Identification and characterization of nucleolin as a COUP-TFII coactivator of retinoic acid receptor β transcription in breast cancer cells

Abstract

Introduction: The orphan nuclear receptor COUP-TFII plays an undefined role in breast cancer. Previously we reported lower COUP-TFII expression in tamoxifen/endocrine-resistant versus sensitive breast cancer cell lines. The identification of COUP-TFII-interacting proteins will help to elucidate its mechanism of action as a transcriptional regulator in breast cancer.

Results: FLAG-affinity purification and multidimensional protein identification technology (MudPIT) identified nucleolin among the proteins interacting with COUP-TFII in MCF-7 tamoxifen-sensitive breast cancer cells. Interaction of COUP-TFII and nucleolin was confirmed by coimmunoprecipitation of endogenous proteins in MCF-7 and T47D breast cancer cells. In vitro studies revealed that COUP-TFII interacts with the C-terminal arginine-glycine repeat (RGG) domain of nucleolin. Functional interaction between COUP-TFII and nucleolin was indicated by studies showing that siRNA knockdown of nucleolin and an oligonucleotide aptamer that targets nucleolin, AS1411, inhibited endogenous COUP-TFII-stimulated RARB2 expression in MCF-7 and T47D cells. Chromatin immunoprecipitation revealed COUP-TFII occupancy of the RARB2 promoter was increased by all-trans retinoic acid (atRA). RARβ2 regulated gene RRIG1 was increased by atRA and COUP-TFII transfection and inhibited by siCOUP-TFII. Immunohistochemical staining of breast tumor microarrays showed nuclear COUP-TFII and nucleolin staining was correlated in invasive ductal carcinomas. COUP-TFII staining correlated with ERα, SRC-1, AIB1, Pea3, MMP2, and phospho-Src and was reduced with increased tumor grade.

Conclusions: Our data indicate that nucleolin plays a coregulatory role in transcriptional regulation of the tumor suppressor RARB2 by COUP-TFII.

Figure 1. Endogenous nuclear nucleolin-COUP-TFII interaction in MCF-7 and T47D cells.

NE (200 µg protein) from MCF-7 cells (A) and (400 µg protein) from T47D (B) cells were immunoprecipitated with COUP-TFII antibody or with rabbit IgG (negative control), followed by western blot analysis for nucleolin and COUP-TFII. 5% input NE serves as loading control. C, Immunofluorescent staining of endogenous COUP-TFII (green) and nucleolin (red) in the nuclei (Hoechst, blue) of MCF-7 cells. Merged images are shown at the right. Bar is 10 µm. D, schematic representation of the N- terminal maltose binding protein (MBP)-tagged recombinant nucleolin proteins used for MBP pull-down assays. MBP was fused to the N-termini of the RNA binding domains (RBD) and/or the arginine/glycine-rich domain (RGG) of nucleolin. E, In vitro transcribed/translated COUP-TFII was incubated with the MBP-nucleolin fragments or MBP. Interacting proteins were captured with amylose resin. Eluted proteins were probed for COUP-TFII (top) and MBP (bottom, control).

Figure 2. COUP-TFII and nucleolin in breast cancer tissue microarrays.

A and B, COUP-TFII immunostaining at 200×: A, benign breast tissue (H-score 30) and B, invasive ductal carcinomas, grade 2 (H-score 153). Bar is 200 mm. C, Average ± SEM of H-score for nuclear COUP-TFII staining in ERα-positive invasive ductal carcinomas by tumor grade. * significantly different from T2N0M0 (p<0.05). D and E, Nucleolin immunostaining at 400×: D, benign breast tissue (H-score 13) and E, invasive ductal carcinomas, grade 3 (H-score 151). Bar is 100 µm. F and G, immunohistochemical localization of COUP-TFII (100×, inset: 200×) on a tissue microarray constructed from archival tissue from 332 breast cancer patients showing positively (F) and negatively (G) stained cores COUP-TFII at 200×, bar is 100 µm. H and I, Kaplan-Meier estimates of disease-free survival functions were computed, and the Wilcoxon test was used to compare survival curves. In addition, the Wilcoxon rank sum test was used to compare two medians. The data are not statistically significant.

Figure 3. COUP-TFII increases RARβ2 transcription in MCF-7 cells.

