Steroid receptor coactivator-1 upregulates integrin α₅ expression to promote breast cancer cell adhesion and migration

Abstract

Metastatic breast cancer remains a lethal disease with poorly understood molecular mechanisms. Steroid receptor coactivator-1 (SRC-1 or NCOA1) is overexpressed in a subset of breast cancers with poor prognosis. It potentiates gene expression by serving as a coactivator for nuclear receptors and other transcription factors. We previously reported that SRC-1 promotes breast cancer metastasis without affecting primary mammary tumor formation. Herein, we found that SRC-1 deficiency in mouse and human breast cancer cells substantially reduced cell adhesion and migration capabilities on fibronectin and significantly extended the time of focal adhesion disassembly and reassembly. In agreement with this phenotype, SRC-1 expression positively correlated with integrin α(5) (ITGA5) expression in estrogen receptor-negative breast tumors whereas SRC-1 deficiency decreased ITGA5 expression. Furthermore, ITGA5 reduction in SRC-1-deficient/insufficient breast cancer cells or knockdown of ITGA5 in SRC-1-expressing breast cancer cells was associated with a disturbed integrin-mediated signaling. Critical downstream changes included reduced phosphorylation and/or dampened activation of focal adhesion kinase, paxillin, Rac1, and Erk1/2 during cell adhesion. Finally, we found that SRC-1 enhanced ITGA5 promoter activity through an AP-1 (activator protein)-binding site proximal to the transcriptional initiation site; both SRC-1 and c-Jun were recruited to this promoter region in breast cancer cells. These results show that SRC-1 can promote breast cancer metastasis by directly enhancing ITGA5 expression and thus promoting ITGA5-mediated cell adhesion and migration. Therefore, targeting ITGA5 in SRC-1-positive breast cancers may result in inhibition of SRC-1-promoted breast cancer metastasis.

Fig. 1. SRC-1 deficiency reduces BC cell adhesion and migration on FN-coated plates

A. Phase contrast images of SRC-1 WT1 and KO1 mammary tumor cell lines after culturing on FN-coated plates for time periods indicated (left panel), and percentages of adhered KO1, KO2, WT1 and WT2 cells at 1 hour culture time (right panel). *, p<0.05 by t test. Western blot results showed SRC-1 protein presence in WT1 and WT2 cells and absence in KO1 and KO2 cells. B. Percentages of adhered SRC-1 KO;PyMT (n=5) and WT;PyMT (n=5) primary mammary tumor cells after culturing on FN-coated plates for 1 hour. C. Left panel: Migration track areas of SRC-1 KO1, KO2, WT1 and WT2 cell lines on FN-coated plates. Average migration area was calculated from the track areas of 50 cells. Middle panel: Knockdown of SRC-1 mRNA in WT1 and WT2 cells using siRNAs; non-targeting siRNA was used as a control. Right panel: SRC-1 knockdown reduced the average migration areas of WT1 and WT2 cells. D. Left panel: knockdown of SRC-1 mRNA in MDA-MB-231 cells using three shRNA constructs; non-targeting shRNA was used as a control. Right panel: MDA-MB-231 cells with SRC-1 knockdown showed reduced cell migration as measured by tracing the migration areas with fluorescent beads.

Fig. 2. The disassembly and reassembly of FACs and ITGA5 expression in SRC-1 KO and WT mammary tumor cells

A. Different distribution patterns of FACs and F-actin in SRC-1 KO1 and WT1 cells as revealed by vinculin immunostaining (green) and philloidin staining (red). Cells were cultured on FN-coated plate for one hour. Arrows indicate FACs. B. FAC disassembly and reassembly in SRC-1 KO1 and WT1 cells were monitored at the time points indicated by vinculin immunostaining. C. FAC disassembly in SRC-1 siRNA-transfected WT1 cells was slower than that in control siRNA-transfected WT1 cells. D. Measurements of ITGB1, ITGB3, ITGAv and ITGA5 mRNA levels by qPCR, as well as ITGA5 protein levels by Western blot in SRC-1 KO1, KO2, WT1 and WT2 cells. *, p<0.05 by t test.

