| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Department of Oncology and Lombardi Comprehensive Cancer Center,
Department of Physiology and Biophysics, Georgetown University School of Medicine, Washington, DC, USA
1Correspondence: Rm. W405A Research Bldg., Department of Oncology, Georgetown University School of Medicine, 3970 Reservoir Rd., NW, Washington, DC 20057, USA. E-mail: clarker{at}georgetown.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: apoptosis • aromatase • cell cycle • Faslodex • SERM • SERD • unfolded protein response • Tamoxifen • proliferation • XBPI
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(ER). Antiestrogens, including selective estrogen receptor modulators (SERMs) and selective estrogen receptor down-regulators (SERDs), prevent estrogen from binding to the estrogen receptor (ER) and regulating cell growth (1)
. The triphenylethylene Tamoxifen (TAM), which is the most frequently used antiestrogen, can reduce mortality and breast cancer recurrence in women with breast tumors that express detectable levels of the estrogen receptor- (ER+, gene symbol ESR1; refs. 2
, 3
). Faslodex (FAS) is effective as a second-line agent in women who have progressed on TAM (4)
and is as effective as some aromatase inhibitors in postmenopausal women (5)
. Unfortunately, up to 50% of ER+ breast cancers do not respond to endocrine therapies, displaying de novo or intrinsic resistance, while breast tumors that initially respond to antiestrogens frequently acquire resistance to these treatments (6)
. Factors that contribute to de novo and acquired antiestrogen resistance and to cell survival in ER+ breast cancer are not completely understood. However, in sensitive cells, antiestrogen and estrogen withdrawal induce both a cell cycle arrest in G0/G1 and an increase the rate of apoptosis (1)
.
Previously, we used a series of estrogen [17β-estradiol (E2)] -independent and antiestrogen-resistant models to identify key components of a broader gene signaling network that contributes to estrogen-independence and antiestrogen resistance (TAM and FAS crossresistance; ref. 6
). Using immunohistochemistry, we have also recently shown that several of these proteins are expressed in ER+ human breast cancers and in patterns of coexpression that are consistent with our proposed signaling network (7)
. Among key nodes in this network, we identified interferon regulatory factor-1 (IRF-1), NF-
B, and X-box binding protein-1 (XBP1; ref. 8
); subsequent studies have shown the functional relevance of IRF-1 (9
, 10)
and NF-B (8
, 11
, 12)
. We have recently shown that the XBP1 protein is expressed in almost 80% of ER+ breast tumors (7)
, an observation consistent with several molecular profiling studies in breast cancer (13
14
15
16
17)
and a study indicating that XBP1 expression is detected in neoplastic but not non-neoplastic breast tissue (18)
. Although we have also observed that XBP1 protein and mRNA are up-regulated in antiestrogen resistant cells and that this is accompanied by an increase in cyclic AMP response element (CRE) activity (8)
, the functional relevance of these associations has remained unknown.
XBP1 is a transcription factor that belongs to the basic region/leucine zipper (bZIP) family of proteins (19
, 20)
. Regulation of transcription by XBP1 is a consequence of its binding to and activating specific CREs that have a conserved ACGT core sequence GATGACGTG(T/G)NNN(A/T)T (19
, 21)
. Two forms of XBP1 have been identified: a spliced form, XBP1(S), with a molecular mass of 
54 kDa and an unspliced form, XBP1(U), with a molecular mass of 33 kDa (22)
. Splicing of the XBP1 RNA results in the removal of a 26 base intron that creates a translational frameshift. The spliced protein product is a potent transcriptional regulator of many target genes in the unfolded protein response (UPR), an adaptive endoplasmic reticulum signaling pathway that allows cells to survive the accumulation of unfolded proteins in the endoplasmic reticulum lumen (23)
. However, although UPR is initially a compensatory mechanism by which cells can reset normal endoplasmic reticulum function, prolonged UPR will induce cell death. Induced by cellular stressors such as hypoxia, UPR can be initiated by each of three molecular sensors, i.e., IRE1, ATF6, and PERK (24)
. Splicing of XBP1 by IRE1 is an obligate component in both IRE1- and ATF6-induced UPR (25
26
27)
.
In the mouse, XBP1 is essential for fetal survival, neurological development, bone growth, immune system activation, and liver development (28
, 29)
. XBP1 is widely expressed in adult animals, whereas it is detected mainly in exocrine glands, osteoblasts, chondroblasts, liver, and bone precursors in the embryonic mouse (28)
. Expression of the tissue inhibitor of metalloproteinases, osteopontin, and osteocalcin (28)
is regulated by XBP1, which also is implicated in affecting plasma cell differentiation (30)
.
In this study, we show that XBP1 is a key factor in affecting antiestrogen responsiveness and estrogen dependence in breast cancer cells. MCF-7 and T47D cells stably overexpressing XBP1(S) no longer require estrogen for cell growth and are resistant to both TAM and FAS. Cells overexpressing XBP1(S) express functional ER that can form ER:XBP1 heterodimers. Finally, we show that XBP1(S) overexpression alters the expression of several apoptotic and cell cycle genes and promotes cell survival through affecting activity of the intrinsic apoptosis pathway.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Assays and analyses
XBP1 transfection, clonal selection, and cell lysis
MCF-7 and T47D cells were seeded in IMEM at a density of 7 x 105 cells/25 cm2 plastic tissue culture flasks. Twenty-four hours later cells were stably transfected with 2 µg of pCDNA3.1/XBP1 and the neomycin-resistance cassette, using Fugene 6 as recommended by the manufacturer (Roche Diagnostics, Indianapolis, IN, USA). Twenty-four hours post-transfection, cells were trypsinized and transferred to plastic cell culture dishes and allowed to grow in IMEM containing 800 µg/ml of G418. Resistant colonies were selected and expanded from 24-well plates to T-75 cm2 plastic tissue culture flasks.
