| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrinology |
Department of Pharmacology, Case Western Reserve University, Cleveland, Ohio 44122 [B. M. W., M. M. M.], and Department of Pathology, University Hospitals of Cleveland, Cleveland, Ohio 44106 [N. W.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
was used in yeast two-hybrid screenings. These screenings resulted in the identification of a novel antiproliferative protein, estrogen down-regulated gene 1 (EDG1), of which the mRNA and protein were shown to be down-regulated directly by estrogens. Our studies additionally suggested an important role for EDG1 in ER-mediated breast cancer development. Analysis of 43 invasive breast cancer samples and 40 adjacent normal breast samples demonstrated EDG1 protein levels to be significantly higher in normal breast epithelial tissue as compared with breast epithelial tumor tissue. EDG1 expression levels were also correlated with the proliferation activity and ER status of the tumors to examine the prognostic value of EDG1 in invasive breast tumors. EDG1 expression was more disassociated from proliferative activity as compared with ER expression in tumor cells. A growth regulatory function for EDG1 is additionally indicated by studies wherein overexpression of EDG1 protein in breast cells resulted in decreased cell proliferation and decreased anchorage-independent growth. Conversely, inhibiting EDG1 expression in breast cells resulted in increased breast cell growth. Thus, we have identified a novel growth inhibitor that is down-regulated by estrogens and colocalizes with ER in breast tissue. These studies support a role for EDG1 in breast cancer. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although estrogens play an important role in the initiation and development of breast cancer, the exact mechanism(s) by which estrogens regulate mammary epithelial cell proliferation is not well-defined (6)
. Some of the genes induced by estrogen, cyclin D1, c-myc, and prothymosin-, may partially explain the ability of estrogen to stimulate cell proliferation (7, 8, 9)
. It is also possible that estrogen could stimulate cell proliferation by down-regulating genes that may inhibit cell proliferation. Protein expression of Hairy and Enhancer of Split homologue-1 (of which the down-regulation appears to be necessary for estrogen-induced cell proliferation) is suppressed by estrogen in breast cancer cells (10)
. Although estrogen suppresses the expression of this protein, only autologous down-regulation of ER has been shown to be a direct effect of estrogen (11)
.
We have identified a protein that inhibits cell growth and is down-regulated directly by estrogens, termed EDG1. We also report on findings that suggest that altered expression (between normal and tumorigenic breast epithelial tissue) of EDG1 is involved in the initiation and/or development of breast tumors. Previous studies have demonstrated that breast tumors expressing ER have better prognosis and are often responsive to hormonal therapies (Ref. 12
; as reviewed in Ref. 13
). Proliferative activity of breast tumors, as measured by Ki67 expression, appears to be an indicator of poor prognosis in node negative breast cancer patients (14)
. We show that breast tumor cells expressing EDG1 were more likely to express ER than Ki67. This suggests a possible role for EDG1 as both an antigrowth and prognostic tool in breast cancer.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Plasmids.
The EDG1 clone, pAD-GAL4-2.1-EDG1, obtained from yeast two-hybrid screening contains the coding sequence cloned in frame with the activation domain of GAL4 in the pAD-GAL4-2.1 phagemid vector (Stratagene, La Jolla, CA). The human homologue clone for EDG1, HEXIM1, was obtained from American Type Culture Collection. The EDG1 coding sequence was released by an NcoI/XbaI digest. The NcoI/XbaI-digested and blunted EDG1 fragment was inserted into EcoRI digested and blunted pBD-Gal4-cam vector (Stratagene) to make pBD-EDG1. pET-15b-EDG1, which encodes the full length EDG1 in frame with six NH2-terminal His-tag sequences, was constructed by inserting the EcoRI/SalI-digested and blunted EDG1 cDNA from pBD-EDG1 into the BamHI-digested and blunted pET-15b vector (Novagen, Madison, WI). The NcoI/XbaI-digested and blunted EDG1 fragment was inserted into BamHI-digested and blunted pBPSTR1 retroviral vector (16)
in the sense or antisense direction to make pBPSTR1-EDG1 or pBPSTR1-EDG1AS, respectively. pEGFP-EDG1, which encodes full-length EDG1 in frame with the coding sequence for GFP, was constructed by inserting NcoI/XbaI-digested and blunted EDG1 fragment into HindIII-digested and blunted pEGFP-C3 vector (Clonetech, Palo Alto, CA).
Yeast Two-Hybrid Screenings.
The yeast two-hybrid screenings used to identify ER-interacting clones were described previously (17)
.
