) receptors, which are members of a superfamily in which each member is specific for a (monomer) class of immunoglobulins. Li addition, the effects of blocking antibodies such as anti-secretory component, anti J chain and anti Fc will be assessed with all three cell types. Where indicated, chemical cross-linking with ¹²⁵I-labeled Ig will be performed to define the mass of the receptors. Optionally, metabolic labeling and/or immunoprecipitation techniques will be used instead, employing well-known techniques.

0595Western immunoblotting with normal, steroid hormone receptor positive and steroid hormone receptor negative cell types will be performed to identify the receptors present. Lnmunohistochemistry will be applied to identify the poly-Ig receptor and Fcγ receptors on all three types of cells using the blocking antibodies. Using a full-length human poly-Ig receptor cDNA clone, SI nuclease protection assays will be conducted with RNA from normal prostate and breast cells, ER⁺ and ER^" breast cancer cells, and AR⁺ and AR^" prostate cancer cells to identify mRNA. Li the cases of ER⁺ and AR⁺ or ER^" or AR^" cells, this method will help to identify truncated or otherwise altered receptors or non-functional receptors. As described in certain of the preceding examples, Western blots have already been conducted, as well as cell growth assays with receptor blocking antibodies. The remaining analyses will be done with normal cells as well as all other ER^" or AR^" lines. All blocking antibodies are dialyzed against buffer containing charcoal to remove interfering steroid hormones. Rabbit polyclonal anti secretory component will be raised (e.g., by HTI BioProducts, Ramona, CA) and rabbit polyclonal anti-human J chain and specific antibodies against the Fc receptors for IgG and IgA are commercially available (Accurate). The specificity of all antiserum will be checked by Western analysis.

0596To identify the receptors mediating the androgen reversible inhibition of normal and/or AR⁺ cells, PCR cloning methods will additionally be used to determine the cDNA sequences of the poly- Ig receptor and Fcγ receptors from normal, AR⁺ and, if indicated, from AR^" cells. This method will provide clear answers to the question of the relationship of the human poly-Ig receptor and Fcγ receptors to immune system negative regulation. It is expected that the receptors will be found to be either identical to known sequences or altered in sequence to convert them to "inhibitory motif receptors. Based on the known cDNA sequence of the poly-Ig receptor from HT-29 cells, PCR cloning technology will be applied to obtain a full-length clone from the LNCaP and T47D cells. Ongoing investigations are directed to comparing receptor sequences from normal prostate and breast cells to identify any changes. Based on the known sequence of the FcγRTJ-Bl receptor, these same studies will be repeated. The receptors identified by cloning will be examined for the immunoreceptor tyrosine-based inhibitory motif (ITIM) amino acid sequence LNxYxxL or related sequences. Concomitantly, the cells will be examined by Western analysis for SHP-1 and SHP-2 phosphatase mediators of the inhibition of growth factor activity. These markers are not only associated with the inhibitory motif but also other inhibitory receptors. More specifically, an LΝCaP and T47D full-length poly-Ig receptor clone will be prepared and compared to the reported sequence of the poly-Ig receptor. The same technology will be applied to the poly-Ig receptor from normal prostate cells, and, if indicated, from the AR⁺ lines. Because these cell lines are expected to express the known poly-Ig receptor, or a related form, the PCR approach is applicable. The same approach will be used with the Fcγ like receptor. However, in this case, because these receptors are predominantly lymphoid origin, the form in epithelial cells may be substantially different. Standard cloning methods will be employed to obtain the complete cDΝA sequence of the Fcγ like receptor from normal and LΝCaP cells. Total RΝA will be extracted and mRΝA purified by oligo dT cellulose chromatography (also for Northern analysis). cDNA synthesis will be done with oligo dT primers and AMV reverse transcriptase followed by Rnase H to remove RNA. Second strand synthesis will be done with hexameric random primers and DNA pol. I. Treatment with T4 DNA pol, Rnase H and Rnase A creates blunt ends. EcoRI methylation is followed by EcoRI linkers and ligation into a cloning vector. (Stragene) vectors based on λgtlO (hybridization screening) and λgtl 1 (secretory component antibody screening). Both vectors will accept inserts larger than the receptor. The cDNA sequence of human poly-Ig receptor known is the genomic sequence. These will be used to prepare sequence specific primers for PCR. The primers will encompass the 5' and 3' non-coding sequences to ensure a complete cDNA. The PCR products will be subcloned using the TA kit from Invitrogen. The sequencing of PCR clones will be done by the dideoxy chain teπnination method (Lone Star Labs, Houston, TX). From these, determination of whether there have been significant alterations in the receptor during the transition from normal to ER^" and AR^" cancer cells is expected. From sequence data, the ITIM amino acid sequences indicating an inhibitory motif receptor will be sought. It is important to note, however, that the absence of these sequences does not necessarily rule out an inhibitory function. The Western analyses for SHP-1 and SHP-2 will be valuable for indicating an inhibitory function even in the absence of ITIM or when the ITIM is in a modified form.