For each colony selected, cells were washed with ice-cold phosphate-buffered saline solution (PBS) and lysed on ice in modified RIPA buffer (150 mM NaCl, 50 mM Tris Base, 1% Igepal, and 0.5% deoxycholate, pH 7.5) containing one complete protease inhibitor tablet (Roche Diagnostics) with shearing applied using a 1 ml syringe and 21-gauge needle. Insoluble cell debris was removed by centrifugation at 4°C. The bicinchoninic acid (BCA) assay (Pierce Biotechnology, Rockford, IL, USA) was used to determine protein concentration in the cell lysates. Western blot analysis was used to screen clones for XBP1 protein expression. MCF-7 cells overexpressing XBP1 were designated MCF7/XBP1; T47D cells overexpressing XBP1 were designated T47D/XBP1. The respective control populations, transfected with the empty expression vector, were designated MCF7/c and T47D/c.
Western blot analysis
Total protein (20 µg) was isolated from subconfluent cell populations, size fractionated by electrophoresis using NuPage 10% Bis-Tris gels, and blotted onto nitrocellulose membranes. Nitrocellulose membranes were washed briefly in a solution of TBS/0.1% Tween-20 (TBST), pH 7.4, and blocked in a milk solution (cow’s milk diluted to 5% in TBST) for 15 min with constant agitation. After blocking, the nitrocellulose membrane was washed with TBST (3x for 15 min) and incubated with the following antibodies: rabbit polyclonal XBP1 primary antibody (1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA), ER mouse monoclonal antibody 6F11 (1:200; Vector Laboratories, Burlingame, CA, USA), and mouse monoclonal BCL2 primary antibody (1:200; Stressgen Biotechnologies, Vancouver, BC, Canada) diluted in TBST overnight. Extensive washes were carried out as before, and the membranes were subsequently incubated for 1 h in rabbit anti-mouse horseradish peroxidase-conjugated IgG (Amersham Biosciences, Piscataway, NY, USA) at a 1:5,000 dilution at room temperature. Nitrocellulose membranes were reprobed with an anti-β-actin antibody (1:5,000; Sigma) to assess equivalence of protein loading. After the final washes of the membranes in TBST, antigen-antibody complexes were visualized using the ECL detection system (Amersham Biosciences) and quantified by densitometry. Data (mean±SE) are presented as the ratio of target:β-actin signals.
Transient transfection and promoter-reporter assays
Cells (7x104 cells/well) were plated in 12-well plates and allowed to proliferate for 24 h before cotransfection with 0.4 µg of CRE-luciferase reporter plasmid (Promega, Madison, WI, USA) or estrogen response element plasmid (ERE) -tk-luciferase reporter plasmid (Promega). Transfection efficiency was determined by cotransfection with 0.1 µg of plasmid containing the Renilla luciferase gene (Promega). Transfection was carried out using the Fugene 6 transfection reagent (Roche Diagnostics). Three hours post-transfection, cell monolayers were washed and refed with the appropriate medium for 24 h; IMEM with 5% fetal bovine serum (to measure CRE-luciferase activity) or CCS-IMEM ± E2 (to measure ERE-tk luciferase activity). Cells were lysed and activation of the CRE or ERE-luciferase constructs was measured using the Dual Luciferase Assay Kit (Promega) according to the manufacturer’s protocol. Luminescence was measured using a Lumat LB 9501 luminator (EG&G Berthold, Bundoora, Australia). Renilla luminescence was used to normalize luciferase values. CRE-luciferase activity in XBP1-transfected cells was subsequently normalized to empty vector control. ERE-luciferase activity was normalized to empty vector control growing in media without E2 supplementation.
Cell proliferation assays
To study estrogen independence, MCF7/XBP1 Clone A, T47D/XBP1 pool, and their respective MCF7/c and T47D/c control cells were estrogen deprived by washing the monolayers with CCS-IMEM and maintaining the cultures for 72 h in T-75 cm2 plastic tissue culture flasks. Cells were then trypsinized and seeded at a density of 3.5 x 104 cells/well into 12-well plastic tissue culture plates in CCS-IMEM ± E2 (day 0); on days 1, 3, and 7, the cells were trypsinized, resuspended in PBS, and counted using a Beckman coulter counter (Beckman Coulter, Fullerton, CA, USA). To study estrogen and antiestrogen response, cells were seeded at a density of 4 x 104 cells/well in 12-well plastic tissue culture plates in IMEM ± E2,TAM, FAS, or 0.1% (v/v) ethanol (vehicle control) as indicated. On day 4 (T47D) or day 6 (MCF7) of treatment, the cells were trypsinized, resuspended in PBS, and counted as above. To study the response to BCL2 inhibition, cells were seeded at a density of 104 cells/well in 24-well plastic tissue culture plates in IMEM ± 100 nM YC-137 or vehicle control (ethanol) On day 3, the cells were trypsinized, resuspended in PBS, and counted as described above.
Cell cycle and apoptosis analyses
Cells stably overexpressing XBP1 or the empty vector control plasmid were seeded in IMEM into 100 mm plastic tissue culture dishes. Twenty-four hours later, cell monolayers were washed with CCS-IMEM, the growth medium was changed as appropriate, and the cells were maintained for a further 72 h. For cell cycle analysis, cells were harvested, fixed, and analyzed for alterations in cell cycle by fluorescence activated cell sorting (FACS) according to the method of Vindelov et al. (31)
. All FACS analyses were performed by the Lombardi Comprehensive Cancer Center Flow Cytometry Shared Resource. To measure apoptosis, cells were harvested and stained as described in the Vybrant Apoptosis Assay Kit #3 (Molecular Probes/Invitrogen, Carlsbad, CA, USA). Programmed cell death or apoptosis was measured by counting cells stained green by annexin V-FITC (Lombardi Comprehensive Cancer Center, Flow Cytometry Shared Resource).
