Cathepsin D is partly endocytosed by the LRP1 receptor and inhibits LRP1-regulated intramembrane proteolysis.

The aspartic protease cathepsin-D (cath-D) is a marker of poor prognosis in breast cancer that is overexpressed and hypersecreted by human breast cancer cells. Secreted pro-cath-D binds to the extracellular domain of the ß-chain of the LDL receptor-related protein-1 (LRP1) in fibroblasts. The LRP1 receptor has an 85-kDa transmembrane ß-chain and a noncovalently attached 515-kDa extracellular α-chain. LRP1 acts by (1) internalizing many ligands via its α-chain, (2) activating signaling pathways by phosphorylating the LRP1ß-chain tyrosine and (3) modulating gene transcription by regulated intramembrane proteolysis (RIP) of its ß-chain. LRP1 RIP involves two cleavages: the first liberates the LRP1 ectodomain to give a membrane-associated form, LRP1ß-CTF, and the second generates the LRP1ß-intracellular domain, LRP1ß-ICD, that modulates gene transcription. Here, we investigated the endocytosis of pro-cath-D by LRP1 and the effect of pro-cath-D/LRP1ß interaction on LRP1ß tyrosine phosphorylation and/or LRP1ß RIP. Our results indicate that pro-cath-D was partially endocytosed by LRP1 in fibroblasts. However, pro-cath-D and ectopic cath-D did not stimulate phosphorylation of the LRP1ß-chain tyrosine. Interestingly, ectopic cath-D and its catalytically inactive (D231N)cath-D, and pro-(D231N)cath-D all significantly inhibited LRP1 RIP by preventing LRP1ß-CTF production. Thus, cath-D inhibits LRP1 RIP independently of its catalytic activity by blocking the first cleavage. As cath-D triggers fibroblast outgrowth by LRP1, we propose that cath-D modulates the growth of fibroblasts by inhibiting LRP1 RIP in the breast tumor microenvironment.

Heat shock cognate 70 protein secretion as a new growth arrest signal for cancer cells.

Earlier studies indicated that density-arrested cancer cells released an unidentified growth inhibitor whose secretion was prevented by overexpression of the lysosomal protease cathepsin D (cath D). In this study, this growth inhibitor was purified by affinity chromatography and identified as the heat shock cognate 70 protein (hsc70) based on its peptide microsequencing and specific antibody recognition. Among intracellular proteins, including other heat shock proteins, only constitutive hsc70 was secreted in response to the high-cell density. Moreover, hsc70 secretion from cancer cells was generated by serum deprivation, whereas its cellular concentration did not change. Prevention of Hsc70 secretion by cath D overexpression was associated with the formation of multilayer cell cultures, thus indicating a loss of contact inhibition. In addition, we showed that supplementing the culture medium with purified hsc70 inhibited cell proliferation in the nanomolar range. Conversely, removal of this extracellular hsc70 from the medium by either retention on ADP-agarose or competition at the Hsc70 binding site restored cell proliferation. Hsc70 appears active in human breast cancer cells and hypersecreted by direct cath D inhibition. These results suggest a new role of this secreted hsc70 chaperone in cell proliferation that might account for the higher tumor growth of cancer cells overexpressing cath D.

Estrogen receptor mediated inhibition of cancer cell invasion and motility: an overview.

In this overview of results from our laboratory, we address the question of the role of estrogens during early steps of metastasis, involving cell invasion through the basement membrane and cell motility. The motility of several estrogen receptor (ER) positive breast (MCF7, T47D) and ovarian (BG-1, SKOV3, PEO4) cancer cell lines was studied using a modified Boyden chamber assay. We observed, in all cases, estradiol induced inhibition of cancer cell invasion and motility. A similar inhibitory effect of estradiol was found when the wild-type ER alpha was stably transfected in the ER-negative MDA-MB231 cells and 3Y1-Ad12 cancer cells. The mechanism of this inhibitory effect is unknown. In ovarian cancer, however, it may involve intermediary proteins such as fibulin-1, an extracellular matrix protein that strongly interacts with fibronectin and which is induced by estrogen and secreted by ovarian cancer cells. We conclude that estrogens in ER-positive breast and ovarian cancers have a dual effect, since they stimulate tumor growth but inhibit invasion and motility. This may be consistent with the good initial prognostic value of ER-positive breast cancers compared to ER negative breast cancers noted in several clinical studies.

