Structure-activity relationships of nonisomerizable derivatives of tamoxifen: importance of hydroxyl group and side chain positioning for biological activity.

The antiestrogen tamoxifen [(Z)-1(p-beta-dimethylaminoethoxy-phenyl)-1,2-diphenylbut-1-ene] is an effective anticancer agent against estrogen receptor (ER)-positive breast cancer. The alkylaminoethane side chain is essential for antiestrogenic activity, but the potency of the antiestrogen can be increased by para hydroxylation of the phenyl ring on carbon 1 of but-1-ene. This compound, 4-hydroxytamoxifen, is a metabolite of tamoxifen and has a very high binding affinity for ER [J. Endocrinol. 75:305-316 (1977)] because the hydroxyl is located in the equivalent position as the 3-phenolic hydroxyl of 17 beta-estradiol. In this study, we have examined the relationship between the relative positions of the hydroxyl and the alkyl-aminoethane side chain and the pharmacological activity of the ligand. A fixed seven-membered ring derivative of the triphenylethylene was used to prevent isomerization. All compounds were tested, with and without 17 beta-estradiol, for their effects on the growth of estrogen-responsive T47D and MCF-7 human breast cancer cells in vitro. The growth of MDA-MB-231 ER-negative breast cancer cells was not affected by any of the compounds tested, at a concentration (1 microM) that had a profound estrogenic or antiestrogenic action in ER-positive cell lines. The relative binding affinity of the compounds was determined using rat uterine ER and was found to be consistent with the observed potencies in vitro. The compounds found to be antiestrogens in vitro were antiestrogenic against estradiol (0.08 micrograms daily) in the 3-day immature rat uterine weight test. All compounds were partial agonists in vivo. In general, the estrogenic and antiestrogenic results obtained in vivo were consistent with the potency estimates obtained with the breast cancer cells in vitro. The results of this extensive structure-activity relationship study demonstrate that the substitution for 4-hydroxytamoxifen appears to be optimal to produce a potent antiestrogen; all other substitutions produced either estrogenic compounds or less potent antiestrogens. The hydroxyl group appears to be critical to locate the alkyl aminoethoxy side chain in the correct position in the steroid-binding site to block estrogen action. Novel antiestrogens were identified that could have been predicted based upon earlier drug-receptor models for the ER.

c-myc and pRB: role in TGF-beta 1 inhibition of keratinocyte proliferation.

The TGF-beta s are potent inhibitors of proliferation of most cell types in culture and in vivo. Previous studies have demonstrated that TGF-beta inhibition of skin keratinocyte proliferation involves suppression of c-myc transcription. Evidence derived from use of expression plasmids for certain DNA viral oncoproteins has suggested that the retinoblastoma gene (RB) may be involved in this process. Transient expression of pRB, like TGF-beta 1, in skin keratinocytes represses expression of a human c-myc reporter plasmid, and the same c-myc promoter region (TCE) is required for repression by either TGF-beta 1 or pRB. We showed here that proliferation and c-myc expression in a cell line lacking normal pRB (DU145 human prostate adenocarcinoma cells) are not inhibited by TGF-beta 1. Oligonucleotides containing the TCE were found to bind to a cellular protein of approximately 106 kD (termed p106) in Southwestern assays, utilizing extracts from both the skin keratinocytes and DU145 cells. TCE binding to p106 was diminished by TGF-beta in TGF-beta-sensitive skin keratinocytes but not in TGF-beta-insensitive SV40-transformed keratinocytes. These data support the hypothesis that pRB is required for TGF-beta 1 suppression of c-myc transcription and suggest the involvement of a cellular factor(s) in addition to pRB in the TGF-beta 1 pathway for inhibition of c-myc transcription and growth inhibition.

TGF beta regulation of epithelial cell proliferation: role of tumor suppressor genes.