Gene expression microarray studies
Total RNA was isolated using the Trizol method from six independent cultures (cell populations grown on different days from different stocks); three from MCF7/XBP1 Clone A cultures and three from the vector control cultures. MIAME 1.1 compliant data were collected as proposed by the Microarray Gene Expression Data (MGED) Society (http://mged.org). RNA concentrations were determined by comparing the optical density ratios (260:280 nm) obtained spectrophotometrically using a Beckman DU640 Spectrophotometer (Beckman, Fullerton, CA, USA). Data on RNA quality were obtained using an Agilent 2100 Bioanalyzer and RNA 6000 LabChip kits (Agilent Technologies, New Castle, DE, USA). RNA quality was assessed by visual inspection of the electropherograms from the Bioanalyzer data and by the calculated RNA integrity numbers (RIN; ref. 32
) and Degradometer values (33)
. Only high quality total RNA was labeled and hybridized to U133A Affymetrix GeneChips using manufacturer recommended procedures (Affymetrix, Santa Clara, CA, USA). Standard "spiked-in" controls also were included in each hybridization. Microarray data quality was assessed using various tools including those recommended by Affymetrix; all array data presented here passed the quality control measures. The raw gene expression data were then preprocessed using the Robust Multiple-Array Average (RMA) method (34)
as implemented in the Bioconductor Project (http://bioconductor.org). Expression data are available through the Gene Expression Omnibus (GEO) database, accession GSE8562.
For this study, the primary goal was to identify an initial series of genes associated with the processes of cell cycle regulation and apoptosis as affected by XBP1. Thus, we applied a simplistic approach to identify a broad range of possible candidate genes for exploration. We first extracted a reduced dimensional data set enriched in the most informative signals by excluding those genes least likely to be differentially expressed between the control and MCF7/XBP1 cell populations. This was accomplished by applying a series of predefined filters such that the reduced dimensional data set comprised 398 genes that exhibit 
1.45-fold change, P < 0.05 (pairwise univariate comparison) and genes with intensity log2 (10)
in both control and experimental groups (Bin 1 in Supplemental Fig. 1). To minimize the proportion of false negative data, the univariate filter was applied without correction for multiple comparisons. We then defined the reduced dimensional data set by estimating both the false discovery rate (FDR; refs. 35
, 36
) and the miss-rate (MR) (37)
. We do not use the FDR and MR to select individual genes, rather these are used to provide estimates of the extent to which the univariate filter included false positive (FDR) and excluded false negative data (MR) from the reduced dimensional data set. This approach maps the potential errors across the entire data set, enabling gene selection to proceed in a manner where the trade-offs between Type I and Type II errors are clearly identified. FDR analysis (38)
estimates that this data set of 398 genes contains 10 genes that may not be truly statistically different. Data visualization before and after dimensionality reduction is facilitated by multidimensional scaling using principal component analysis (PCA) and our discriminant component analysis (DCA) method (39)
. This ensures that the global structure of the data has not been compromised during the dimensionality reduction procedures (not shown).
RNA isolation and real-time quantitative PCR (qPCR)
Total RNA was isolated, cleaned, quantified, and analyzed by the Agilent Bioanalyzer 2100 as detailed above. One microgram of RNA was then treated with DNase I (Invitrogen) before reverse transcription using SuperScript II Reverse Transcriptase (Invitrogen) and oligo d(T)16 primers (Applied Biosystems, Foster City, CA, USA). qPCR reactions for each cDNA sample and a standard curve were established using TaqMan Universal PCR Master Mix and the following TaqMan Gene Expression Assay primers (Applied Biosystems): APBB2 = Hs00300268_m1; BECN1 = Hs00186838_m1; CRK = Hs00180418_m1; CSPG6 = Hs00271322_m1; IL24 = Hs01114274_m1; MYC = Hs00153408_m1; PHLDA2 = Hs00169368_m1; S100A6 = Hs00170953_m1; TFDP1 = Hs00830594_s1; TOP1 = Hs00243257_m1; XRCC6 = Hs00750856_s1; and the housekeeping gene RPLP0 = Hs99999902_m1. Ten microliter reactions were run in triplicate in 384-well plates on an ABI Prism 7900HT Sequence Detection System, using the absolute quantification protocol specified by the manufacturer. Expression data for each gene were estimated relative to the housekeeping control, and these data were used to calculate the ratio of expression in MCF7/XBP1 Clone A relative to that in the parental MCF7/c cell line; data are mean ± SE.
Coimmunoprecipitation
Cell lysate (400 µg) was incubated with 2 µl of XBP1 antibody at 4°C overnight with rotation. Thirty microliters of protein A-Sepharose beads (Amersham Bioscience) were then added, and the tubes were returned for additional rotation for 1 h at 4°C to remove immune complexes. Samples were centrifuged at 4°C, the supernatant was removed, and beads were washed twice with Tris-saline (150 mM NaCl, 50 mM Tris Base, pH 7.5). After final aspiration, immune complexes were resuspended in 2x Laemmli sample buffer and boiled for 5 min. The immune complexes and 20 µg of corresponding cell lysates were then resolved by electrophoresis using NuPage 10% Bis-Tris polyacrylamide gels.
Mitochondrial membrane permeability assay
Cells were seeded at a density of 104 cells/well in black, 96-well dishes one day before the addition of TAM, FAS, or 0.1% (v/v) ethanol vehicle. The cells were then cultured for 18–20 h before 100 µl of MitoSensor (Clontech, Mountain View, CA, USA) reagent per well were added (final concentration 5 µg/ml) for 15 min at 37°C. Cells were washed with PBS, and green fluorescence (485 nm excitation/535 nm emission) was measured on a Wallac Viktor2 1420 Multilabel Counter (Perkin-Elmer, Boston, MA, USA).