A, ChIP of COUP-TFII-FLAG to the RARB2 promoter in MCF-7 cells transfected with empty vector (EV) or COUP-TFII (CII) and treated with EtOH or 1 µM atRA for 6 h. * P<0.05 versus E control, ‡ P<0.05 versus CII-EtOH. B-C, Cells were transfected with parental or expression plasmids for COUP-TFII or nucleolin for 24 h and were treated with EtOH, 10 µM CRO or AS1411 for 24 h post-transfection. For C, cells were transfected with 0.5 µg and 1 µg of nucleolin and COUP-TFII expression vector, respectively. Q-PCR was performed to determine RARB2 expression. Values are the average of 6 separate experiments ± SEM. D, Cells were transfected with pcDNA or pcCOUP-TFII and treated with 10 µM RO or AS1411 for 24 h. Dual luciferase activity was expressed relative to the pcDNA-transfected, no-treatment control. Values are mean ± S.E.M. of two separate experiments. For B–D, * P<0.05 versus vector control, ** COUP-TFII alone, or ▴ between the indicated values.

Figure 4. Regulation of RARβ2 transcription.

A and B, Q-PCR for RARB2 (B RARβ2) and NR2F2 (C COUP-TFII) in MCF-7 or T47D cells treated with EtOH, 1 µM atRA, or 1 µM 9-cis-RA for 24 h. Values are the average of 3–5 separate experiments. * P<0.05 versus EtOH. C, MCF-7 and D, T47D cells were transfected with 2 µg pCMV-tag2 (−) or pCMV-tag2-nucleolin (+) for 24 h prior to 24 h treatment with EtOH, 10 µM CRO, or 10 µM AS1411. Where indicated, cells were treated with 1 µM atRA, 100 nM 4-OHT, or 100 nM ICI 182,780 for 6 h. Q-PCR for RARB2 expression. Values are the average of 6 (MCF-7) and 4–10 (T47D) separate experiments ± SEM. * P<0.05 versus EtOH or ** between the indicated values.

Figure 5. Reduction of COUP-TFII or nucleolin decreases RARβ2 transcription in MCF-7 cells.

MCF-7 (A) and T47D (B) cells were transfected with control siRNA or an siRNA targeting nucleolin for 48 h. T47D cells were treated with EtOH or 1 µM atRA for 24 h. Q-PCR for nucleolin (NCL) and RARB2. Values are the average of triplicates. C, Western blot showing COUP-TFII and RARβ2 expression after transfection with siCOUP-TFII. Values are relative to β-actin. MCF-7 were transfected with siControl or siCOUP-TFII for 48 h and treated with 1 µM atRA for 6 h. Q-PCR was also performed for RRIG1. P<0.001 * versus control or ** versus atRA.

Figure 6. Effects of AS1411 on nuclear nucleolin-COUP-TFII interaction.

A, Representative western blots of CE and NE (30 µg) from MCF-7 cells: untreated (untx) or treated with 10 µM CRO or AS1411 for 24 h were probed for nucleolin or COUP-TFII, then stripped and reprobed for HDAC1 and α-tubulin. B, Relative nucleolin and COUP-TFII expression (normalized to respective loading controls and untreated sample protein ratios were set to 1 for NE and CE). Bars are the mean ± SEM of 4–8 separate experiments. C, NE or CE (200 µg) from MCF-7 cells treated as above were IPed with COUP-TFII antibody or rabbit IgG. The Ab-no NE lane was a negative control: COUP-TFII antibody incubated with beads and buffer without NE. D, The ratio of nuclear nucleolin/COUP-TFII is the mean ± SEM of 3 experiments.

Figure 7. Model of nucleolin-COUP-TFII interaction and upregulation of RARB2 expression.

COUP-TFII binds its response elements as either a homodimer or as a heterodimer with RXR . Previous reports demonstrated that 1) 9cisRA and atRA bind COUP-TFII and increase COUP-TFII transcriptional activity ; 2) Nucleolin acts as a transcriptional coregulator by interacting with cJun and Sp1 ; 3) COUP-TFI and COUP-TFII increase RARB2 expression in cooperation with RARα and CBP . Here we demonstrated that 1) nucleolin interacts directly with nuclear COUP-TFII; 2) atRA and 9-cisRA increased RARB2 mRNA; 3) AS1411, used as a functional inhibitor of nucleolin , , inhibited COUP-TFII-upregulation of RARB2 gene transcription; 4) siRNA knockdown of nucleolin reduces induction of RARB2 and reduced RARβ2 protein.