0597Discussion of Example 26. Without wishing to be bound by a particular theory, it is proposed that the inhibitory effect of IgGl is more marked with normal cells than with ER⁺ or AR⁺ cancer cell lines and an early step in the pathway to malignancy involves loss by the cell of IgGl regulation. From preliminary investigations, it is suggested that the IgA and IgM receptors are likely to be a common poly-Ig receptor or poly-Ig like receptor, which in normal cells is expected to be the same as in steroid hormone receptor positive cell lines. In contrast, the IgGl receptor, likely an Fc gamma type receptor, is expected to either be altered or its expression greatly reduced in ER⁺ and AR⁺ cell lines. The demonstration that IgGl has a major growth inhibiting effect on normal cells may lead to immunization against breast cancer by administering or enhancing IgGl in at-risk tissues. Characterization of an inhibitory role for IgGl via an Fcγ-like receptor is expected to lead to important innovations in medical diagnosis, treatment and prevention of cancers of mucus epithelial tissues.

0598Example 27: Conceptual Model for Cascading Loss of Cell Growth Inhibition in Cancer Cells

0599Concept. The isolated inhibitors, now identified as IgA, IgM and IgGl, controlled breast and prostate cell growth by acting as a steroid hormone reversible inhibitor even when tested under the very rigorous conditions of serum-free defined culture. These active natural inhibitors are present in blood, bodily secretions and mucosal epithelial tissues. The isolated inhibitors readily prevented the growth of these types of cancer cells when they were still in the early (i.e., hormone responsive) stage, but not in the late, non-hormone responsive stage. These results have many implications with regard to the diagnosis, genetic screening, treatment and prevention of breast, prostate, colon and other mucosal cancers. Without wishing to be bound by a particular theory, considering the present experimental results and discoveries, a new conceptual model for understanding how estrogens cause ER⁺ breast cancer cell growth and for understanding how the natural progression of breast cancers occurs to give rise to highly malignant (and dangerous) hormone autonomous forms is proposed. This same model is applicable to other mucosal tissues that respond to the steroid hormone family of hormones, including androgens and thyroid hormones.

0600Progression Concept based on the Breast Cancer Model - Generally Applicable to

0601Mucosal Tissue Cancers. It is well established that breast cancers pass through a characteristic natural history that involves a gradual evolution from near normal growth patterns into cancers that are completely steroid hormone autonomous (i.e. they are no longer stimulated by steroid hormones). These are usually designated estrogen receptor negative (ER^"). As disclosed herein, it has been found that autonomous (ER^") breast cancer is accompanied by a loss in sensitivity to IgA or IgM. Fully autonomous breast cancers are not inhibited by these secretory immunoglobulins. Li light of the outcome of the present investigations, it appears that autonomous breast cancers lack the poly-Ig receptor that mediates the growth inhibiting effects of IgA and IgM. These results are of extraordinary significance because for the first time they pinpoint a specific genetic change (i.e. in the poly-Ig receptor) that might account for the majority (i.e. approximately 75%) of breast cancers termed "sporadic" and for which there is as yet no clear genetic change identified. Indeed, these results also provide an excellent opportunity to implement gene therapy based on reintroduction of the poly-Ig or poly-Ig like receptor into completely autonomous cancers to regain immunological regulation.

0602It is well established in the literature that IgGl is present in serum during childhood, when breast tissue growth is precisely regulated to body size (isometric growth). The other inhibitors, IgA and IgM, are very low at this time, but increase in serum at puberty. Because adult women have increased positive stimuli for breast cell proliferation due to estrogen production, the presence of IgA and IgM may provide additional protection.