Statistical evaluation of data
One-way ANOVA was used to determine overall significant differences following treatment in the cell proliferation, cell cycle, apoptosis, and mitochondrial membrane permeability assay. Paired t test was used to determine differences between empty vector controls and stable transfectants for XBP1 overexpression, ER, BCL2 expression, and luciferase promoter-reporter assays. Coexpression in the microarray studies was assessed by pair-wise application of the Pearson correlation coefficient. Calculation of IC50s was performed using XLfit4 software from IDBS (Alameda, CA, USA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
54 kDa] was confirmed by Western blot analysis in both MCF7/XBP1 and T47D/XBP1 cells (Fig. 1
). Thus, breast cancer cells endogenously generate the XBP1(S) protein product implicated in regulating the UPR. Furthermore, the resulting phenotypes are primarily the result of increased XBP1(S) expression, since XBP1 (U) expression remains at levels similar to controls.
|
XBP1 binds to and activates only those cAMP response elements (CREs) with the conserved ACGT core sequence: GATGACGTG(T/G)NNN(A/T)T (19)
. Thus, we measured CRE activity in the XBP1 stably transfected cells using an appropriate CRE-luciferase promoter-reporter assay; transfection efficiency was measured by cotransfecting cells with a constitutively expressed Renilla luciferase gene. As expected in MCF7/XBP1 cells, CRE activity is increased 3-fold (P<0.05) in Clone A and 1.5-fold in Clone 2 (Fig. 2
A; P<0.05). Similar increases are observed in the T47D/XBP1 cells, where a 2-fold increase in CRE activity is observed in the pooled population (Fig. 2B
; P<0.05). MCF7/XBP1 Clone A and the T47D/XBP1 pooled population were utilized for all subsequent experiments.
|
XBP1 reduces E2-dependence for cell proliferation
To determine whether XBP1(S) overexpression modifies the dependence of some breast cancer cells on estrogen for proliferation, control and XBP1 transfected cells were cultured for 72 h in media devoid of estrogens (CCS-IMEM) before being plated into 12-well plates. Proliferation in the absence of E2 was measured by counting the cells on days 1, 3, and 7. Consistent with their E2 dependence for proliferation, cell number at day 7 is similar to that on day 1 for both MCF7/c (Fig. 3
A) and T47D/c control cells (Fig. 3B
) in the absence of E2. In marked contrast, MCF7/XBP1 (Clone A; P<0.001, ANOVA) and T47D/XBP1 cells (pooled population; P=0.003, ANOVA) continue to proliferate and have acquired an E2-independent phenotype. We also examined the effect of E2 stimulation on the growth of MCF7/XBP1 and T47D/XBP1 cells (Fig. 3C, D
). As shown above in Fig. 3A
, MCF7/XBP1 cells proliferate more than MCF7/c cells in the absence of E2 (0 nM), although the difference between the cell lines appears smaller due to the use of a log scale, and MCF7/c and MCF7/XBP1 Clone A cells both respond similarly to increasing concentrations of E2 (Fig. 3C
; P<0.001 for both cell lines, ANOVA). However, T47D/XBP1 cells proliferate so much more rapidly than T47D/c cells that no additional effect of E2 stimulation is seen (Fig. 3D
; P<0.004 for control, ANOVA). This may be a dose-dependent effect of XBP1(S) overexpression, as MCF7/XBP1 Clone A cells are estimated to express 8-fold more XBP1(S) as compared to 25- to 30-fold more XBP1(S) in T47D/XBP1 cells relative to their respective controls (see Fig. 1
).
|
Sensitivity to the antiproliferative effects of antiestrogens is reduced by XBP1
Since ER+ breast tumors that initially respond to TAM frequently acquire a resistant phenotype (1)
, we determined the effects of XBP1(S) overexpression on the responsiveness of MCF-7 and T47D cells to inhibition by the antiestrogens TAM and FAS. Control and XBP1 transfected cells were seeded in 12-well plates 1 day before treatment with 0.1% EtOH (vehicle control), TAM, or FAS. Consistent with prior studies, the control T47D cells (Fig. 4
A, B, closed symbols) are growth inhibited by both TAM and FAS. However, T47D/XBP1 cells are significantly less responsive to both antiestrogens (Fig. 4A, B
, open symbols; P<0.001, ANOVA) when compared with their respective controls, with an increase in the IC50 of at least three orders of magnitude. MCF7/XBP1 Clone A cells also exhibit a statistically significant reduction in sensitivity to TAM and FAS (Fig. 4A, B
, white bars; P<0.05). Given that MCF7 breast cancer cells express wild-type p53 and T47D cells carry a mutated p53 (40)
, these data suggest that XBP1 induces CRE activity, E2 independence, and antiestrogen-resistant growth in a manner that does not require wild-type p53.
|
XBP1 alters cell cycle progression and reduces antiestrogen-induced apoptosis
Antiestrogens inhibit cell cycle progression, resulting in a reduction in the proportion of cells in S-phase and an increase in the proportion in G0/G1 (41)
. To investigate the role of XBP1(S) in regulating cell cycle progression, cells treated with E2, TAM, FAS, or EtOH (control) were analyzed for alterations in the relative proportions of cells in S-phase using FACS. Consistent with the ability of XBP1(S) to confer E2 independence for proliferation, in the absence of E2 20% of the MCF7/XBP1 cells are in S-phase compared with only 6% of control cells (Fig. 5
A; P<0.05). E2 induces a significant increase in the proportion of control cells in S-phase but only a modest increase in MCF7/XBP1 cells. Reflecting their reduced sensitivity to antiestrogens, both TAM and FAS are significantly less effective in reducing the proportion of cells in S-phase in MCF7/XBP1 compared with control cells (Fig. 5B
; P<0.001, ANOVA). These data clearly show that XBP1(S) increases the proliferative fraction (S-phase) in the absence of E2 and partially protects breast cancer cells from antiestrogen-mediated cell cycle arrest.