Expression and Purification of Recombinant EDG1-His Tagged Fusion Protein.
pET15b-EDG1 was transformed into BL21 bacteria (Novagen), and expression of His-tagged EDG1 fusion protein was induced by 1 mM isopropyl-1-thio-ß-D-galactopyranoside (Promega, Madison, WI). His-tagged EDG1 fusion proteins were isolated by running the bacterial lysate over a Ni-NTA agarose column (Qiagen, Valencia, CA) and competed off with increasing concentrations of imidazole (10, 25, 100, 250, and 500 mM) containing buffer (50 mM NaH2PO4 and 300 mM NaCl). Eluted His-tagged EDG1 was resolved by SDS-PAGE and detected using the Qiaexpress Detection System (Anti-His antibody chromogenic method) according the manufacturer’s protocol (Qiagen).
Western Analysis.
Total protein was extracted from MCF7 cells using M-PER Mammalian Protein Extraction Reagent according to the manufacturer’s protocol (Pierce, Rockford, IL). EDG1 was detected using the EDG1 (peptide 154–171) polyclonal rabbit antibody at a 1:1000 dilution (2.2 µg/ml) and an horseradish peroxidase-conjugated antirabbit IgG secondary antibody (Amersham Biosciences, Piscataway, NJ). Cytokeratin 18 or GAPDH were detected as recommended using the DC-10 monoclonal mouse anticytokeratin 18 IgG antibody (Santa Cruz Biotechnology, Santa Cruz, CA) or the 6C5 monoclonal mouse anti-GAPDH IgG antibody (Chemicon International, Inc., Temecula, CA), respectively, and an HRP-conjugated goat antimouse IgG secondary antibody (Amersham Biosciences). EDG1, cytokeratin 18, and GAPDH signals were detected using enhanced chemiluminescence Western Blotting Analysis System (Amersham Biosciences).
Northern Blot Analysis.
RNA was extracted from breast epithelial cells using TRIzol (Life Technologies, Inc. Invitrogen, Carlsbad, CA) and was subjected to Northern Analyses as described previously (15)
.
Retroviral-mediated Transfection.
Retroviruses were made by transfecting PA317 cells with the pBPSTR1 plasmid alone or pBPSTR1 containing EDG1 cDNA in the sense or antisense orientation. Breast epithelial cell lines were infected with retrovirus-containing supernatants in the presence or absence of 3 µg/ml tetracycline. When tetracycline was added, expression of the viral gene was inhibited. Changes in EDG1 protein were verified by immunofluorescence staining or immunoblot analyses.
Immunofluoresence Staining of Breast Tissue and Cells.
Breast tissue samples were obtained from the Cooperative Human Tissue Network-Western Division and University Hospitals of Cleveland, Department of Pathology. Breast tissue samples were fixed in formalin, embedded in paraffin, and sectioned at 5 µm thickness. To unmask epitopes we used heat-induced antigen retrieval technique using 10 mM citrate buffer. After blocking with PBS (140 mM sodium chloride, 2.6 mM potassium chloride, 1.5 mM potassium phosphate monobasic, and 10 mM sodium phosphate) containing 10% normal goat serum and 0.3% Triton X-100, sections were incubated with EDG1 (peptide 154–171) polyclonal rabbit antibody (1:100 dilution; 22 µg/ml) and goat, antirabbit IgG Alexa 488 fluorescence secondary antibody (1:500 dilution; Molecular Probes, Eugene, OR). As a negative control, duplicate sections were immunostained with nonspecific rabbit IgG. To detect ER, tissue was immunostained with ER 1D5 IgG monoclonal mouse antibody (1:100 dilution; Lab Vision, Fremont, CA) and goat, antimouse Alexa 594 secondary antibody (1:500 dilution; Molecular Probes). To detect Ki67, tissue was immunostained with Ki67 IgG monoclonal mouse antibody (1:100 dilution; Lab Vision) and goat, antimouse Alexa 594 secondary antibody (1:500 dilution; Molecular Probes). As a negative control, duplicate sections were immunostained with nonspecific mouse IgG.
Cells grown on coverslips were fixed in PBS containing 4% paraformaldehyde and nuclei were permeabilized using 0.3% Triton X-100. After blocking with serum, samples were incubated with EDG1, and Ki67 primary and secondary antibody as described above. Coverslips were mounted on slides using Vectashield mounting medium with 4',6-diamidino-2-phenylindole for nuclear staining (Vector Laboratories, Burlingame, CA).
Proliferation Assays.
Breast epithelial cancer cells were infected with retrovirus-containing supernatant as described above and weaned off estrogen-containing medium as described previously (18)
. Cell proliferation was determined using Cell Titer 96 Aqueous One Solution Proliferation Assay as recommended by the manufacturer (Promega).
Anchorage-independent Growth.
