Changes in expression of steroid receptors, their downstream target genes and their associated co-regulators during the sequential acquisition of tamoxifen resistance in vitro.

Tamoxifen resistance (TAMr) in breast cancer is a serious clinical dilemma, with no satisfactory explanation. We hypothesised that changes in the expression of steroid hormone receptors (ERalpha, ERbeta), their downstream target genes (PR, pS2) and their associated co-regulators (AIB-1, SRC-1, SRA, NCoR-1, SMRT and REA) could be related to the acquisition of TAMr. To test this hypothesis, we developed in vitro TAMr cell line models by continuous exposure of MCF-7 cells to 4-hydroxytamoxifen (4-HT) over 12 (MCF-7MMU1) and 21 (MCF-7MMU2) months, respectively and examined the expression of the above by Western blotting and immunohistochemistry. In addition, we further examined the changes in global gene expression in TAMr cells in comparison with TAM-sensitive cells by microarray analysis. We report here that acquisition of TAMr is associated with changes in the expression of PR, pS2 and several co-activators, but not ERs. In addition, genes associated with cell cycle, cell adhesion and extracellular matrix, were up-regulated while those associated with apoptosis or growth factors/hormones were down-regulated. Based on our results, it appears that increased co-activator expression, in concert with alterations in genes associated with controlling cell proliferation and survival contribute to TAMr in breast cancer.

Expression of alternatively spliced estrogen receptor alpha mRNAs is increased in breast cancer tissues.

We previously identified 20 different alternatively spliced estrogen receptor alpha (ERalpha) mRNAs that have deletions in various combinations of exons in breast cancer cell lines using a novel 'Splice Targeted Approach' [J. Steroid Biochem. Mol. Biol. 72 (2000) 249]. In the current study, we compared the frequency of alternatively spliced ERalpha variant expression in 35 reduction mammoplasty and 38 breast cancer tissues with known ERalpha, ERbeta status using this highly specific 'Splice Targeted Approach'. A total of 16 different alternatively spliced variants were identified that have deletions in various combinations of exons in normal, as well as cancer tissues. However, not all 16 variants were present in every tissue. The frequency and type of variants in normal and cancer tissues was significantly different. Majority of normal tissues expressed only single exon deletion variants with the exception of those in combination with exon 2Delta and 7Delta. Tumor tissues, on the other hand, showed increased frequency of multiple exon deletion mRNAs (P<0.019). In addition, cancer tissues also showed an increased frequency of all variants compared with normal tissues (P<0.044). Among the 16 variants, the dominant negative variant, exon 3Delta, showed the most significant increase in cancer tissues (P=0.000032). Specifically, we detected four different mRNAs that have exon 2 deletion-exon 2Delta; 2 and 4Delta; 2 and 5Delta; and 2, 4-5Delta in various combinations in both normal and cancer tissues. A large number of normal tissues expressed two transcripts-exon 2Delta and 2, 4-5Delta. The multiple exon deletion 2, 4-5Delta were predominant in cancer tissues. Only the single exon 3 deletion variant, exon 3Delta, was detected in normal tissues. Cancer tissues showed the presence of a double exon deletion variant, exons 3 and 7Delta, in addition to exon 3Delta. A small fraction of normal tissues showed exons 2-3Delta mRNAs, whereas, cancer tissues showed increased frequency of exons 2-3Delta expression in addition to a triple exon deletion variant, exons 2-3, and 5Delta. The expression of exon 4Delta; or 4 and 7Delta or both was equivalent in normal and cancer tissues. Exon 5Delta transcripts were present at very low levels in both normal and tumor tissues. A small percentage of cancer tissues but not normal tissues showed exon 6Delta mRNA. The presence of single, double, triple and quadruple exon deletion mRNAs, exon 7Delta; 7 and 4Delta; 7, 3-4Delta; 7, 3-5Delta, respectively, were detected in normal as well as cancer tissues. Each normal and cancer tissue had a distinct profile of ERalpha wild type, ERbeta wild type and ERalpha splice variants. Heterogeneity in ER isoform profiles may result in variations in estrogen/anti-estrogen binding and activation/inactivation of estrogen-dependent genes, and, therefore, may have implications in the risk of developing breast cancer, survival with the disease and response to anti-estrogen, as well as other therapies.