|
Since both TAM and FAS can also induce apoptosis, we studied the effects of E2 deprivation and treatment with TAM and FAS on apoptosis by measuring FITC-annexin V and propidium iodide staining by FACS analysis. As expected, the control cells exhibit a high level of apoptosis in the absence of E2 when compared with E2-treated cells (Fig. 5C
; P<0.05). Neither TAM nor FAS is able to induce apoptosis in MCF7/XBP1 cells (Fig. 5D
), whereas TAM induces apoptosis 3-fold and FAS induces apoptosis 6-fold in the respective control cells (Fig. 5D
; P<0.05, ANOVA). We see similar effects of XBP1 overexpression on cell cycle inhibition and the induction of apoptosis by antiestrogens in the T47D cell line. TAM and FAS reduce the percentage of T47D/c cells in S-phase by 50%, but have no effect on cell cycle progression in T47D/XBP1 cells (Fig. 6
A; P
0.005). Furthermore, neither E2 deprivation (Fig. 6B
) nor TAM and FAS treatment (Fig. 6C
) induce apoptosis in T47D/XBP1 cells, whereas significant cell death occurs in T47D/c cells (P<0.05). Thus, XBP1(S) confers an E2-independent and antiestrogen-crossresistant phenotype with respect to both cell cycle arrest and cell survival. Furthermore, XBP1(S) can protect MCF-7 and T47D breast cancer cells from antiestrogen-induced apoptosis and reduce sensitivity to antiestrogen-induced cell cycle arrest.
|
Key genes associated with estrogen responsiveness, apoptosis, and cell cycle are differentially regulated in MCF7/XBP1 cells
Since XBP1(S) is a transcription factor and we have shown (above) that XBP1(S) affects both apoptosis and cell cycle distribution, we performed gene expression microarray analysis on MCF7/XBP1 Clone A cells and evaluated the expression of those genes associated with these two functions within the resulting data set. We used the Gene Ontology database (http://www.geneontology.org) as the source database for determining gene annotations for the categories "apoptosis" (GO:0006915) and "cell cycle" (GO:0007049), since these are the components of the phenotype most relevant to this study. Ontology searches were performed using AmiGo and other tools from the Gene Ontology Project (http://geneontology.org/go.tools.shtml). Gene symbols are those approved by the human gene ontology nomenclature committee (HUGO; http://www.gene.ucl.ac.uk/nomenclature). To find those genes most likely to be directly regulated by XBP1(S), rather than secondary or tertiary downstream events, we then used the MatInspector algorithm (42)
to scan the apoptosis/cell cycle annotated gene list for putative XBP1 CREs in the sequences 2–4 kb upstream of the transcriptional start site of each gene. Within this final gene list (Table 1
), genes were selected for validation either by real-time qPCR or Western hybridization.
|
Of the genes we selected for validation by qPCR, APBB2, CRK, IL24, MYC, PHLDA2, S100A6, and XRCC6 were confirmed to be differentially expressed (Table 1)
. Importantly, we also found several other genes/functions already implicated in our earlier studies on antiestrogen resistance (Table 1)
. For example, we have recently functionally implicated specific BCL2 family members in endocrine resistance (unpublished results), and BCL2 is up-regulated in the XBP1(S) overexpressing cells (1.45-fold; P=0.037; Table 1
). Since BCL2 is included in both the cell cycle and apoptosis functional categories and is predicted to have multiple XBP1 CREs in the upstream regulatory region, we selected BCL2 as a high priority gene for independent validation.
Another gene that is not specifically annotated with the cell cycle or apoptosis functions, but is of clear relevance, is ESR1 (ER). A 1.6-fold up-regulation of ER is detected in MCF7/XBP1 cells (P=0.038). MatInspector analysis did not locate an ACGT-containing CRE in the upstream region out to 9 kb. However, four consensus binding sites were identified for activating transcription factor 6 (ATF6; another bZIP family member) upstream of ESR1. Whereas ATF6 is equivalently expressed in both MCF7/c and MCF7XBP1 cells (qPCR, data not shown), XBP1(S) interacts with ATF6 (43)
to form a heterodimer that also acts as a transcription factor (44)
.
ER expression is increased in MCF7/XBP1 and T47D/XBP1 cells
ER mRNA expression is increased in the MCF7/XBP1 gene expression microarray study, ER is the primary molecular target for estrogens and antiestrogens, and the levels of ER expression can be altered by specific ligands. Thus, we determined whether ER protein expression is altered by XBP1(S) expression and whether TAM or FAS treatment affects ER protein expression in these cells. Western blot analysis confirmed the elevated levels of basal ER expression in both MCF7/XBP1 (Fig. 7
A, B; P=0.04) and T47D/XBP1 cells (Fig. 7C
; P=0.03) when compared to controls. TAM increased ER expression in both control and MCF7/XBP1 cells (P=0.001), whereas FAS inhibited ER expression in both control and MCF7/XBP1 cells. The levels of expression in FAS treated MCF7/c and MCF7/XBP1 cells were not significantly different. Thus, whereas XBP1 can increase ER expression, the regulation of ER protein expression by a SERD is not affected by XBP1(S) overexpression.
|
It has been reported that XBP1(S) can bind to and activate ER in a ligand-independent manner (21)
. Therefore, we performed coimmunoprecipitation analysis to confirm the ability of XBP1(S) to bind to ER in MCF-7 cells. As expected, greater levels of XBP1(S) are immunoprecipitated from MCF7/XBP1 cells than MCF7/c cells (Fig. 8
: compare lanes 3 and 4, top panel). Consequently, more ER is associated with XBP1 in the MCF7/XBP1 cell line (Fig. 8
: compare lanes 3 and 4, bottom panel), consistent with its overall increased expression (Fig. 8
: compare lanes 1 and 2, middle panel). To explore the functional relevance of increased ER expression and XBP1 binding, we also measured ER activity using an ERE-reporter assay. Although we detect minor regulation of the control (Renilla) reporter signal, consistent with published data, a modest increase in ER activity is observed in MCF7/XBP1 cells after accounting for this regulation (not shown). Thus, the ability of XBP1(S) to confer E2-independence and antiestrogen resistance is likely to be partly related to its effects on ER expression and function.