Four days after infection cells were detached and suspended at a concentration of 1 x 104 in medium containing 0.3% agar and then plated onto a six-well plate precoated with 0.7% agar base layer. At 24 h and 21 days after plating, colonies >50 µm were counted.
Quantification and Statistical Analysis.
Expression levels for EDG1 in Western and Northern blots were quantified as a percentage of loading control by densitometry using NIH image 1.62 software. To quantify EDG1 immunostaining levels, mean histogram readings of cell nuclei were averaged from 
15 cells per sample using Adobe PhotoShop 6.0. Five background readings were taken for each sample, averaged, and subtracted from the EDG1 reading. To analyze the tissue data, 2 x 2 contingency tables were used to obtain 
2 and Ps. Significant differences were comparisons with Ps < 0.05.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
as bait. Sequence analysis of the EDG1 cDNA clone indicated an open reading frame of 1077 bp (359 amino acids) encoding a Mr 41,000 protein with a nuclear localization sequence at amino acids 150–177. His-tagged recombinant EDG1 was affinity purified and eluted with 100 mM imidazole, and observed on SDS-PAGE as a Mr 41,000 protein, along with an additional larger Mr 70,000 protein, suggesting either post-translational modifications or dimer formation (data not shown). To determine the specificity of our EDG1 (peptide 154–171) polyclonal rabbit antibody, the 10 mM imidazole eluate (containing high levels of bacterial proteins and no His-tagged EDG1) and the 100 mM imidazole eluate (containing low levels of bacterial proteins and high concentration of His-tagged EDG1) were subjected to Western analysis. No bands were observed in the 10 mM imidazole elute, whereas the Mr 41,000 and Mr 70,000 His-tagged EDG1 fusion proteins were detected by the EDG1 polyclonal antibody (data not shown). Fluorescence studies indicated that transfected EDG1 (in CHO cells) and endogenous EDG1 (in nontumorigenic MCF10A cells) localized to the nucleus (Fig. 1A)
.
|
70,000 (20)
. EDG1 mRNA expression was also prevalent in other human endocrine tissues such as the ovary and prostate (Fig. 1B)
, which supports a role for EDG1 in endocrine signaling.
EDG1 Expression Is Increased by the Differentiating Agent HMBA and Decreased by the Mitogen Estrogen.
Consistent with findings in vascular endothelial cells (19)
, up-regulation of EDG1 mRNA expression by HMBA was also evident in MCF7 breast cancer cells (Fig. 2A)
. It was shown previously that MCF7 cells treated with HMBA have morphological characteristics of a more differentiated cell (21)
. Viable MCF7 cells exhibit decreased cell growth and anchorage-independent growth, enlargement of cytoplasm, and increased expression of the epithelial differentiation marker, epithelial membrane antigen, after treatment with HMBA (21)
. Conversely E2, which induces breast cancer cell growth, had an inhibitory effect on EDG1 mRNA expression even in the presence of cycloheximide, suggesting a direct effect by ER (Fig. 2B)
. Endogenous EDG1 from breast cancer cells ran on SDS-PAGE exclusively as the Mr 70,000 size protein. Treatment of breast cancer cells with E2 resulted in a gradual decrease in EDG1 protein levels from a 20% decrease after 1 day and a 76% decrease after 6 days of E2 treatment. The decrease in EDG1 protein coincided with, although at a slower rate, the decrease observed in the EDG1 mRNA levels from 1- to 6-day E2 treatments (Fig. 2C)
.
|
, and Ki67 Expression, and Colocalization in a Small Breast Tissue Set., and Ki67 (Figs. 3
and 4
). Analysis of our initial breast tissue set suggested that some (5 of 16) invasive breast tumors (Fig. 4A
, Patient II, row II) and (1 of 3) ductal carcinoma in situ (Fig. 3
, Patient III, row 2) express EDG1. We also observed many (11 of 16) invasive tumors (Fig. 3
, Patient I and II, row 2) and some (2 of 3) ductal carcinoma in situ (Fig. 4B
, Patient IV, row IV; yellow arrow) that did not express EDG1. This was in contrast with the adjacent normal breast samples, in which all of the samples contained some epithelial cells that expressed nuclear EDG1 (Fig. 3
, Patient I, II, III, row 4). These luminal epithelial cells give rise to the majority of breast tumors (reviewed in Ref. 22
). Also, as expected, we observed EDG1 expression in the nuclei of endothelial cells (Fig. 3
, Patient III, row 4, indicated by arrowhead).
|
|
positive (Fig. 4A
, Patient II, row II and Patient III, row III). ER-positive tumors are considered to be more differentiated and have better prognosis than ER-negative tumors (12)
. We also observed colocalization of EDG1 within normal breast epithelial cells that expressed ER (Fig. 4A
, Patient I, row I).