0603It is now proposed that alterations in immune regulation lead to the progression of breast and prostate cells from normal control to ER⁺ and AR⁺ cancer cells and that additional alternations in immune control contribute to the development of fully autonomous cancers, according to the following model presented in TABLE 13:

0604TABLE 13 Model for Progression of Steroid Hormone Dependent Cancers from Normal Growth Regulation by the Immune System to Steroid Responsive Cancers and on to Fully Hormone Autonomous Cancers

0605Normal epithelial cells

0606Inhibitory receptors for IgGl, IgA and IgM Inhibition caused by all Igs

I

0608ER⁺ breast cancers or AR⁺ prostate cancers (hormone responsive)

0609Inhibitory receptors for IgA and IgM Inhibition caused primarily by IgA and IgM

I

0611ER^" breast cancers or AR^" prostate cancers (autonomous)

0612No functional receptors for IgA, IgM and IgGl

0613• No inhibition by Ig

0614Inhibitory Motif Receptors. The receptors mediating the immune response regulation must be at or very near the beginning of the onset of breast cancer. Using the tools developed in the present series of investigations, it is expected that inhibitory motif receptors for these immunoglobulins will be identified. It is now proposed that the mediating receptors are members of the Ig superfamily, which includes Fc receptors and a new class of Ig inhibitory motif receptors. This new class of receptors has emerging importance because of the increasing recognition of the role of negative regulation of cell growth. These receptors have both common and unique properties. They bind immunoglobulins via the Fc domains and hence can be classified as Fc receptors. One of these is, in fact, FcγRITβ that binds IgGl (TABLE 12) and causes inhibition of antigen activation of B cells. There are many other examples (Cambier JC (1997) Proc Natl Acad Sci USA 94, 5993-5995). Among these are more than 15 receptors now designated Signal- Regulatory Proteins (SIRPs). These all express a special inhibitor motif of six amino acids (WxYxxL) that is now referred to as the "immunoreceptor tyrosine-based inhibitory motif or ITIM. One of the most marked characteristics of the ITIM containing SIRPs is that this motif recruits two phosphatases (SHP-1 and SHP-2) to result in the inhibition of all growth factor dependent proliferation. This is similar to what was observed with IgGl, IgA and IgM and ER⁺ breast cancer cells and AR⁺ prostate cancer cells serum-free defined medium. This work is expected to aid in the identification of the missing genes for sporadic breast cancers and a more complete understanding of the cascade of gene changes that lead to complete loss of immune control of breast cell growth.

0615Similarly, it is suggested that alterations in immune regulation also lead to the progression of prostate cells from normal control to AR⁺ cancer cells and that additional alterations in immune control contribute to the development of AR^" fully autonomous cancers. Further studies are directed at identifying a cascade of gene changes leading to complete loss of immune control of cell proliferation. .

0616Similarly, it is also proposed that alterations in immune regulation also lead to the progression of colon cancer cells from thyroid hormone receptor (THR) normal control to THR⁺ cancer cells and that additional alterations in immune control contribute to the development of THR^" fully autonomous cancers. Further studies are directed at identifying a cascade of gene changes leading to complete loss of immune control of cell proliferation

0617In continuing investigations, tests to determine whether steroid hormone independent breast and prostate cancer cell growth results from either the loss of the poly-Ig receptor or an inactivation of its function are being carried out. A series of steroid hormone dependent and steroid hormone independent breast and prostate cancer cell lines will be compared for their inhibitory growth responses to IgA, the presence of poly-Ig receptor m-RNA, the expression of the receptor by ¹²⁵I-IgA binding analysis and immunohistochemistry localization of receptor. Detection of an absence of the receptor or an inability to bind IgA will suggest that cancer cell autonomy arises due to a loss of secretory immune system regulation. Such a result would be entirely new in the field of hormone dependent cancers and would provide a new immune mechanism responsible for conversion from hormone dependence to autonomy. New immunotherapies can be developed based on activating the receptor in hormone responsive cancers and new gene therapies based on reestablishing the function of this receptor in autonomous breast cancers.

0618Ongoing investigation is directed at resolving whether hormone autonomous breast cancer cell lines have functional poly-Ig receptors. The ER^" cell lines to be studied are the MDA- MB-231, BT-20, MDA-MB-330 the non-tumorus HBL-100, and the Hs578t and Hs578Bst. Each will be evaluated for growth in serum-free medium ± IgA and ± E₂. This study will determine if autonomous cells have lost immune system negative regulation. To determine if the receptor is lost, the SI nuclease protection assays will be used to seek its mRNA. A kit from AMBION will be used. Li addition, ¹²⁵I-I labeled IgA will be used to determine specific binding characteristics as described above. Immunohistochemistry will be employed to confirm and/or extend the binding data. If the receptor mRNA and protein are absent, these methods should confirm that fact. If they are present but nonfunctional, these methods should confirm that fact as well.