|
BCL2 protein is up-regulated and mitochondrial membrane permeabilization is inhibited in MCF7/XBP1 cells
The antiapoptotic gene BCL2 was also differentially expressed in MCF7/XBP1 as compared to the control cell line in the microarray data. Antiestrogens down-regulate BCL2 expression while increasing apoptosis or programmed cell death, whereas E2 induces BCL2 expression and protects breast cancer cells from apoptosis (reviewed in (45)
. To confirm the microarray data and determine how BCL2 expression is affected by antiestrogens in XBP1(S) overexpressing cells, appropriate cell lysates were subjected to Western blot analysis. BCL2 protein expression is significantly elevated in MCF7/XBP1 cells compared with controls (Fig. 9
; P<0.029). FAS and TAM inhibit BCL2 expression in both MCF7/c and MCF7/XBP1 cells. However, the remaining level of BCL2 expression in FAS-treated MCF7/XBP1 cells is broadly comparable to that in untreated MCF7/c cells. This outcome was not observed in the T47D cell line, in which the BCL2 protein has been previously reported to be undetectable (46
, 47)
; we saw no basal BCL2 protein expression in either T47D/c or T47D/XBP1 cells (data not shown).
|
BCL2 family members control the integrity of the mitochondrial membrane; the induction of apoptosis is frequently accompanied by perturbation of mitochondrial membrane integrity and the release of cytochrome c (48
49
50)
. Thus, to determine how the validated up-regulation of BCL2 in MCF7/XBP1 cells might regulate antiestrogen-mediated apoptosis, we measured the effects of TAM and FAS on mitochondrial membrane permeability. The data in Fig. 10
A show that there is no significant increase in mitochondrial membrane permeability (MMP) in response to either TAM or FAS treatment in MCF7/XBP1 cells. In contrast, MCF7/c control cells exhibit significant increases in mitochondrial membrane permeability in the presence of either antiestrogen (P<0.002, ANOVA).
|
MCF7/XBP1 cells are more sensitive to growth inhibition by the small-molecule BCL2 inhibitor YC-137
Disruption of BCL2-mediated prosurvival signaling is an attractive approach to induce cell death, and breast cancer cell lines expressing high levels of BCL2 are often more sensitive to pharmacological inhibitors of BCL2 (51)
. To determine whether the up-regulation of BCL2 in response to XBP1 overexpression is associated with increased sensitivity to BCL2-specific inhibition, we measured the effects of the small-molecule BCL2 inhibitor YC-137 on cell proliferation (Fig. 10B
). MCF7/c cells are unresponsive to 100 nM YC-137 (black bars), whereas MCF7/XBP1 Clone A cells are significantly growth inhibited (white bars, P=0.03). Overall, these data suggest that the mechanism by which XBP1(S) overexpression protects MCF-7 cells from antiestrogen-mediated apoptosis is through the up-regulation of BCL2 and the prevention of mitochondrial membrane disruption and that BCL2 functions contribute directly to the increased proliferation and survival of these cells.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
7
8)
. XBP1 is readily detectable in breast tumors that are predominantly ER+ (7
, 16)
. Analysis of XBP1 mRNA levels in several breast cancer cell lines has shown that both the spliced and unspliced forms of XBP1 are expressed (21)
. In this study, we show that overexpression of the appropriately spliced XBP1(S) reduces the dependence of both p53 wild-type MCF-7 and p53 mutant T47D human breast cancer cells on estradiol for proliferation. XBP1(S) also reduces sensitivity of these cells to the antiproliferative and proapoptotic activities of TAM and FAS. Using gene expression microarray analysis, we have further identified significant changes in the expression of several genes associated with cell cycle progression, apoptosis, and estrogen independence in XBP1(S) overexpressing cells. These changes include increased expression of both ER and the antiapoptotic molecule BCL2. Finally, we have shown that XBP1(S) overexpression protects MCF-7 cells from antiestrogen-mediated disruption of mitochondrial membrane potential, implying a role for this molecule in the intrinsic apoptotic pathway.
XBP1 induces antiestrogen resistance
The specific mechanisms that control antiestrogen responsiveness in ER+ breast cancers are not completely understood. We now show that overexpression of XBP1(S) results in cross-resistance to the antiestrogens TAM (SERM) and FAS (SERD) by markedly reducing apoptosis and preventing cell cycle arrest in the G0/G1 phase. The reduction in the percentage of cells entering the S-phase observed in our empty vector-transfected control cells supports previously published studies (52)
showing that MCF-7 and T47D cell proliferation is attenuated by antiestrogen treatment. In addition to cell cycle arrest, antiestrogens are also known to induce apoptosis in ER+ breast cancer cells (45)
. Significant apoptosis is induced in MCF7/c and T47D/c control cells, but antiestrogen-mediated cell death is inhibited in the presence of overexpressed XBP1(S). Therefore, our data suggest that XBP1(S) overcomes the inhibitory effects of antiestrogen treatment by stimulating both basal and estrogen-independent cell cycle progression and by preventing drug-induced apoptosis (Fig. 5)
. Importantly, we have found that XBP1(S) overexpression reduces breast cancer cell sensitivity to two different classes of antiestrogen and in two different ER+ human breast cancer cell lines. TAM is a nonsteroidal, partial antagonist of ER (SERM) that is frequently used as a first-line antihormonal therapy in ER+ breast cancer patients. FAS is a steroidal antagonist (SERD) that has activity comparable to some aromatase inhibitors in postmenopausal women (5)
and is effective in tumors that have progressed on TAM (4)
. In addition to inhibiting the transcriptional activity of ER, FAS has also been shown to accelerate the degradation of ER protein, leading to its classification as a SERD.