Breast cancer cell invasiveness: correlation with protein kinase C activity and differential regulation by phorbol ester in estrogen receptor-positive and -negative cells.

Increased protein kinase C (PKC) activity in malignant breast tissue and in most aggressive breast cancer cell lines has suggested a possible role of PKC in breast carcinogenesis and tumor progression. We have investigated here the involvement of PKC in the in vitro invasiveness and motility of several breast cancer cell lines. Modulation of PKC activity by treatment with a phorbol ester (TPA), drastically increased the invasiveness of 2 estrogen receptor-positive (ER+) lines (MCF7 and ZR 75.1), whereas it markedly decreased the invasiveness of 2 ER- cell lines (MDA-MB-231 and MDA-MB-435). A PKC inhibitor (H7) reversed the TPA effects in MCF7 cells, whereas it mimicked TPA action in MDA-MB-231 cells. All of these effects of TPA also were observed to a similar extent for cell chemotaxis, and they were not dependent on protein neo-synthesis. In parallel, short TPA treatment induced cell spreading and microtubule organization in MCF7 cells and inverse morphological changes in MDA-MB-231 cells. In ER+ cells, constitutive PKC activity and PKCalpha expression were very low as compared to ER- cells, and this correlated with the invasive potential of the cells. The opposed effects of TPA in ER+ and ER- cells could be due to the abnormal TPA regulation of PKCalpha observed in ER- cells.

Both estradiol and tamoxifen decrease proliferation and invasiveness of cancer cells transfected with a mutated estrogen receptor.

Previous studies have shown that, after wild-type estrogen receptor (ER) transfection in ER-negative breast cancer cells, estradiol but not tamoxifen prevents growth, invasiveness and metastasis of these cells in mice. Because an ER mutation at position 400 converts the triphenylethylene antiestrogen, OH-tamoxifen into a full estrogen agonist, we transfected this mutated form of human ER in an ER-negative rat cancer cell line. This was aimed at inducing an inhibitory, estrogen-like response of tamoxifen in these cells. In two stable ER-positive transfectants, OH-tamoxifen inhibited cell growth and invasiveness in vitro as efficiently as estradiol. The pure antiestrogen, ICI 164,384, was not agonistic alone and antagonized estrogen action. In contrast, the three compounds were ineffective in control mock-transfected cells. When injected into ovariectomized nude mice, ER-negative mock-transfected cells formed tumours which were significantly stimulated by estradiol and inhibited by tamoxifen treatment. This indicates that estradiol and tamoxifen altered the growth of ER-negative tumours via a general effect on the host response. Surprisingly, the hormone responsiveness of ER-positive tumours developed from ER-transfected cells did not significantly differ from that of ER-negative (mock-transfected) tumours. We conclude that transfection of a mutated human estrogen receptor inhibited, through an estrogenic activity of tamoxifen, the growth and invasiveness of these cancer cells in vitro. However, the low expression of ER did not allowed us to obtain the same effect of tamoxifen in vivo.

Biological and clinical significance of cathepsin D in breast cancer metastasis.

Cathepsin D (cath-D) is an aspartyl lysosomal protease expressed in all tissues. Most metastatic breast cancer cell lines, unlike normal cells, secrete high levels of pro-cath-D. This abnormal secretion is due to both overexpression of the cath-D gene and to an altered processing of the precursor protein. Cath-D gene transcription is increased by estrogen and growth factors in estrogen-receptor-positive breast cancer cells and by an unknown mechanism in estrogen-receptor-negative cells. A large number of independent clinical studies associated high cath-D concentrations in the cytosol of primary breast cancers with increased risk of subsequent metastasis. The amino acid sequence of cath-D analyzed in two breast cancer cell lines is normal, but glycosylation appears to be different with more acidic isoforms. To assess the potential role of this protease in cancer metastasis, we transfected a human cDNA cath-D expression vector in 3Y1-Ad12 embryonic rat tumorigenic cells which did not secrete the proenzyme. A moderate overexpression of human cath-D was sufficient to increase the metastatic potential of these cells in nude mice. The mechanism of cath-D-induced metastasis seems to require maturation of the proenzyme, in endosomes and in large acidic compartments identified as phagosomes. Rather than increase cancer cell escape from the primary tumor through basement membrane degradation as proposed for neutral proteinases, cath-D appears to facilitate cell growth at distant sites. The mechanism of this indirect mitogenic effect is discussed from results obtained in different models. Different cath-D substrates (growth inhibitors, precursors of growth factors, etc.) are proposed to mediate this activity.