0619Discussion of Example 27. The proposed model for progression of mucosal cancers from normal cells to fully autonomous cancers is based on the experimental results presented and is unique. No previous recognition has been published of the roles of IgA, IgM and IgGl in breast, prostate, or other mucosal cancers. Application of this model has diagnostic implications. Breast, prostate and other cancers can be examined for content of the receptors for IgA, IgM and IgGl to determine stage of the cancer. This information can be compared to the determination of estrogen receptor and progesterone receptor status to aid in decisions regarding immunotherapy with immune modulators or the immunoglobulins or the use of combined anti-hormone and immune therapy modalities. Tumors that are negative for all of the immunoglobulin receptors are prime candidates for gene therapy to replace the receptors and thereby reestablish immune surveillance, as further described in U.S. Patent App. No. (Atty. Dkt. No. 1944-00800)/PCT/US2001/ (Atty. Dkt. No. 1944-00801) entitled "Compositions and Methods for the Diagnosis, Treatment and Prevention of Steroid Hormone Responsive Cancers," which is incorporated herein by reference.

0620Example 28. IgA IgM Based Test to Detect Lowered Levels of Steroid Hormone Reversible Cell Growth Inhibitors in Plasma or Body Secretions

0621Toward identifying individuals with high susceptibility to breast cancer or prostate cancer, the level of the inhibitory form of IgA (i.e., IgA dimer) will be measured in an individual's plasma, or the secretory IgA and polymeric IgM will be measured in a bodily secretion. Decreases in plasma levels of IgA or decreased secretory capacity into milk or structural alterations in IgA may confer greater susceptibility to breast cancer. Levels are expected to be low in susceptible individuals and to fall with increasing age in normal individuals, substantially mirroring the age distribution pattern associated with breast and prostate cancer incidence. An antibody raised against the D5 domain disulfide regions, with IgA attached, is an example of an assay for the dimeric/polymeric IgA. Li secretory fluids, direct measure of slgA can be done along with a measure of secretory component by radioimmunoassay or other methods using ELISA or biotin- avidin technology. The levels of IgM can be measured directly although their levels are more subject to wide variations. Also, "J" chain can be measured, but only in samples treated to remove the free (unbound) form known to be in plasma.

0622Another useful test process is use of rectal or nasal passage antigen challenge and measurement of the appearance of the specific antibody against the antigen in plasma and secretory fluids by standard high capacity clinical test methods. This will directly measure the immune status of the individual. Those with optimum capacity can be separated from individuals with impaired secretory immune system function. Impaired function of the secretory immune system may indicate susceptibility to cancer. The testing is carried out by first treating a plasma specimen to deplete or substantially remove the steroid hormone content without inactivating or removing the endogenous poly IgA dimer and poly IgM molecules. The hormone depleted specimen is then tested for cell growth inhibitory activity in the presence of added steroid hormone in an in vitro assay employing cultured tumor cells incubated in a defined serum-free medium. Procedures for preparing the steroid hormone depleted plasma or serum and for conducting the assay are described in the preceding examples. XAD™-4 is particularly suited for treating small biological specimens. These extraction methods yield steroid hormone depleted serum that allows identification of 30 to 100-fold estrogen and androgen growth effects (cell number measurement) in culture in 7 to 14 days with human breast and human prostate cancer cells as well at rat mammary, rat pituitary and Syrian hamster kidney tumor cells.

0623The results are compared to similar tests using positive and negative control plasmas or serums, which have defined levels of IgA dimer and poly IgM. Li this way the tumor cell growth inhibitory activity of the individual's plasma is measured. Because the in vitro assay system employs a cell line that forms breast or prostate tumors when implanted in vivo, the in vitro assay results are believed to be suggestive of the in vivo condition of the individual.

0624Alternatively, or additionally, plasma and bodily fluids may be monitored for autoimmune antibodies that block the inhibitory action of IgA and IgM. An expected increase in autoimmune antibodies with increasing age is expected to coincide with increased cancer incidence, or the incidence of cancer may be high in individuals with early onset disease.

0625Each and every claim is incorporated into the specification as an embodiment of the present invention. Thus the claims are a further description and are an addition to the preferred embodiments of the present invention. While the preferred embodiments of the invention have been shown and described, modifications thereof can be made by one skilled in the art without departing from the spirit and teachings of the invention. The embodiments described herein are exemplary only, and are not intended to be limiting. Many variations and modifications of the invention disclosed herein are possible and are within the scope of the invention. Accordingly, the scope of protection is not limited by the description set out above, but is only limited by the claims which follow, that scope including all equivalents of the subject matter of the claims. The disclosures U.S. Provisional Patent Application Nos. 60/203,314; 60/208,348; 60/208,111; 60/229,071 and 60/231,273, and all patents, patent applications and publications cited herein are hereby incorporated herein by reference.