The observation that XBP1(S) overexpression confers resistance to both SERM and SERD antiestrogens is consistent with our previously published finding that both XBP1 expression and CRE promoter-reporter activity are increased in FAS-resistant, TAM-crossresistant cells as compared with antiestrogen-sensitive cells (8)
. These data support the idea that XBP1(S) is an important node in a wider gene network that confers resistance to FAS. The prosurvival activities of XBP1(S) suggest that either treatment naive or breast tumors recurring after initial endocrine therapy may be less responsive to many subsequent ER-targeted endocrine therapies if they express high levels of the appropriately spliced XBP1.
XBP1 and estrogen independence
XBP1(S) overexpression also induces estrogen independence in both MCF-7 and T47D cells. Transfected cells continue to proliferate under estrogen-deprived conditions, and although control cells show a reduction in S-phase entry and increased apoptosis, XBP1-transfected cells are unaffected. Molecular mechanisms that have been proposed to control estrogen independence vary widely. One suggestion has been that estrogen-independent (and antiestrogen-resistant) breast cancer cells have either down-regulated or lost ER expression. However, XBP1-transfected cells have increased levels of ER mRNA and protein (Table 1
; Fig. 7
). Both antiestrogen resistance and estrogen independence have been attributed to alterations in signal transduction pathways that either do not directly interface with ER or participate in crosstalk with ER through signaling intermediates (45
, 53)
. These mechanisms can directly affect breast cancer cell cycle progression and apoptosis, enabling these cells to switch from an ER-driven growth phenotype to one dependent on alternative proliferation and survival pathways.
The ability to confer E2-independent growth on estrogen-dependent cells strongly implicates XBP1 activity in resistance to aromatase inhibitors. Since XBP1(S) can confer resistance to SERMs, SERDs, and possibly aromatase inhibitors, XBP1 may be useful in identifying those ER+ breast cancers that have a relatively poor response to endocrine therapy. For example, 25% of ER+/PgR+ tumors exhibit de novo endocrine resistance, and tumors with acquired resistance to first line therapy generally exhibit lower overall response rates and responses of shorter duration to subsequent endocrine therapies than they did to their first line therapy (1)
.
XBP1 regulates genes associated with the unfolded protein response, cell cycle, and apoptosis
XBP1-mediated antiestrogen resistance is characterized by increased cell cycle progression and reduced levels of apoptosis. As a transcription factor, it is likely that XBP1(S) controls the resistance phenotype by regulating the transcription of cell cycle- and apoptosis-associated genes. In our gene expression microarray studies, we found several differentially expressed genes that are annotated with these functions and also contain putative ACGT CREs in their 5' sequences. For example, IL24 is strongly downregulated by XBP1(S) overexpression (–2.68-fold by microarray and –9.7 fold by qPCR). Also known as melanoma differentiation-associated gene 7 (mda-7) or interleukin-10B (IL-10B), IL24 can induce apoptosis in breast cancer cells (54)
and is also implicated in the unfolded protein stress response (55)
. Although this gene has not been previously associated with estrogen or antiestrogen responsiveness, down-regulation of proapoptotic molecules like IL24 may plays a functional role in resistance. Similarly, PHLDA2 (also known as IPL or TSSC3) is significantly down-regulated in MCF7/XBP1 cells (–1.8-fold by microarray, and –3.3-fold by qPCR). PHLDA2 is one of the few apoptosis-associated genes known to be susceptible to genetic imprinting. Moreover, PHLDA2 is located on chromosome 11p15, a region thought to harbor tumor suppressor activities and that is also altered in several cancers including breast cancer (56
, 57)
. In the mouse, where expression of maternal PHLDA2 is highest in extraembryonic structures, knockout of this gene leads to overgrowth of the placenta (58)
, suggesting that dysregulation of PHLDA2 can promote hyperproliferation.
In contrast to the inhibition of tumor suppressor-like activities, we observe up-regulation by XBP1(S) of the proproliferative and/or prosurvival genes c-Myc and XRCC6 and the calcium-binding protein S100A6 (calcyclin). Increased S100A6 expression has been demonstrated in several different neoplastic vs. normal tissues (59)
and in 32% of breast tumors (60)
. In pulmonary fibroblasts, S100A6 antisense oligonucleotides inhibit cell proliferation, (61)
. In WI-38 embryonic lung cells, any delay in the up-regulation of S100A6 transcripts in late G1 is accompanied by a delay in S-phase entry (62)
. The molecular mechanism(s) by which S100A6 regulates proliferation is unclear, but its calcium binding function may be important. XRCC6 (also known as the Ku70 autoantigen) is a DNA repair molecule essential to the end-joining process (reviewed in ref. 63
). Two single nucleotide polymorphisms in the XRCC6 gene are strongly associated with breast cancer risk (64)
, and XRCC6 overexpression has recently been implicated in acquired cisplatin resistance in ovarian cancer cells (65)
. Finally, c-Myc is a well-established prosurvival molecule in human breast cancer and mouse mammary tumorigenesis (66)
. Capable of modulating both cell growth and cell death, c-Myc is overexpressed in 70% of breast tumors and amplified at least 3-fold in 16% of breast tumors (67)
. Moreover, inducing c-Myc expression in MCF-7 cells can confer resistance to FAS (68)
. To our knowledge, this is the first report that XBP1(S) can regulate c-Myc expression in breast cancer cells. Whether the regulation of c-Myc signaling by XBP-1(S) contributes to the resistance phenotype in T47D cells is currently under investigation.