Quantitation of alternatively spliced estrogen receptor alpha mRNAs as separate gene populations.

Estrogen receptor (ER) mRNA undergoes alternative splicing generating transcripts that have deletions in various combination of exons. Although several reports have shown that the spliced variant mRNAs are expressed in both normal and malignant tissues, the exact functional role(s) of these molecules have not been established in estrogen induced signal transduction processes mainly due to practical limitations involved in their detection and quantitation. We have recently described a 'Splice Targeted Primer Approach' that can specifically detect splice variants without amplifying the wild type ERs [12]. In the current report, we describe strategies to quantify individual splice variant mRNAs as separate gene populations independent of wild type or other variants using ER exon 7Delta and exon 2Delta as models. We describe the methods of quantifying the exon 7Delta and exon 2Delta transcripts in two breast cancer cell lines, MCF-7 and LCC2, and a breast tumor using the splice-targeted primers in combination with template competition RT PCR. The exon 2Delta splice specific sense primer along with an anti-sense primer in exon 4 amplified a 412 bp product in both cell lines and the tumor that could be quantitated. The exon 7Delta splice targeted anti-sense primer along with a partner primer in exon 2 amplified four transcripts that have deletions in exon 7, exons 7 and 4, exons 7 and 3-4, and exons 7 and 3-5. These four transcripts could be simultaneously quantified by the template competition method described here. Our results also show that the estrogen-independent LCC2 cells express significantly higher levels of the above 7Delta transcripts compared to the estrogen-dependent MCF-7 cells.

Alterations in the estrogen receptor alpha mRNA in the breast tumors of African American women.

Several recent reports have shown that the mortality rate with breast cancer is about three times higher in African American women than in other populations. In addition, the available data also indicate that the tumors are very aggressive and poorly differentiated with a very low frequency of hormone receptors. To gain an insight into the factors that may be responsible for their aggressive tumors, we investigated the transcript profiles of the estrogen receptor (ER), the most important prognostic factor in breast cancer, in the tumors derived from African American women. We analyzed 24 immunohistochemically ER+ and 6 ER- malignant tumors for ER mRNA by reverse transcription polymerase chain reaction using a number of primer pairs. For comparative purposes, 20 ER- malignant tumor issues derived from Caucasian patients were also included. Our results showed that only 15 of the ER+ tumors from African American women patients had full-length wild-type receptor transcripts and the others exhibited alterations/truncations in exon 8. We also found that the majority of tumors that had alterations/truncations in exon 8 did not express the naturally occurring, more abundant exon 7 deletion transcript. Most of the tumors expressed exon 2, exons 2-3, and exon 5 deletion variant transcripts. Unexpectedly, 2 of the 6 immunohistochemically ER- tumors showed full-length wild-type receptor mRNA but none of the variant transcripts.

Identification of twenty alternatively spliced estrogen receptor alpha mRNAs in breast cancer cell lines and tumors using splice targeted primer approach.

Estrogen receptor (ER) alpha splice variant transcript profiles were analyzed by RT PCR in six ER positive breast cancer cell lines, MCF-7, T47D, ZR-75, LCC1, LCC2 and LCC9, three ER negative cell lines, MDA-MB-435, MDA-MB-235 and LCC6, and three ER positive malignant breast tumors using targeted primers which specifically anneal to the splice junctions of exon 2Delta, exon 3Delta, exons 2-3Delta, exon 4Delta, exon 5Delta, exon 6Delta and exon 7Delta. The partner primers were chosen such that largest possible transcripts were amplified between exons 1 and 8. The results described here show that each splice specific primer amplified not only the single exon deleted transcript but also a number of related transcripts that have deletions in various combinations of exons. The exon 2Delta specific primer amplified five transcripts that have deletions in exon 2, exons 2 and 7, exons 2, 5, and 7, exons 2 and 4-5, and exons 2 and 4-6. The exon 3Delta specific primer amplified two transcripts that have deletions in exon 3, and exons 3 and 7. The exon 2-3Delta specific primer amplified three products that have deletions in exons 2-3, exons 2-3 and 7 and exons 2-3, 5 and 7. The exon 4Delta specific primer amplified two products that have deletions in exon 4, and exons 4 and 7. The exon 5Delta specific primer amplified three transcripts, that have deletions in exon 5, exons 5 and 2, and exons 5, and 2-3. The 6Delta specific primer amplified only one transcript that has a deletion in exon 6. The 7Delta specific primer amplified four transcripts, that have deletions in exon 7, exons 7 and 4, exons 7 and 3-4, and exons 7 and 3-5. None of the above splice specific primers amplified the wild type ER sequences. The six ER positive cell lines differed in the patterns of the variant transcripts and among the three ER negative cell lines analyzed, only MDA-MB-435 showed the presence of exon 2Delta and exon 4Delta transcripts. Analyses in the tumor samples indicated that the above transcripts are extensively modified.