Transfected cathepsin D stimulates high density cancer cell growth by inactivating secreted growth inhibitors.

Cathepsin D, a lysosomal protease, is overexpressed in primary breast cancer and associated with increased risk of metastasis. We have shown previously by transfection in rat tumor cells that overexpression of cathepsin D increased both experimental metastasis in nude mice and in vitro proliferation under low-serum conditions. In this study, we used the transfected cell lines to investigate the mechanism by which cathepsin D prevents density-dependent arrest of cell proliferation. This effect was not associated with a general alteration of cell-substratum or cell-cell adhesiveness. As shown by coculture and conditioned media experiments, control cells reaching saturation density released inhibitory activity that was able to prevent the growth of control or cathepsin D transfectants and decreased the cloning efficiency of normal rat kidney fibroblasts in agar. By contrast, in media from two cathepsin D-transfected cell lines, this inhibitory activity was markedly reduced. Cathepsin D overexpression did not affect cell sensitivity to the inhibitor but modified the secretion of several proteins. The increase in cell density appeared to be due to intracellular maturation of cathepsin D since it was reversed by amine treatment that neutralizes the pH of acidic compartments within the cells. Moreover, the addition of secreted pro-cathepsin D was unable to increase the saturation density of control clones. Finally, the inhibitory factor was partially characterized as a heat-labile, secreted protein. We conclude that cathepsin D overexpression increases the growth of cancer cells to a higher density via an intracellular mechanism, leading to a decreased secretion of growth inhibitor(s).

Anti-proliferative and anti-estrogenic effects of ICI 164,384 and ICI 182,780 in 4-OH-tamoxifen-resistant human breast-cancer cells.

The effects of the anti-estrogens 4-hydroxytamoxifen (OHTam), ICI 164,384 and ICI 182,780 were tested on the MCF-7/LCC2 breast-carcinoma cell line, which grows significantly in the presence of OHTam and serves as a model for studying anti-estrogen resistance of estrogen-receptor-positive breast cancer. Cell proliferation and cathepsin-D secretion were strongly inhibited by either ICI 182,780 or ICI 164,384 alone or ICI 164,384 in combination with 17-beta-estradiol (E2) or OHTam. ICI 164,384 alone did not affect the cathepsin-D and pS2 mRNA levels, but antagonized the stimulatory effects of E2 or OHTam on these 2 mRNAs. OHTam was more effective than E2 in increasing cathepsin-D mRNA levels, supporting the idea that anti-estrogen-resistant breast cancer continues to overexpress cathepsin-D. These data show that the steroidal anti-estrogens ICI 164,384 and ICI 182,780 retain their ability to inhibit cell proliferation and the estrogen-responsiveness of cathepsin-D and pS2 genes in the OHTam-resistant MCF-7/LCC2 cell lines. These pure anti-estrogens may thus be efficient second-line treatments of some Tamoxifen-resistant tumors.

Activation of estrogen receptor transfected into a receptor-negative breast cancer cell line decreases the metastatic and invasive potential of the cells.

Breast cancers containing estrogen receptors are responsive to antiestrogen treatment and have a better prognosis than estrogen receptor-negative tumors. The loss of estrogen and progesterone receptors appears to be associated with a progression to less-differentiated tumors. We transfected the human estrogen receptor into the estrogen receptor-negative metastatic breast cancer cell line MDA-MB-231 in an attempt to restore their sensitivity to antiestrogens. Two stable sublines of MDA-MB-231 cells (HC1 and HE5) expressing functional estrogen receptors were studied for their ability to grow and invade in vitro and to metastasize in athymic nude mice. The number and size of lung metastases developed by these two sublines in ovariectomized nude mice was not markedly altered by tamoxifen but was inhibited 3-fold by estradiol. Estradiol also significantly inhibited in vitro cell proliferation of these sublines and their invasiveness in Matrigel, a reconstituted basement membrane, whereas the antiestrogens 4-hydroxytamoxifen and ICI 164,384 reversed these effects. These results show that estradiol inhibits the metastatic ability of estrogen receptor-negative breast cancer cells following transfection with the estrogen receptor, whereas estrogen receptor-positive breast cancers are stimulated by estrogen, indicating that factors other than the estrogen receptor are involved in progression toward hormone independence. Reactivation or transfer of the estrogen receptor gene can therefore be considered as therapeutic approaches to hormone-independent cancers.