Together, these five genes may comprise a putative signaling subnetwork, controlled by XBP1(S), which can regulate breast cancer cell proliferation, survival, and endocrine therapy response. Figure 11
shows the relationships among XBP1(S) and downstream molecules as they appear to occur in MCF7/XBP1 cells. Pair-wise correlation coefficient analyses of the gene expression microarray data show that IL24/mda-7 and BCL2 exhibit a significant negative correlation in MCF7/XBP1 cells overexpressing XBP1(S) (r=–0.95; P=0.01). An inverse relationship between these two genes occurs in both melanoma and prostate cancer cell lines (69
, 70)
. The IL24/mda-7 secreted by normal cells may selectively induce apoptosis in adjacent tumor cells, and IL24 plus radiation kills tumor cells overexpressing BCL2 that are normally resistant to single agent treatment (55)
. Consistent with their opposing activities with respect to proliferation (PHLDA2 is growth inhibitory; XRCC6 is growth stimulatory), the expression of XRCC6 and PHLDA2 also is significantly negatively correlated (r=–0.95; P=0.02).
|
XBP1 regulates ER expression
Gene expression microarray analysis also revealed that ER mRNA expression is increased in MCF7/XBP1 cells, despite the apparent lack of an ACGT CRE in its upstream regulatory region. Subsequent experiments showed that ER protein expression is significantly enhanced by XBP1(S) in both MCF-7 and T47D cells, exhibits greater transcriptional activation, and shows increased coimmunoprecipitation with XBP1(S) in MCF-7 cells. Importantly, the regulation of ER protein expression by antiestrogens is not affected by XBP1(S) overexpression, since FAS treatment down-regulates ER in both control and XBP1-transfected cells. Ding et al. (21)
have previously reported that XBP1(S) binds ER and that XBP1-mediated enhancement of ER transcriptional activity occurs in a ligand-independent fashion. However, up-regulation of ER expression by XBP1(S) was not observed in their study, most likely because they utilized transient transfection techniques rather than the generation of cell lines stably overexpressing transcriptionally active XBP1. Lack of ACGT consensus sequences upstream of the ER start site suggests that transcription of this gene may be indirectly regulated by XBP1(S), possibly through heterodimerization with ATF6. ATF6 is expressed in both MCF7/c and MCF7XBP1 cells and several ATF6 binding sites are present in the ER promoter. Regardless of whether the up-regulation of ER is a direct or indirect effect of XBP1 expression, our data suggest that increasing the levels of ER under conditions in which ligand-independent activation of the receptor occur (as could arise in the presence of an aromatase inhibitor) is a likely contributor to the endocrine resistance phenotype induced by XBP1.
XBP1 regulates BCL2 expression and mitochondrial membrane permeability
The gene expression microarray data show that the prosurvival/antiapoptotic gene BCL2 is also up-regulated by XBP1(S) in MCF-7 cells. This outcome was not observed in the T47D cell line, in which the BCL2 protein has been previously reported to be undetectable (46
, 47)
; we saw no basal BCL2 protein expression in either T47D/C or T47D/XBP1 cells. These data imply that the mechanism by which XBP1-overexpressing cells escape apoptosis may differ depending on cellular context.
BCL2 has two promoters (P1 and P2; ref. 71
), and within both promoters there are predicted ACGT-containing CREs. Thus, XBP1(S) likely regulates BCL2 expression by affecting its rate of transcription. BCL2 is located in the outer membrane of the mitochondria, and mitochondria play an important role in the regulation of apoptosis (72
, 73)
. In the intrinsic cell death pathway, Bax and Bad homodimerize in the outer membrane, causing membrane permeabilization that in turn leads to cytochrome c release, activation of the proapoptotic molecules Smac/DIABLO and Omi/HtrA2 (74)
, and ultimately increased caspase activity (48)
. BCL2 prevents intrinsic apoptosis by binding to and inhibiting Bax or Bad function, thus preventing mitochondrial membrane permeabilization, cytochrome c release, and caspase activation (48)
. Importantly, XBP1(S) over-expression robustly inhibits antiestrogen-induced permeability of the mitochondrial membrane in MCF-7 cells.
The ability of breast cancer cells to undergo apoptosis when challenged with antiestrogens or other therapeutic agents is a critical factor in the success of breast cancer treatment. Antiestrogen-induced apoptosis likely plays a larger role in promoting disease regression than cell cycle arrest. If antiestrogen therapy leads primarily to arrest in G0/G1, then significant tumor shrinkage and the widespread elimination of tumor cells would likely not be observed (45)
. However, it has been repeatedly demonstrated that antiestrogen treatment concomitantly decreases breast tumor size and improves overall survival, strongly suggesting that drug-induced cell death is a key component of effective antiestrogen action. We propose that the ability of XBP1(S) to abolish antiestrogen-mediated apoptosis, possibly through up-regulating BCL2 expression and perturbing the intrinsic apoptotic pathway, is an important mechanism of antiestrogen resistance that merits further investigation in the clinical setting.
|
ACKNOWLEDGMENTS |
|---|
Received for publication December 20, 2006. Accepted for publication June 7, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B inhibitor parthenolide restores ICI 182,780 (Faslodex; Fulvestrant)-induced apoptosis in antiestrogen resistant breast cancer cells. Mol. Cancer Ther. 4,33-41This article has been cited by other articles:
|
|
 |
|
  B. Zhang, H. Li, R. B. Riggins, M. Zhan, J. Xuan, Z. Zhang, E. P. Hoffman, R. Clarke, and Y. Wang Differential dependency network analysis to identify condition-specific topological changes in biological networks Bioinformatics, February 15, 2009; 25(4): 526 - 532. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. B. Riggins, J. P-J. Lan, U. Klimach, A. Zwart, L. R. Cavalli, B. R. Haddad, L. Chen, T. Gong, J. Xuan, S. P. Ethier, et al. ERR{gamma} Mediates Tamoxifen Resistance in Novel Models of Invasive Lobular Breast Cancer Cancer Res., November 1, 2008; 68(21): 8908 - 8917. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
= induction; 
= inhibition.