Finally, although EDG1-positive tumors still contained proliferating cells, the amount of cells undergoing proliferation was low (Fig. 4B
, Patient II, row II). Of note, highly proliferating tumors either did not express EDG1 or expressed low levels of EDG1 (Fig. 4B
, Patient III, row III). We observed high Ki67 expression in both of the ductal carcinoma in situ, which did not express nuclear EDG1 (Fig. 4B
, Patient IV, row IV; yellow arrow). Conversely, Ki67-positive cells were rarely observed in the 16 samples of normal breast epithelial ducts, which were highly positive for nuclear EDG1 (Fig. 4B
, Patient I, row I and Patient IV, row IV). It has been shown previously that ER and the proliferation antigen, Ki67, rarely colocalized in normal breast epithelial cells, whereas in a proportion of human breast tumors there is increased colocalization between ER and Ki67 (23)
. The data from this initial tumor study suggested that additional investigation of breast tumors was necessary because of the small sample size and under-representation of both BR grade I and grade II tumors. Therefore, we did a more in-depth analysis of a larger set of invasive tumors for each BR grade.
Analysis of 43 Invasive Breast Tumors and Adjacent Normal Breast Epithelial Cells for EDG1, ER, and Ki67 Expression.
Additional studies were conducted on a total of 43 invasive breast cancer samples and 40 adjacent normal breast samples (1 grade III tumor and 3 normal samples were unavailable for the study). Analyses were done on 15 BR grade I, 15 BR grade II, and 13 BR grade III breast tumors, and the adjacent normal breast epithelial tissue samples for 13 BR grade I, 13 BR grade II, and 14 BR grade III. Spatial expression of ER and EDG1 or Ki67 and EDG1 was compared for each sample. At least 1000 epithelial cells per tumor sample and up to 500 epithelial cells per normal sample were counted (10 fields of view were randomly chosen for each sample). Positive expression for EDG1, ER, and Ki67 were quantified as cells with complete nuclear staining (weak or strong) above background. The samples were then categorized as either negative (0%), low expressers (<10% of cells expressing antigen), moderate expressers (10–50% of cells expressing antigen), or high expressers (>50% of cells expressing antigen). The data are summarized in Fig. 5, A–E
, and Table 1
.
|
|
[53% (5% low expressers, 39% moderate expressers, and 9% high expressers) and 80% (45% low expressers, 35% moderate expressers, and 0% high expressers), respectively] and Ki67 [93% (33% low expressers, 58% moderate expressers, and 2% high expresser) and 62.5% (57.5% low expressers, 5% moderate expressers, and 0% high expressers), respectively]. As expected, tumor and normal samples were different in that the majority of ER- and Ki67-positive tumors were moderate expressers (39% and 58%, respectively), whereas the majority of adjacent normal samples were low expressers (45% and 57.5%, respectively; data not shown). This resulted in the overall percentage of cells expressing ER and Ki67 to be significantly higher in tumor cells as compared with the adjacent normal ducts (ER: 2 = 1023, P < 0.001; Ki67: 2 = 2993, P < 0.001; Table 1
).
Thirty-nine of 43 samples had tumors with adjacent normal samples. We observed a decrease in EDG1 expression in 37 of these 39 tumor samples as compared with their adjacent normal samples. The 2 samples in which a decrease was not observed were both BR grade I tumor samples (Fig. 5A)
. Nine of 43 tumors (21%) were negative for EDG1 expression, and 20 tumors (47%) were either EDG1 negative or EDG1 low expressers. The majority of tumors [20 of 43 (47%)] were EDG1 moderate expressers, and very low percentage [3 of 43 (6%)] were high EDG1 expressers (Fig. 5B)
. As was observed in the initial set of 16 samples, none of the 40 normal breast samples were negative or low expressers for EDG1, and the majority [32 of 40 (80%)] were high EDG1 expressers (>50% expression) with 20% (8 of 40) moderate EDG1 expressers (Fig. 5C)
.
When we compared the number of breast tumors that were 100% EDG1 negative (9 of 43) to the normal breast samples that were 100% EDG1 negative (0 of 40), the two sets of samples were significantly different (2 = 8.2; P = 0.004). These data supports the cell by cell counts where the majority of adjacent normal breast epithelial cells expressed EDG1 (63%), and the proportion was significantly lower in tumor cells (16%; EDG1: 2 = 12,717, P < 0.001; Table 1
).