Overexpression of transfected cathepsin D in transformed cells increases their malignant phenotype and metastatic potency.

Cathepsin D, an aspartic lysosomal proteinase, is overexpressed by breast cancer cells and highly correlated with the occurrence of metastasis in patients. We used an adenovirus transformed rat cell line that does not secrete cathepsin D, to study the consequences of transfecting human cathepsin D cDNA. Overexpression of human cathepsin D in stable transfectant clones, results in higher initial growth rates in low serum conditions, overgrowth at high cell densities resulting in stellate aggregates, and greater anchorage-independent growth in soft agar. The metastatic activity (mostly in liver) of cathepsin D clones injected into athymic mice was significantly higher than that of control clones. These results show that overexpression of cathepsin D increases the transformed phenotype of malignant cells in vitro and their metastatic potency in vivo.

Cloning and sequencing of the 52K cathepsin D complementary deoxyribonucleic acid of MCF7 breast cancer cells and mapping on chromosome 11.

Two lambda gt11 libraries containing complementary DNAs from human breast cancer MCF7 cells were screened by expression with monoclonal antibodies to the secreted 52K protein and with a 36-mer oligonucleotide derived from the N-terminal amino acid sequence of the secreted 52K protein. Four overlapping clones were sequenced, and found to be extensively homologous to the cathepsin D of normal human kidney, except for 5-point mutations resulting in one amino acid change (Ala to Val) in the profragment of cathepsin D. Northern blot analysis showed the 2.2 kilobase (kb) cathepsin D mRNA to be induced by estradiol in MCF7 cells and produced constitutively at high levels in the estrogen-receptor-negative BT20 cell line. A simple restriction pattern consistent with the restriction map of cathepsin D cDNA was obtained in Southern blot analysis of MCF7 cell DNA. In situ hybridization of the 52K-9 cDNA probe on normal lymphocytes assigned the 52K cathepsin D gene at the extremity of the short arm of chromosome 11, in the p15 band, close to the H-ras gene and in the region whose deletion increases the risk of invasive breast cancer. We conclude that the estrogen induced 52K protein has the same sequence as normal pro-cathepsin D and we propose that the 52K protein correspond to the only pro-cathepsin D expressed in MCF7 cells.

Difference between R5020 and the antiprogestin RU486 in antiproliferative effects on human breast cancer cells.

We have compared the effects of the progestin R5020 and the antiprogestin RU486 on the growth of the MCF-7 and T47D breast cancer cell lines. Differences between the two compounds were demonstrated in several parameters. 1. Estradiol was required for the efficient inhibition of cell growth of both lines by R5020 but not by RU486. Therefore in the total absence of estrogen (phenol-red free medium), the effects of the two drugs on cell growth were dissociated, RU486 remaining inhibitory while R5020 was inactive. 2. The proteins secreted by cells were differently affected, since R5020 induced a 48K protein and decreased the production of the estrogen-regulated 52K protein, while RU486 had no effect on these two parameters. 3. The morphology of cells treated by R5020 was more altered in the presence of estradiol than in its absence, while that of cells treated by RU486 was not affected whether or not estradiol was present. 4. There was a greater reduction of estrogen receptor sites in MCF-7 cells produced by R5020 than by RU486. Even though the two drugs appear to act through the same progesterone receptor and to inhibit total protein secretion, it is likely that they exert their antiproliferative effects on cultured breast cancer cells by different mechanisms. R5020 antagonizes the stimulation produced by estradiol. RU486 by contrast exerts a more direct progesterone receptor mediated inhibitory effect requiring no synergism by estradiol and therefore does not act through a partial progestin activity.

Immunohistochemical distribution of the 52-kDa protein in mammary tumors: a marker associated with cell proliferation rather than with hormone responsiveness.