Transforming growth factor beta 1 (TGF beta 1) is the prototype of a large family of polypeptides involved in growth control, extracellular matrix production, and development. The TGF beta s have marked stimulatory effects on connective tissue formation. They are chemotactic for fibroblasts, indirect mitogens for certain mesenchymal cells and stimulators of extracellular matrix deposition. The TGF beta s are also potent inhibitors of proliferation of most cell types in culture, and in vivo studies have indicated that the predominant effect of TGF beta 1 on cell proliferation is inhibition. We have investigated the mechanism of TGF beta 1 inhibition of skin keratinocyte growth. Earlier studies demonstrated that TGF beta 1 inhibition of keratinocyte proliferation involves suppression of c-myc transcription, and indirect evidence suggested that the product of the retinoblastoma tumor susceptibility gene, pRB, may be involved in this process. More recently, we have shown that transient expression of pRB in skin keratinocytes can repress human c-myc promoter/CAT transcription as effectively as TGF beta 1. The same c-myc promoter region, termed the TGF beta control element (TCE), was required for regulation by both TGF beta 1 and pRB. Oligonucleotides containing the TCE bound to several nuclear factors in mobility shift assays and a cellular protein of approximately 106 kD in Southwestern assays. Binding of these factors could be demonstrated in cells with or without normal pRB, and the binding of some factors was rapidly inhibited by TGF beta 1 treatment of TGF beta-sensitive but not TGF beta-insensitive cells. These data indicate that pRB can function to inhibit c-myc transcription and suggest the involvement of cellular factor(s) in addition to pRB in the TGF beta 1 pathway for suppression of c-myc transcription. Studies with other cell types have shown that another tumor suppressor gene, p53, can also regulate transcription of c-myc in transient assays. Whereas wild type p53 markedly suppressed transcription, four different mutant p53 clones caused transactivation. The data support the hypothesis that pRB and p53 can both cause growth inhibition by blocking the expression of the protooncogene, c-myc, and indicate that tumor suppressor genes may function in the response pathway for diffusible negative growth regulators such as TGF beta.

Characterization of a receptor-negative, hormone-nonresponsive clone derived from a T47D human breast cancer cell line kept under estrogen-free conditions.

We have established an estrogen receptor- and progesterone receptor-negative, hormone-nonresponsive breast cancer cell line from a receptor-positive, hormone-responsive line grown under estrogen-free conditions. T47D breast cancer cells were cultured under estrogenized conditions (in phenol red-containing medium supplemented with whole fetal bovine serum) and cloned to produce line T47D:A18. The parental T47D line was also estrogen deprived (in phenol red-free medium supplemented with dextran-coated charcoal-treated fetal bovine serum) for more than 1 year and subsequently clone T47D:C4 was established. T47D:A18 was estrogen receptor and progesterone receptor positive as determined by both ligand binding assay analysis and enzyme immunoassay analysis. T47D:C4 cells were estrogen receptor and progesterone receptor negative and mRNA for these receptors was not detected. Incubation of hormone-responsive T47D:A18 cells with 17 beta-estradiol caused a 3-fold increase in cell growth over 8 days when compared to control. This stimulation of growth was completely inhibited by the anti-estrogens 4-hydroxytamoxifen (0.1 microM) and ICI 164,384 (1.0 microM). Receptor-negative T47D:C4 cells were refractory to the effects of both 17 beta-estradiol and the antiestrogens. T47D:A18 cells grown under both estrogen-containing and estrogen-free conditions expressed low levels of transforming growth factor (TGF)-alpha and epidermal growth factor receptor mRNA. In the presence of estrogen, high levels of TGF-beta 1 mRNA were detected in T47D:A18 cells. These levels decreased when T47D:A18 cells were grown in estrogen-free media. Conversely, TGF-beta 2 mRNA was not detected in T47D:A18 cells cultured under estrogenic conditions; however, message was detected after the cells were cultured under estrogen-free conditions. T47D:C4 cells expressed low levels of TGF-alpha, epidermal growth factor receptor, TGF-beta 1, and TGF-beta 2 mRNA. These studies characterize a novel hormone-nonresponsive cell line which has been established from a hormone-responsive cell line grown under estrogen-free and drug-free conditions. Further analysis of these lines should provide valuable information concerning the development of antiestrogen-resistant breast cancer.

Structure-function relationships of hydroxylated metabolites of tamoxifen that control the proliferation of estrogen-responsive T47D breast cancer cells in vitro.