No correlation between BR tumor grade and the level of EDG1 expression within the tumors was observed (Fig. 5B
; Table 1
). Sixty-seven percent of normal samples adjacent to BR grade I and 77% adjacent to BR grade II tumors contained >70% of their cells expressing nuclear EDG1. In comparison, only 14% of those normal samples adjacent to BR grade III tumors contained >70% of their cells expressing nuclear EDG1 (2 = 5.5, P = 0.004; Fig. 5C
). This correlated with the cell by cell counts in which we observed that the difference for EDG1-positive expression in adjacent normal breast epithelial ducts, between grade I and II (66% and 78%, respectively; 2 = 183; P < 0.001), was not as strong as the difference between grade I and II versus grade III (71% and 51%, respectively; 2 = 386; P < 0.001; Table 1
). This may also partially explain why the difference in the percentage of proliferating adjacent normal breast cells from grade I (0.5%) and II (0.8%) tumors was not as significant (2 = 4.0; P > 0.025) when compared with the difference in the percentage of proliferating adjacent normal breast cells from grade I and II (0.7%) versus grade III (2.0%; 2 = 98; P < 0.001). These findings are consistent with our observation that a higher percentage of grade III adjacent normal breast epithelial cells were proliferating.
As we stated previously, our initial small sample of 16 breast tumors suggested that EDG1-positive breast tumors were correlated with ER-positive breast tumors. Therefore, we analyzed the breast tumors from our larger sample to determine the percent of tumors that contained cells expressing both ER and EDG1 within the same cells (Fig. 5D)
, and the percentage of cells from all of the tumors expressing both EDG1 and ER within the same cells (colocalization; Table 1
). Forty-six percent (20 of 43) of tumors contained cells expressing both ER and EDG1 within the same cells at various percentages. The majority of these tumors [65% (13 of 20)] had >10% colocalization (Fig. 5D)
. Because some tumors were either negative for both EDG1 and ER, only EDG1 positive, or only ER positive, it was necessary to consider these samples as a separate population to optimally determine whether EDG1-expressing and ER-expressing tumors were correlated. Fifty-nine percent of the tumors that were EDG1-positive expressed ER (20 of 34), and 86% (20 of 23) of the samples that were ER-positive expressed EDG1 (data not shown).
As stated previously EDG1 expression was not completely disassociated from breast tumors undergoing proliferation in the initial 16 tumor samples. Therefore, we analyzed the larger tumor set for Ki67 and EDG1 expression within the same tumor cells as we did with ER and EDG1. We observed 74% of the tumors to contain cells expressing both EDG1 and Ki67 within the same cells at various percentages, but in contrast to tumors expressing both ER and EDG1, only 5% of EDG1- and Ki67-positive tumors had >10% of cells expressing both EDG1 and Ki67 within the same cells (Fig. 5D)
. This was observed although the majority of tumors were EDG1 (47%) and Ki67 (58%) moderate expressors (10–50% of the cells).
Although the difference between percentage of total tumor cells expressing ER (18%) versus Ki67 (16%) was significant (2 = 32; P < 0.001; Table 1
), it was not as significant as the difference between tumor cells expressing both ER and EDG1 (7%) as compared with tumor cells expressing both Ki67 and EDG1 (1.6%; 2 = 1396; P < 0.001; Table 1
). This suggests that there is a stronger disassociation between proliferating breast tumor cells and EDG1-positive tumor cells as compared with ER-positive and EDG1-positive breast tumor cells.
We also determined whether normal breast epithelial cells adjacent to tumors were also more likely to express ER and EDG1 within the same cells as compared with EDG1 and Ki67. Eighty percent (32 of 40) of adjacent normal samples contained cells expressing both ER and EDG1 within the same cells. In contrast, 40% of normal samples contained cells expressing both EDG1 and Ki67 within the same cells (Fig. 5E)
. This observation was supported by cell by cell counts in which a significantly higher percentage (6%) of total normal breast epithelial cells counted expressed both ER and EDG1 within the same cells as compared with 0.3% of normal breast epithelial cells expressing EDG1 and Ki67 (2 = 1144; P < 0.001; Table 1
). We also observed that 9% of EDG1-positive normal cells were ER positive, and 83% of ER-positive normal cells were EDG1 positive. In contrast, only 0.4% of EDG1-positive normal cells were Ki67 positive, and only 22% of Ki67-positive normal cells were EDG1 positive (data not shown). This study additionally supports decreased expression of EDG1 in breast tumors and a stronger disassociation with proliferating breast epithelial cells (normal and malignant) as compared with ER-positive breast epithelial cells (normal or malignant).
EDG1 Inhibits Breast Cell Proliferation.