We have previously described a secreted glycoprotein of mol. wt 52,000 (52-kDa protein) which is induced by estrogen in some human breast cancer cell lines. This protein has been identified as the proenzyme of a lysosomal cathepsin-D-like protease which is secreted in large proportions in breast cancer cells. To determine which information may be generated by this marker when detected in mammary tumors, in comparison with hormone receptors, we used monoclonal antibodies interacting specifically with the 52-kDa protein and its related cellular processed products (mols. wts 48 and 34 kDa). A high concentration of this protein has been shown in proliferative ductal mastopathies and cysts, suggesting its value in detecting high-risk mastopathies. We now present the immunoperoxidase distribution of this protein in breast carcinoma compared to the cytosolic hormone receptors assayed in parallel. In 232 breast cancers, no correlation was found between the cellular 52-kDa protein content and cytosolic estrogen or progesterone receptor concentrations. This absence of correlation was also shown by the constitutive production of this protein by estrogen-receptor-negative breast cancer cell lines and confirmed by double immunostaining of breast cancer cell aspirates showing a dissociation between the cytoplasmic staining of this 52-kDa lysosomal protease and the nuclear staining of the estrogen receptor. These clinical results, associated with the in vitro mitogenic and proteolytic activities of this protein, strongly suggest that the 52-kDa protein staining in tissue is associated with tumor proliferation and/or invasion, rather than with hormone responsiveness.

Characterization of monoclonal antibodies to the estrogen-regulated Mr 52,000 glycoprotein and their use in MCF7 cells.

The Mr 52,000 glycoprotein is regulated by estrogen and released by breast cancer cells in culture (B. Westley and H. Rochefort, Cell, 20: 352-362, 1980). This rare protein was partially purified from 25 liters of medium conditioned by MCF7 cells and injected into Biozzi's selected mice. The spleen lymphocytes of one immunized mouse was fused with the murine myeloma P3-X63-Ag8-653. Sixteen hybridomas producing monoclonal antibodies to the Mr 52,000 protein were isolated, and seven of them were cloned and purified. The seven monoclonal antibodies were all of the immunoglobulin G1 isotype, and their dissociation constants ranged from 0.35 to 2.3 nM. The antibodies specifically recognized the secreted Mr 52,000 protein as evidenced by double immunoprecipitation and by immunoblotting after electrophoretic separation and transfer. Double-determinant immunoradiometric assay indicated that the seven purified monoclonal antibodies recognized three distinct regions of the Mr 52,000 protein, and it was used to assay the Mr 52,000 protein in biological fluids. These antibodies did not react with the external plasma membrane of MCF7 cells, as shown by immunofluorescence analysis. By contrast, the cytoplasm of MCF7 cells (but not T47D and RBA cells) was stained by the peroxidase-immunoperoxidase complex after plasma membrane permeation, indicating that the protein is secreted by exocytosis rather than shed from the plasma membrane.

Induction of two estrogen-responsive proteins by antiestrogens in R27, a tamoxifen-resistant clone of MCF7 cells.

In an attempt to approach the mechanism by which estrogen and antiestrogen regulate the growth of estrogen receptor-positive breast cancer, we have studied the effects of the antiestrogens tamoxifen (TAM) and 4-hydroxytamoxifen (OH-TAM) on the induction of two estrogen-specific proteins in a variant of MCF7 cells termed R27, cloned for its ability to grow in the presence of TAM. We found that, in the R27 variant, antiestrogens as well as estradiol were able to increase specifically the production of a Mr 52,000 protein which was released into the culture medium. This protein was shown to be identical to the Mr 52,000 glyco-protein induced by estrogen and released by MCF7 cells on the basis of its specific inducibility by physiological concentrations of 17 beta-estradiol and of its resolution in two-dimensional gel analysis. Dose-response analysis showed that, in the R27 variant, TAM and OH-TAM acquired the ability to induce the Mr 52,000 protein at concentrations compatible with their relative affinities for the estrogen receptor, while these antiestrogens were inefficient in the wild MCF7 cells. Whereas the relative increase of the Mr 52,000 protein was similar with TAM, OH-TAM, and 17 beta-estradiol, the general production of Mr 52,000 and of total labeled proteins was less with the antiestrogens than with 17 beta-estradiol. Moreover, OH-TAM displayed a biphasic dose-response curve with inhibitory effects at concentrations above 10 nM, suggesting an additional mechanism. Neither TAM nor OH-TAM had any antiestrogenic effect when added in the presence of 17 beta-estradiol. In the R27 variant, both estradiol and antiestrogens induced the progesterone receptor sites; however, the extent of the stimulation was lower with antiestrogens than with estradiol. This study shows that, in addition to the classical antiestrogen-resistant breast cancer cells, in which the estrogen receptor is absent or inactive, there is another class of antiestrogen-resistant cells in which the drug becomes a full estrogen agonist as evidenced by the induction of the Mr 52,000 estrogen-specific protein, which is not induced in the wild-type cells.