Expression of neuron-specific enolase, chromogranin A, synaptophysin and Leu-7 in lung cancer cell lines.

Recent reports indicate that Small Cell Lung Cancer (SCLC) responds favorably to chemotherapy, and this response is associated with the expression of neuroendocrine (NE) markers. However, the currently available reports on the expression of NE markers in SCLC and Non-Small Cell Lung Cancer (NSCLC) are highly inconsistent and need to be reevaluated. Therefore, we have undertaken the current study to clearly define the expression of NE markers in SCLC and NSCLC using representative cell lines. We studied the NE markers, chromogranin A, Neuron specific enolase (NSE), Leu-7 and synaptophysin using various immunochemical and molecular biological methods. By employing Western blotting method, we found that both SCLC and NSCLC cells express large amounts of NSE and chromogranin A. The natural killer cell surface associated antigen (Leu-7) was expressed specifically only by SCLC cell lines. This finding was further confirmed by immunofluorescence staining of the SCLC cells. We could not detect any synaptophysin levels in any of the above cell types by Western blotting technique. Therefore, we employed reverse transcription polymerase chain reaction (RT.PCR) to study the expression of synaptophysin mRNA and found that both SCLC and NSCLC cells express it, albeit in different amounts.

Quantitation of estrogen receptor mRNA copy numbers in breast cancer cell lines and tumors.

Several clinical studies have suggested that the content of estrogen receptor (ER) in breast tumors influences the survival, tumor recurrence, and response to antiestrogen therapies. Therefore, the ability to precisely quantitate the ER content in tumor tissues will be of significant benefit to women with breast cancer. Although immunohistochemical and polymerase chain reaction (PCR) methods have been described for the detection and semiquantitation of ER, none of them precisely quantitate ER copy numbers in tumor samples. In the present report we describe a molecular approach to accurately quantitate ER mRNA copy numbers using a reverse-transcription PCR (RT-PCR) template competition method. A competitor template was devised by inserting unrelated nucleic acid sequences into an ER cDNA clone. A template competitive RT-PCR analysis was then performed to determine the number of copies of ER mRNA. As a standard of reference for the ER mRNA copy numbers from various samples, the mRNA copy numbers of a constitutively expressed gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), were also quantitated. The ER quantitations were performed in three positive cell lines, MCF-7, T47D, and ZR-75, and two positive tumor tissues by this approach. Our results described here show that among the cell lines studied, T47D expresses the highest copy numbers of ER. We also present here that ER as low as 10(3) copies per 10(5) copies of GAPDH can be detected and quantitated in tumor samples by the template competition method. In addition, the molecular approach can simultaneously detect, distinguish, and quantitate exon deletion variant copy numbers of ER. The results described in this report indicate that the ratios of exon 7 deletion variant to wild type in the tumor tissues are significantly higher than in the cell lines studied.

An estrogen inducible 104 kDa chaperone glycoprotein binds ferric iron containing proteins: a possible role in intracellular iron trafficking.

We have previously described an estrogen inducible, intracellular, homodimeric membrane glycoprotein (subunit Mr 104 kDa) which is structurally related to 'chaperone' proteins (Poola, I. and Kiang J.G., J. Biol. Chem. 269 (1994) 21762-21769). In this report we describe a novel finding that the 104 kDa chaperone protein exhibits affinity for iron containing proteins such as transferrins from several species, human lactoferrin and microbial ferric binding protein (FBP). A single ferric ion in the above proteins appears to be sufficient for binding. It also binds to immobilized ferritin. However, it does not exhibit any affinity for apotransferrins, apolactoferrin, apoferritin and apoFBP. This is the first report of a chaperone protein that exhibits affinity for iron containing proteins.