Although the majority of proliferating and EDG1-positive breast epithelial cells were disassociated, it is not conclusive from these studies that EDG1 can regulate cell proliferation. Therefore, a self-contained tetracycline-regulated retroviral vector system (17)
was used to modulate EDG1 expression in breast epithelial cell lines. Infection of MCF7 cells with EDG1 retrovirus resulted in increased nuclear expression of EDG1 and decreased expression of Ki67 as compared with control cells (Fig. 6A)
. In contrast, inhibiting EDG1 expression in MCF7 cells resulted in decreased nuclear expression of EDG1 and an increased number of cells expressing Ki67 as compared with control cells (Fig. 6A)
. Proliferation assays indicated that a minor increase (50%) in EDG1 expression resulted in an inhibition of growth of MCF7 cells, whereas a decrease (90%) in EDG1 expression enhanced growth rate (Fig. 6B)
. It appears that tetracycline did not fully repress EDG1AS retroviral expression (Fig. 6B)
. This could explain why an increase in proliferation was observed in EDG1AS cells with tetracycline, but to a lesser degree than EDG1AS cells without tetracycline. These findings support our results that breast tissue cells positive for nuclear EDG1 are less likely to express Ki67 as compared with breast tissue cells negative for nuclear EDG1. Studies presented in Fig. 6, A and B
, also indicate that EDG1 inhibited E2-induced cell proliferation.
|
.
To determine whether the effects of EDG1 on cell proliferation were cell line-specific, EDG1 and EDG1AS retroviruses were infected into a normal breast epithelial cell line (MCF10A) or another breast cancer cell line (MDA-MB-231). Decreased nuclear expression of EDG1 in MCF10A after infection with EDG1AS retroviruses was associated with 4.5-fold increase in proliferation, whereas a slight increase in EDG1 nuclear expression inhibited proliferation markedly (Fig. 7A)
. No significant effects on proliferation of MDA-MD-231 cells were evident after infection with EDG1AS retroviruses. However, after infection with EDG1 retroviruses we saw a 64% decrease in proliferation (Fig. 7B)
.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Tumors expressing EDG1 appeared to be correlated with tumors expressing ER. Although this may appear to contradict what one would expect if estrogens directly down-regulate EDG1, there is a logical explanation as to why cells in which ER is detected could still express EDG1. Numerous studies have demonstrated autoregulation of ER, wherein estrogen induces a rapid decline in both ER mRNA and protein. Although the mechanism is not well defined, studies support estrogen induction of ER-mediated transcriptional down-regulation of ER gene transcription, and/or degradation of ER protein via a proteosome-mediated proteolytic mechanism (11
, 24, 25, 26)
. Regardless of the mechanism, ER may not be detected by immunostaining in cells wherein EDG1 has been down-regulated by estrogens. The correlation between EDG1 and ER expression in the same tumors provides some evidence that EDG1 expression can be correlated with better prognosis, suggesting correlative studies with other prognostic factors are necessary (Ref. 12
; reviewed in Ref. 13
).
A more thorough understanding of the role of estrogens in the initiation or promotion of breast cancer is relatively lacking. Some have proposed ER-independent pathways involving cytochrome P450-mediated metabolism of estrogens, which elicit direct genotoxic effects by increasing mutation rates (27)
. Whereas estrogen induces expression of growth factors, proto-oncogenes, proteases, and basement membrane receptors (reviewed in Refs. 3
, 28
), the exact mechanism(s) by which estrogens regulates mammary epithelial cell proliferation and the relative role of gene induction on the proliferative effects of estrogens is still not well-defined. More recent studies using Serial Analysis of Gene Expression technology have provided additional insights into estrogen gene regulation in normal ER breast cells and breast cancer cells (29
, 30)
. Gene transcripts that have a putative role in cell proliferation were induced or down-regulated by estrogens. Two genes in particular, one encoding a secreted protein lipocalin 2 and another protein of unknown function, EIT-6, appear to be induced directly by estrogen, and enhance proliferation in normal breast cells and breast cancer cells, respectively (29
, 30)
.
Our results support the down-regulation of EDG1 as being involved in estrogen-induced proliferation of breast cells. In particular, EDG1 expression is inhibited by estradiol and EDG1 overexpression (which bypasses the inhibitory effects of estradiol on EDG1 expression) attenuated estradiol-induced cell proliferation. We did not see enhancement of estradiol-induced cell proliferation upon down-regulation of EDG1 expression. It is possible that in this case maximal stimulation of cell proliferation by estrogens has already been achieved.
The fact that only 21% of the breast tumor samples in the larger study showed complete loss of EDG1, and the remaining tumors only showed a severe reduction in EDG1, suggests that the loss of EDG1 in the majority of tumors was not clonal in origin. It is possible that for some tumors, loss of EDG1 is a genetic or epigenetic event, warranting the investigation of possible genetic or epigenetic alterations of EDG1 in breast cancer patients. However, for the majority of the tumors, the decrease in EDG1 expression can be attributed to other factors.