The antiproliferative effect of tamoxifen in breast cancer cells: mediation by the estrogen receptor.

The effects of tamoxifen (Tam) and its 4-hydroxylated metabolite (OH-Tam) on the growth of two human breast cancer cell lines ( MCF7 and BT20 ) were evaluated by fluorometric DNA assay. The effects of the antiestrogens were dependent upon their concentrations and the nature of the cells. At concentrations below 4 microM, the degree of inhibition was related to their relative affinities for the estrogen receptor and was totally reversed by estradiol in MCF7 cells. No inhibition was observed in the estrogen receptor negative cell line BT20 . This supports and extends the idea that the antiproliferative effect of Tam at these concentrations is mediated by the estrogen receptor even in the absence of measurable estradiol concentration. At concentrations greater than 4 microM, Tam was cytotoxic on MCF7 and BT20 mammary cell lines within 2 days of treatment. The cytotoxic effect was irreversible and was not prevented by occupation of the estrogen receptor with estradiol, suggesting that it was not mediated by the estrogen receptor. The cytotoxicity of the triphenylethylene drugs, however, has some specificity since it was not observed in a fibroblast rat cell line ( 49F ) or in the two mammary cell lines with similar high concentrations of estradiol and diethylstilbestrol.

[Estrogen-induced proteins secreted by MCF7 human breast cancer cells stimulate their proliferation]. / Les protéines oestrogéno-induites sécrétées par les cellules mammaires cancéreuses humaines MCF7 stimulent leur prolifération.

In the MCF7 human breast cancer cells, estrogens stimulate in vitro cell proliferation and biosynthesis of specific proteins which are released into the culture medium; the most abundant of these is a 52,000 dalton glycoprotein. Conditioned media from E2-treated MCF7 cells stimulate the growth of E2-deprived MCF7 cells, whereas conditioned media from untreated control cells are ineffective. At protein concentration greater than 5 micrograms/ml, an inhibitory effect was also obtained with control media. The mitogenic factors of these conditioned media are non-dialysable, sensitive to temperature and proteases and adsorbed on Con A Sepharose. The cell surface of MCF7 cells, examined by scanning electron microscopy, is differentially affected by these two types of conditioned media. We conclude that estrogens may stimulate the growth of breast cancer cells indirectly via the induction of glycoproteins acting as growth factors.

Effect of estradiol on the ultrastructure of the MCF7 human breast cancer cells in culture.

We have analyzed the effect of estradiol and of two classes of antiestrogens on the morphology of the MCF7 human breast cancer cell line by scanning and transmission electron microscopy. Estradiol progressively increased the number and the length of microvilli at the cell surface. The density of the microvilli network increased between 2 and 11 days of estrogen treatment, while the cells became more granular and less tightly attached to the surface of the dish. Estradiol also progressively transformed cells into secretory cells containing, at Day 2, large, clear mitochondria and, at Day 4, rough endoplasmic reticulum and Golgi complex. At Day 6, secretory granules (diameter, 0.2 microM), which mainly contained glycoproteins, were first developed in the cytoplasm. By Day 8, they were concentrated at the cell membrane and being liberated into the medium. Larger granules (diameter, 0.8 microM), which probably contained lipids, were obtained later (Day 11). Cell cultures in 10% fetal calf serum not treated by charcoal contained secretory granules. The modifications were induced by physiological concentrations of estradiol but not 5 alpha-dihydrotestosterone. Progesterone (10 nM for 8 days) completely inhibited the effect of estradiol on the microvilli and secretory activity. Tamoxifen or hydroxy-tamoxifen did not induce secretory activity but did alter the cell morphology compared to control cells. The effects of estradiol were observed in other estrogen receptor-positive breast cancer cell lines (ZR 75-1, T 47 D) but not in an estrogen receptor-negative cell line (BT 20). This morphological evidence that estrogens modify the cell surface of breast cancer cells in culture and transform them into "secretory cells" complements evidence that a molecular weight of approximately 50,000 into the culture medium (Cell, 24: 352-362, 1980). (The molecular weight was found first to be 46,000. It seems to be closer to 52,000 in a 10% polyacrylamide gel and by using the NEN-labeled proteins as molecular weight markers.