Fig. 3. SRC-1 expression positively correlates with ITGA5 expression

A. Adenovirus-mediated SRC-1 expression in SRC-1 KO1 and KO2 cells enhanced ITGA5 expression. Adenovirus-mediated GFP expression served as a control (left panel). Conversely, siRNA-mediated knockdown of SRC-1 decreased ITGA5 mRNA and protein levels in SRC-1 WT1 and WT2 cells. The non-targeting siRNA served as a control (middle and right panels). *, p<0.05 by t test. B. The shRNA-mediated stable knockdown of SRC-1 in MDA-MB-231 cells reduced ITGA5 expression. C. Immunohistochemical staining (brown color) of ITGA5 in SRC-1 WT;PyMT and KO;PyMT tumors isolated from 8- and 14-week-old mice. Slides were counterstained with hematoxylin. D1. Western blot analysis of SRC-1 and ITGA5 proteins in ERα+;PR+ (n=11), ERα+;PR- (n=5) and ERα-;PR- (n=8) human breast tumors was performed as shown in Supplementary Figure 1. β-actin was assayed as a loading control. Band intensities were measured by densitometry. Relative ITGA5 and SRC-1 levels were normalized to β-actin and presented here as bar graphs. *, p<0.05, and **, p<0.01 by F test. D2. SRC-1 mRNA levels correlate with ITGA5 mRNA levels in human ductal breast carcinoma. The microarray data were collected by M. Bittner and downloaded from Oncomine Database. The SRC-1 and ITGA5 expression data shown in Supplementary Figure 2 were plotted and statistically analyzed by Pearson correlation. The p value indicates that the expression levels of SRC-1 significantly correlate with the expression levels of ITGA5 in the human ductal breast carcinomas.

Fig. 4. Effects of ITGA5 reduction in SRC-1 deficient cells on integrin-signaling pathways and cell migration

A. Western blot analysis of pY397-FAK, total FAK (t-FAK), p-Paxillin, t-Paxillin, active Rac1 (a-RAC1) and t-RAC1 in SRC-1 WT1, WT2, KO1 and KO2 cells and WT1 cells transfected with siRNAs against SRC-1 (siSRC-1) or ITGA5 (siITGA5) mRNA or with non-targeting siRNA (control). Cells were allowed to adhere to FN-coated plates for 30 or 60 minutes prior to the analysis. The relative band intensities (indicated) were normalized to total protein band intensities. B. Western blot analysis of pY397-FAK, p-Paxillin and β-actin in MDA-MB-231 cells with stable expression of non-targeting shRNA (Ctrl) or one of the three shRNAs targeting different regions of the SRC-1 mRNA. C. Western blot analysis of pY925-FAK, t-FAK, p-Erk1/2 and t-Erk1/2 in SRC-1 WT1, WT2, KO1 and KO2 cells after cultured on the FN-coated plates for 30 or 60 minutes. Relative band intensities are indicated. D. WT1 and WT2 cells were transfected with either control siRNA or siRNAs targeting ITGA5 mRNA. ITGA5 mRNA levels in these cells were measured by qPCR (left panel). These cells were subjected to migration assays on FN- and fluorescence bead-coated plate (middle panel). The track areas of individual cell migration were traced and averaged from about 50 cells for each group. *, p<0.05 by t test.

Fig. 5. SRC-1 regulates ITGA5 promoter activity

A. The pGL3 plasmid constructs containing different ITGA5 promoter fragments, F1, F2 or F3, which are linked to the luciferase (Luc) reporter sequence. The computer-mapped binding sites for AP-1, NF-κB, PEA3 and C/EBP are indicated. The AP-1 site at bp -9 is deleted in pGL3-MF1-Luc, pGL3-MF2-Luc and pGL3-MF3-Luc constructs. B. Expression of SRC-1 in HeLa cells enhances pGL3-F1-Luc, pGL3-F2-Luc and pGL3-F3-Luc activities in a dose dependent manner. C. Effects of SRC-1 co-expression with PEA3, NF-κB, C/EBPα or C/EBPβ on the activities of pGL3-F1-Luc, pGL3-F2-Luc and pGL3-F3-Luc in HeLa cells. D. Co-expression of SRC-1 and c-Jun in HeLa cells significantly enhanced the activities of pGL3-F1-Luc, pGL3-F2-Luc and pGL3-F3-Luc, but not the mutant pGL3-MF1-Luc, pGL3-MF2-Luc and pGL3-MF3-Luc.

Fig. 6. SRC-1 and AP-1 are associated with a proximal region of the ITGA5 promoter

A. ChIP assays were designed to detect the a, b, c, d and e regions of the ITGA5 promoter with specific PCR primer pairs as indicated. Predicted AP-1 binding sites are indicated. B and C. ChIP assays were performed with MDA-MB-231 and MDA-MB-435 cells and antibodies against SRC-1 (panel B) and c-Jun (panel C). DNA extracted from cross-linked and sonicated cell lysate was used as input control. IgG served as a control for antibody specificity in ChIP assays. No template (-template) PCR reaction served as a negative control for PCR. PCR-amplified fragments a, b, c, d and e are indicated. The P indicates the primer bands. The ns indicates a non-specific PCR product.