Several hydroxylated derivatives of tamoxifen were tested for their effects on the growth of T47D human breast cancer cells in vitro. Compounds containing a fused seven-membered ring were used to prevent isomerization of the triphenyl-ethylenes at the double bond. This stable structure permitted the determination of the activity of the cis and trans forms of tamoxifen and the true activity of two of its metabolites, 4-hydroxytamoxifen and metabolite E. Estradiol stimulates the growth of T47D cells 3-4-fold over control after 6 days of treatment (EC50 = congruent to 3 x 10(-12) to 3 x 10(-11) M, depending upon the particular experiment). The fixed ring form of the trans isomer of tamoxifen is an antiestrogen, whereas the cis isomer is estrogenic. Fixed ring-trans-4-hydroxytamoxifen is a potent antiestrogen, and its cis isomer is a weak antiestrogen (IC50 congruent to 4 x 10(-8) to 2 x 10(-7) M). The fixed ring form of trans-metabolite E (tamoxifen without the dimethylaminoethane side chain) is only a weak partial estrogen agonist, whereas the fixed ring derivative of its cis isomer is a potent estrogen agonist (EC50 congruent to 4 x 10(-12) to 1 x 10(-11) M). These studies have determined the true biological activities of the hydroxylated derivatives of tamoxifen. This information will be valuable for the development of drug receptor models and will be particularly useful when the three-dimensional structure of the receptor complex is determined.

The biological significance of the interaction of estrogen agonists and antagonists with the estrogen receptor.

The identification and isolation of a soluble estrogen receptor protein from estrogen target tissues, e.g., uterus, more than 25 years ago provided a unique opportunity for the biochemical pharmacologist to study the interaction of agonist and antagonist molecules in a well-defined organ. Nonsteroidal antiestrogens have proved to be valuable research tools to study cell replication or estrogen specific protein synthesis, e.g., prolactin synthesis or progesterone receptor induction. During the past two decades the antiestrogen tamoxifen has become the endocrine treatment of choice for all stages of breast cancer. This fact has provided an additional incentive to study breast cancer cell replication. In the laboratory breast cancer cells can be used to study the direct actions of antiestrogens without the complicating aspects of pharmacokinetics and metabolism. Studies of the structure-activity relationships of antiestrogens compliment the current investigation of steroid hormone receptors as part of a superfamily of transcription factors. Antiestrogens will, in the future, provide valuable information (from X-ray crystallography) about the three-dimensional folding required if a steroid receptor complex is to prevent cell replication. Clearly this information, and the description of additional transcription factors in hormone-independent cells may prove to be critical for the development of new therapeutic opportunities.

Short- and long-term estrogen deprivation of T47D human breast cancer cells in culture.

The effects of short- and long-term estrogen deprivation on T47D human breast cancer cells was studied. Cells were routinely grown in an estrogenized environment in media containing fetal bovine serum with phenol red indicator. Cells were estrogen deprived (grown in media containing dextran-coated charcoal-stripped fetal bovine serum without phenol red) for either 10 days or at least 8 months, and effects on genotype, receptor content, and cell growth responsiveness were studied. Cells grown in an estrogenized environment are hypertetraploid, whereas long-term estrogen-deprived cells have become hyperdiploid. Short-term estrogen-deprived cells exhibit a decreased growth rate and progesterone receptor (PgR) content, while estrogen receptor (ER) content is not significantly altered. ER mRNA levels are significantly decreased in these cells. Incubation of these cells with estradiol (10(-10) M) for 6 days causes a 5-fold stimulation in cell growth and this stimulation can be inhibited by the antiestrogens 4-hydroxytamoxifen (4-OHT), ICI 164,384, and RU 39411. Cells cultured under long-term estrogen deprivation exhibited an increased growth rate and were refractory to the effects of estradiol and of 4-OHT on cell growth. These cells were ER negative with low levels of PgR; however, one clone of this line was found to be ER and PgR negative. No mRNA for the ER was detected in this line or its clone. With these cell lines it is possible to study the biological characteristics necessary for the outgrowth of a receptor negative, hormone nonresponsive cell population from a receptor positive, hormone-responsive population grown in a estrogen-free environment.

Structural components necessary for the antiestrogenic activity of tamoxifen.