As stated previously, for some tumors, the likely cause is estrogen down-regulation. Although this seems likely for some tumors, six tumors showing severe loss of EDG1 expression (<10% of the cells expressing EDG1) were ER negative. Recent studies have shown that some breast tumors are not clonal in origin and are instead polyclonal (31
, 32)
. Thus, it is possible that tumors that were 100% negative for ER expression and only showed a severe reduction in EDG1 expression, instead of 100% loss, may have been tumors of polyclonal origin. It has been proposed that clones with certain somatic mutations may have a growth advantage over other clones (31)
. It could be speculated that clones which originally had or subsequently acquired alterations in normal EDG1 expression might have a growth advantage over other originating clones because of the ability of EDG1’s to inhibit cell growth. This is reasonable to expect because of the loss of EDG1 expression in the majority of the cells within most tumors. Although we only screened a small set of samples, we found that 2 of 3 preinvasive ductal carcinoma in situ observed were 100% negative for EDG1. This suggests that in some tumors the loss of EDG1 occurs before tumor invasion, but additional investigation that is beyond the scope of this paper is necessary.
Database searches indicated that EDG1 is localized to chromosome arm 17q.4 LOH in 17p (containing the p53 locus) and in 17q (including loss of the whole 17q arm) is common in breast cancers (33 , 34) . Whereas the BRCA1 locus is a common site of LOH in breast cancer there are other sites of LOH outside the BRCA1 locus (35) . Although we have not yet established whether the EDG1 locus undergoes genetic alterations in breast cancer patients, studies do indicate that EDG1 is surrounded by markers that have been reported to frequently undergo LOH in breast tumors (36, 37, 38) .
Regardless of whether or not EDG1 undergoes LOH in breast cancer patients, the decreased expression of EDG1 in the majority of breast tumor cells and the ability of EDG1 to inhibit breast cell growth, strongly support a role for EDG1 in breast cancer development. The normal function of EDG1 in normal breast epithelial cells may be to inhibit cell growth and play a role in breast tissue differentiation, as suggested by the increase in EDG1 expression by the differentiating agent HMBA. The decrease in expression of EDG1 in breast tumors could be attributed to down-regulation by estrogens, LOH, hypermethylation of the EDG1 promoter, a combination of these factors, or other unknown factors. Studies are under way to determine the mechanism(s) by which EDG1 expression is lost in the majority of breast cancer cells and how EDG1 inhibits cell growth. Understanding these mechanisms should provide additional insights into how estrogens play a role in the initiation of breast tumorigenesis.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by NIH Grants CA80959 and CA92440, and Gustavus and Louis Pfeiffer Research Foundation (to M. M. M.). 
2 To whom requests for reprints should be addressed, at Case Western Reserve University School of Medicine, Department of Pharmacology, H.G. Wood Building W307, 2109 Adelbert Road, Cleveland, OH 44106. Phone: (216) 368-3378; Fax: (216) 368-3395; E-mail: mxm126{at}po.cwru.edu
3 The abbreviations used are: ER, estrogen receptor; E2, 17ß-estradiol; EDG1, estrogen down-regulated gene 1; BR, Bloom-Richardson; LOH, loss of heterozygosity; CHO, Chinese hamster ovary; GFP, green fluorescent protein; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HMBA, hexamethylene-bis-acetamide; EDG1AS, estrogen down-regulated gene 1 antisense.