Ligand binding kinetics of a soluble full-length murine erythropoietin receptor.

The full-length murine erythropoietin receptor was expressed in Sf9 cells using a baculovirus vector. Erythropoietin receptors in solubilized Sf9 cell lysates bound erythropoietin with high affinity (92 pM). Erythropoietin receptor-125I-labeled erythropoietin association and dissociation kinetics using solubilized Sf9 cell lysates revealed a ka of 0.16 nM-1 min-1 and a kd of 0.00055 min-1 giving an observed KD of 3.45 pM. The erythropoietin receptors was partially purified from Sf9 cell lysates by chromatography on Con A Sepharose. When erythropoietin receptors were crosslinked to 125I-labeled erythropoietin and analyzed by SDS-7.5% PAGE protein complexes of 90 and 125 kDa were observed with receptors in solubilized lysate, and 170 and 190 kDa with the partially purified receptors.

The estrogen-inducible transferrin receptor-like membrane glycoprotein is related to stress-regulated proteins.

It was previously shown that estrogen induces a membrane glycoprotein (molecular mass, 95 kDa) in the chicken oviducts, which exhibits several properties similar to transferrin receptors (Poola, I., and Lucas, J. J. (1988) J. Biol. Chem. 263, 19137-19146). In the present study, we have further investigated its molecular and transferrin binding properties. We have sequenced several internal peptides isolated from the purified protein by endopeptidase Lys-C. We have found that it has a high degree of sequence homologies with those of chicken heat-shock protein (cHsp108), mouse endoplasmic reticulum protein (mERp99), hamster glucose-regulated protein (hagrp94), and human tumor rejection antigen (hTRAgp96), all of which are shown to be highly homologous to each other and to yeast hsp90. We demonstrate here that the [35S]methionine-labeled immunoaffinity-purified estrogen-inducible membrane glycoprotein binds to the transferrin affinity columns similar to iron-modulated transferrin receptors. Indirect immunofluorescence microscopic studies indicate that it is an intracellular glycoprotein unlike transferrin receptors. We have isolated two molecular forms of the protein, with molecular masses of 116 and 104 kDa, by immunoaffinity column purification, immunoprecipitation, Western blotting, and pulse-chase labeling analyses. Both 116-and 104-kDa species bind transferrin. This protein can be induced by heat-shocking the oviduct cells at 45 degrees C for 3h and recovering at 37 degrees C for 2-3 h. It is also expressed in the human breast cancer cell lines, MCF-7 and T-47D. All these properties taken together strongly suggest that the estrogen-inducible membrane glycoprotein is a novel transferrin-binding protein, structurally related to the stress-regulated proteins.

The chicken oviduct and embryonic red blood cell transferrin receptors are distinct molecules.

We recently described an estrogen-inducible transferrin receptor from the chicken oviduct. We now report on the comparison of the oviduct transferrin receptor with the transferrin receptor obtained from chick embryo red blood cells. Western blot analysis reveals that rabbit polyclonal antibodies raised against one receptor do not cross react with the heterologous receptor. Furthermore, peptide map analyses of either affinity purified, native [125I]-labelled transferrin receptors (dimers) or dissociated, and repurified monomers obtained from oviducts and embryonic red blood cells yield distinct patterns. Therefore, the estrogen-modulated oviduct transferrin receptor appears to be structurally distinct from the iron-modulated red cell transferrin receptor.

Purification and characterization of an estrogen-inducible membrane glycoprotein. Evidence that it is a transferrin receptor.