The biological activities of tamoxifen derivatives that contain various side chain alterations were studied using a T47D breast cancer cell growth assay in vitro. We studied the activity of various analogs to determine the important aspects of side chain composition and aryl ring positioning on antiestrogenic activity. Previous studies utilizing a rat pituitary cell prolactin synthesis assay have shown that substitution of the aminoethoxy side chain for an allyl side chain resulted in agonist activity, whereas the addition of a glyceryl side chain produced antiestrogenic activity. In the present study utilizing T47D cells, compounds with alkyl or allyl substitutions were partial agonists, as were compounds with bulky para-substituted benzyl group constituents. A tamoxifen derivative with a side chain containing an ethyl ester was antiestrogenic (IC50 = 2 x 10(-6) M) and effectively inhibited estradiol (10(-10) M) stimulation of growth. However, a compound with a short similar methyl ester-containing side chain did not possess any activity. Compounds with carbinol-containing side chains were antiestrogenic (IC50 = 2.8-3.5 x 10(-7) M). All of the compounds displaying antiestrogenic activity could be "rescued" by incubation with estradiol (10(-8) M) and therefore were not nonspecifically toxic to the cells. These results support the hypothesis that the presence of a lone pair of electrons within the side chain region of tamoxifen may be required for antiestrogenic activity. Also, nonplanar placement of the aryl ring of the triphenylethylene-type of compound is critical for potency.

Estrogenic activity of phenol red.

It has recently been reported that phenol red, a pH indicator present in most tissue culture media, is a weak estrogen that can stimulate some estrogen-sensitive cells. However, the relative impact of phenol red on various cell lines is controversial. We examined the effect of phenol red on several estrogen-responsive cell systems that we use to study estrogen action. These included estrogenic stimulation of progesterone receptor and growth in human breast cancer-derived MCF-7 cells, stimulation of growth in human breast cancer-derived T47D cells, stimulation of prolactin synthesis in primary cultures of immature rat pituitary cells, and stimulation of progesterone receptor in primary cultures of immature rat uterine cells. Estrogenic responses in MCF-7 cells were the most sensitive to the presence of phenol red, while the other three cell cultures showed lesser effects of the indicator. In addition to intrinsic differences in cell responses, there were several other factors involved. These included differences in the estrogenic activity of phenol red-containing media and phenol red itself from different commercial suppliers, and differences in the concentration of free phenol red in final media due to binding of the indicator by serum. Higher concentrations of serum reduced the impact of phenol red on estrogenic responses in primary pituitary cells. Phenol red added to rat uterine cytosol competed with estradiol for binding to the estrogen receptor (relative binding affinity (RBA) approx. 0.001), and the acidic and basic forms of the indicator showed similar activity. Some commercial phenol red samples inhibited cell growth at levels of 100 mg/l; these effects were toxic rather than antiestrogenic, because growth inhibition could not be competitively reversed by an excess of estradiol. The amount of the indicator bound to serum in the final media, the source of the phenol red and the sensitivity of different cell types to the indicator ultimately determine its influence to the response of cells in tissue culture.

Cyclic AMP-resistant mutants of S49 mouse lymphoma cells hemizygous for expression of regulatory subunit of type I cyclic AMP-dependent protein kinase.

For use in studies of the functional organization of regulatory (R) subunit of type I cAMP-dependent protein kinase, 84 independent cyclic AMP-resistant mutants were isolated from sublines of S49 mouse lymphoma cells that are hemizygous for expression of the R subunit. Mutants were characterized by two-dimensional gel analysis of the R subunits, assays of kinase activation, and assays of cAMP-binding. All but eight of the mutants had kinases with increased apparent Kas for cAMP-dependent activation, and studies with site-selective cAMP analogs revealed considerable phenotypic diversity among these mutants. Forty-nine of the mutants had "charge-shift" lesions that mapped to regions of the R subunit polypeptide implicated in cAMP-binding. Twenty-five of the "charge-shift mutants" expressed only mutant R subunits, and the lesions in most of these isolates inhibited binding of cAMP to mutated cAMP-binding sites. The remainder of the charge-shift mutants expressed both mutant R subunit and R subunit with wild-type gel mobilities. The origin of these "heterozygous" mutants from parental "hemizygous" cells remains a puzzle.

Stimulation of breast cancer cells in vitro by the environmental estrogen enterolactone and the phytoestrogen equol.