4 Internet address: http://www.ncbi.nlm.nih.gov/cgi-bin/entrez/maps.cgi.
Received 10/29/02. Revised 5/30/03. Accepted 6/ 5/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
by estrogen receptor : molecular mechanisms and relevance in estrogen-mediated breast cell growth. Oncogene, 21: 5233-5244, 2002.[Medline]
and estrogen receptor-ß. J. Biol. Chem., 273: 25443-25449, 1998.This article has been cited by other articles:
|
|
 |
|
  J. Kohoutek, Q. Li, D. Blazek, Z. Luo, H. Jiang, and B. M. Peterlin Cyclin T2 Is Essential for Mouse Embryogenesis Mol. Cell. Biol., June 15, 2009; 29(12): 3280 - 3285. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Shimizu, N. Yoshikawa, T. Wada, H. Handa, M. Sano, K. Fukuda, M. Suematsu, T. Sawai, C. Morimoto, and H. Tanaka Tissue- and Context-Dependent Modulation of Hormonal Sensitivity of Glucocorticoid-Responsive Genes by Hexamethylene Bisacetamide-Inducible Protein 1 Mol. Endocrinol., December 1, 2008; 22(12): 2609 - 2623. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Ogba, L. J. Chaplin, Y. Q. Doughman, K. Fujinaga, and M. M. Montano HEXIM1 Regulates 17{beta}-Estradiol/Estrogen Receptor-{alpha}-Mediated Expression of Cyclin D1 in Mammary Cells via Modulation of P-TEFb Cancer Res., September 1, 2008; 68(17): 7015 - 7024. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Asakura Vascular Endothelial Growth Factor Gene Regulation by HEXIM1 in Heart Circ. Res., February 29, 2008; 102(4): 398 - 400. [Full Text] [PDF]
|
|
|
|
|
|
M. M. Montano, Y. Q. Doughman, H. Deng, L. Chaplin, J. Yang, N. Wang, Q. Zhou, N. L. Ward, and M. Watanabe Mutation of the HEXIM1 Gene Results in Defects During Heart and Vascular Development Partly Through Downregulation of Vascular Endothelial Growth Factor Circ. Res., February 29, 2008; 102(4): 415 - 422. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. C. Sedore, S. A. Byers, S. Biglione, J. P. Price, W. J. Maury, and D. H. Price Manipulation of P-TEFb control machinery by HIV: recruitment of P-TEFb from the large form by Tat and binding of HEXIM1 to TAR Nucleic Acids Res., July 26, 2007; 35(13): 4347 - 4358. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Barboric, J. H. N. Yik, N. Czudnochowski, Z. Yang, R. Chen, X. Contreras, M. Geyer, B. Matija Peterlin, and Q. Zhou Tat competes with HEXIM1 to increase the active pool of P-TEFb for HIV-1 transcription Nucleic Acids Res., March 19, 2007; 35(6): 2003 - 2012. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. He, A. C. Pezda, and Q. Zhou Modulation of a P-TEFb Functional Equilibrium for the Global Control of Cell Growth and Differentiation. Mol. Cell. Biol., October 1, 2006; 26(19): 7068 - 7076. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Q. Zhou and J. H. N. Yik The Yin and Yang of P-TEFb Regulation: Implications for Human Immunodeficiency Virus Gene Expression and Global Control of Cell Growth and Differentiation Microbiol. Mol. Biol. Rev., September 1, 2006; 70(3): 646 - 659. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Egloff, E. Van Herreweghe, and T. Kiss Regulation of Polymerase II Transcription by 7SK snRNA: Two Distinct RNA Elements Direct P-TEFb and HEXIM1 Binding Mol. Cell. Biol., January 15, 2006; 26(2): 630 - 642. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W.-J. HE, R. CHEN, Z. YANG, and Q. ZHOU Regulation of Two Key Nuclear Enzymatic Activities by the 7SK Small Nuclear RNA Cold Spring Harb Symp Quant Biol, January 1, 2006; 71(0): 301 - 311. [Abstract] [PDF]
|
|
|
|
|
|
D. Blazek, M. Barboric, J. Kohoutek, I. Oven, and B. M. Peterlin Oligomerization of HEXIM1 via 7SK snRNA and coiled-coil region directs the inhibition of P-TEFb Nucleic Acids Res., December 23, 2005; 33(22): 7000 - 7010. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Dulac, A. A. Michels, A. Fraldi, F. Bonnet, V. T. Nguyen, G. Napolitano, L. Lania, and O. Bensaude Transcription-dependent Association of Multiple Positive Transcription Elongation Factor Units to a HEXIM Multimer J. Biol. Chem., August 26, 2005; 280(34): 30619 - 30629. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Shimizu, R. Ouchida, N. Yoshikawa, T. Hisada, H. Watanabe, K. Okamoto, M. Kusuhara, H. Handa, C. Morimoto, and H. Tanaka HEXIM1 forms a transcriptionally abortive complex with glucocorticoid receptor without involving 7SK RNA and positive transcription elongation factor b PNAS, June 14, 2005; 102(24): 8555 - 8560. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A. Byers, J. P. Price, J. J. Cooper, Q. Li, and D. H. Price HEXIM2, a HEXIM1-related Protein, Regulates Positive Transcription Elongation Factor b through Association with 7SK J. Biol. Chem., April 22, 2005; 280(16): 16360 - 16367. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. N. Yik, R. Chen, A. C. Pezda, and Q. Zhou Compensatory Contributions of HEXIM1 and HEXIM2 in Maintaining the Balance of Active and Inactive Positive Transcription Elongation Factor b Complexes for Control of Transcription J. Biol. Chem., April 22, 2005; 280(16): 16368 - 16376. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