A number of N-linked membrane glycoproteins are induced during chick oviduct differentiation. We have purified a major estrogen-inducible glycoprotein (Mr = 91,000) to homogeneity by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Comparison of partial NH2-terminal sequence data with membrane glycoproteins having similar Mr showed a limited homology with human and murine transferrin receptors. We observed that oviduct membranes contain estrogen-inducible transferrin receptor activity (Kd = 2-8 x 10(-8) M). Analytical purification of the putative receptor on an ovotransferrin-Affi-Gel affinity column and sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis reveals a protein of Mr, 180,000, which contains two disulfide-linked subunits of Mr 91,000. The receptor reacts very strongly with antibodies prepared against the 91-kDa glycoprotein on Western blots. Western blot analysis confirms that the 91-kDa glycoprotein is induced by estrogen. The protein has 2% total carbohydrate with Man, GlcNAc, Gal, GalNAc, and NeuAc in a molar ratio of 6:4:2:1:1. The protein contains at least one O-linked moiety. Analysis of the O-linked moiety by glycosidase digestions and gel filtration indicates there are sialo tetra- and trisaccharides and a neutral disaccharide(s). Labeled N-linked glycopeptides were prepared by pronase digestion, beta-elimination, and 3H-acetylation. The N-linked oligosaccharides include high mannose and complex neutral nonbisected biantennary types in an approximate ratio of 3:1 as determined by serial lectin affinity chromatography.

Interaction of asparagine-linked oligosaccharides with an immobilized rice (Oryza sativa) lectin column.

The carbohydrate-binding specificity of rice (Oryza sativa) lectin was investigated by testing the ability of radioactively labelled glycopeptides and oligosaccharides to bind to a rice lectin-Sepharose 4B column. Rice lectin binds asparagine-linked oligosaccharides through the core NN'-diacetylchitobiose moiety. Whereas beta 1-4-mannose enhances the binding strength only to a small extent, alpha 1-3-linked core mannose increases it considerably. A core alpha 1-6-linked mannose residue has a slightly inhibitory effect. Binding is not affected when one or both of the alpha-mannose residues are substituted with mannose at C-2, C-3 and C-6 or with N-acetylglucosamine (GlcNAc) at C-2 positions. The presence of an alpha 1-6-fucose residue attached to the asparagine-linked GlcNAc also does not affect the binding. The binding of complex biantennary glycopeptides is not altered by the presence of one or two galactose residues in the non-reducing terminus, but the presence of one or two sialic acid residues decreases the binding capacity. A bisecting beta 1-4-linked GlcNAc attached to beta-linked mannose residue enhances the binding of sialo, asialo and asialoagalacto complex biantennary-type glycopeptides. Bisected hybrid-type glycopeptides bind very tightly to a rice lectin-Sepharose 4B column: Substitution of alpha 1-3-mannose residue at C-2 and C-4 with GlcNAc completely inhibits the binding of both bisected and non-bisected complex asparagine-linked glycopeptides. O-Glycosidically linked oligosaccharides containing GlcNAc bind very weakly to a rice lectin column. The applicability of immobilized rice lectin columns in the fractionation of asparagine-linked oligosaccharides is discussed.

Purification and saccharide-binding characteristics of a rice lectin.

A lectin was purified from rice flour by aqueous extraction followed by precipitation by ammonium sulfate and affinity chromatography on p-aminobenzyl 2-acetamido-2-deoxy-1-thio-beta-D-glucoside-succinyl-aminohexylaminyl -Sepharose 4B. The molecular weight of the lectin is approximately 36,000, as determined by sedimentation-equilibrium analysis. It is a tetramer consisting of two different subunits (Mr = 12,000 +/- 1,000 and 9,000 +/- 1,000). Amino acid analysis indicated that the lectin contains very high proportions of half-cystine, glycine, and glutamic acid. All of the half-cystines are present as -S-S- bridges. The lectin agglutinates human A, B, AB, and O erythrocytes, rabbit erythrocytes, human leukocytes, and is mitogenic to human lymphocytes. The hemagglutinating activity of rice lectin is inhibited by 2-acetamido-2-deoxy-D-glucose, methyl 2-acetamido-2-deoxy-beta-D-glucoside, chitobiose, and chitotriose. N-Acetylneuraminic acid was a noninhibitor, but N-acetylneuramin-(2----3)-lactose showed weak inhibition. The agglutinating activity was also inhibited by various sialoglycoproteins. The immobilized rice-lectin bound glycophorin, alpha 1-acid glycoprotein, and fetuin. Asialoglycophorin, asialofetuin, ovomucoid, and human chorionic gonadotropin were bound only partially to the column.