The phenolic lignans enterolactone and enterodiol appear periodically in women's urine, dependent upon synthesis from plant-derived lignans by the intestinal microflora. The phytoestrogen equol is also present in women's urine, and is also derived from a vegetarian diet. Antiestrogenic or antiproliferative actions of these compounds have been postulated and related to the observation that there is a reduced incidence of breast cancer associated with diet. We evaluated the estrogenic and antiestrogenic activity of these compounds using four sensitive assays in tissue culture, including the use of human breast cancer cell lines T47D and MCF-7. Unexpectedly, we found that enterolactone and enterodiol, as well as equol, are weak estrogens, and that enterolactone and equol could stimulate the growth of estrogen-dependent breast cancer cell lines. We suggest that these environmental agents can promote the growth of breast cancer, particularly hormone-dependent metastases that may be located near the gut or in the mesenteries or liver, where the concentration of these intestinally produced compounds would be highest. Treatment with an antiestrogen such as tamoxifen blocks the estrogenic activity of these compounds. In the absence of treatment with an antiestrogen such as tamoxifen, hormonal therapy to block steroidal estrogen synthesis in a patient with breast cancer could conceivably be circumvented by a vegeterian diet rich in the precursors to estrogenic compounds such as enterolactone and equol.

Activation of type I cyclic AMP-dependent protein kinases with defective cyclic AMP-binding sites.

Two S49 mouse lymphoma cell variants hemizygous for expression of mutant regulatory (R) subunits of type I cyclic AMP-dependent protein kinase were used to investigate functional consequences of lesions in the putative cAMP-binding sites of R subunit. Kinase activation properties of wild-type and mutant enzymes were compared using cAMP and six site-selective analogs of cAMP. Kinases from both mutant sublines were relatively resistant to cyclic nucleotide-dependent activation, but they were fully activable by at least some effectors. Relative resistances of the mutant kinases varied from about 5-fold for analogs selective for their nonmutated sites to as much as 700-fold for analogs selective for their mutated sites; resistance to cAMP was intermediate. Apparent affinities of wild-type and mutant R subunits for [3H]cAMP were not appreciably different, but competition experiments with site-selective analogs of cAMP suggested that binding of cAMP to mutant R subunits was primarily to their nonmutated sites. Analyses of cooperativity in cyclic nucleotide-dependent activation of mutant kinases, synergism between site I- and site II-selective analogs in activating the mutant enzymes, and dissociation of bound cAMP from mutant R subunits provided additional evidence that the mutations in these strains selectively inactivated single classes of cAMP-binding sites: phenomena attributable in wild-type enzyme to intrachain interactions between sites I and II were always absent or severely diminished in experiments with the mutant enzymes. These results confirm that R subunit sequences implicated in cAMP binding by homology with other cyclic nucleotide-binding proteins actually correspond to functional cAMP-binding sites. Furthermore, occupation of either cAMP-binding site I or II is apparently sufficient for activation of cAMP-dependent protein kinase. The presence of four functional cAMP-binding sites in wild-type kinase enhances the cooperativity and sensitivity of cAMP-mediated activation.

Hotspots for spontaneous and mutagen-induced lesions in regulatory subunit of cyclic AMP-dependent protein kinase in S49 mouse lymphoma cells.

From an S49 mouse lymphoma cell subline that carries an electrophoretic marker mutation in one allele for a regulatory (R) subunit of cyclic AMP-dependent protein kinase, 130 cyclic AMP-resistant mutants were isolated and characterized. Of the 77 independent spontaneous and mutagen-induced isolates identified, 74 had kinases with increased apparent activation constants (KaS) for cyclic AMP-dependent activation. The "Ka" phenotype was invariably correlated with an apparent structural lesion in one R subunit allele. "Charge-shift" lesions in 43 independent isolates were mapped to small regions within the R subunit by two-dimensional gel analysis of partial proteolysis peptides. Nine Ka mutations were distinguished by differences in charge or peptide maps of mutant R subunits, and the mutations were clustered in two regions associated with the cyclic AMP-binding sites of the R subunit. The relative frequencies of different mutations differed among spontaneous, ethyl methanesulfonate-induced, and N-methyl-N'-nitro-N-nitrosoguanidine-induced isolates. Mutation frequencies were also markedly different for the two R subunit alleles; this allele preference was strongest for mutagen-induced lesions in the more carboxy terminal cyclic AMP-binding site.