A systematic evaluation of miRNA:mRNA interactions involved in the migration and invasion of breast cancer cells

Abstract

In this study we performed a systematic evaluation of functional miRNA-mRNA interactions associated with the invasiveness of breast cancer cells using a combination of integrated miRNA and mRNA expression profiling, bioinformatics prediction, and functional assays. Analysis of the miRNA expression identified 11 miRNAs that were differentially expressed, including 7 down-regulated (miR-200c, miR-205, miR-203, miR-141, miR-34a, miR-183, and miR-375) and 4 up-regulated miRNAs (miR-146a, miR-138, miR-125b1 and miR-100), in invasive cell lines when compared to normal and less invasive cell lines. Transfection of miR-200c, miR-205, and miR-375 mimics into MDA-MB-231 cells led to the inhibition of in vitro cell migration and invasion. The integrated analysis of miRNA and mRNA expression identified 35 known and novel target genes of miR-200c, miR-205, and mir-375, including CFL2, LAMC1, TIMP2, ZEB1, CDH11, PRKCA, PTPRJ, PTPRM, LDHB, and SEC23A. Surprisingly, the majority of these genes (27 genes) were target genes of miR-200c, suggesting that miR-200c plays a pivotal role in regulating the invasiveness of breast cancer cells. We characterized one of the target genes of miR-200c, CFL2, and demonstrated that CFL2 is overexpressed in aggressive breast cancer cell lines and can be significantly down-regulated by exogenous miR-200c. Tissue microarray analysis further revealed that CFL2 expression in primary breast cancer tissue correlated with tumor grade. The results obtained from this study may improve our understanding of the role of these candidate miRNAs and their target genes in relation to breast cancer invasiveness and ultimately lead to the identification of novel biomarkers associated with prognosis.

Figure 1

Differential miRNA expression in 12 breast cancer cell lines. (A) A volcano plot shows that 19 out of 847 miRNAs were differentially expressed (p < 0.01 and fold-change >2). 11 miRNAs (labeled) have an average fold-change above 10 between aggressive and less-aggressive groups. (B) Heatmap representing the expression values of 11 miRNAs that are differentially expressed between aggressive and less aggressive groups. (C) qRT-PCR confirmation of the miRNA array results for miR-200c, miR-205, miR-375, and miR-146a in the 12 breast cell lines.

Figure 2

Migration and invasion assays for miR-200c, miR-205, miR-375 mimic transduced MDA-MB-231 cells. (A) Migration assay for miRNA-transduced MDA-MB-231 cells (100×). (B) Invasion assay for miRNA-transduced MDA-MB-231 cells (100×). (C) Bar graph representing the percentage of reduction in cell migration after the respective miRNA transfection. (D) Bar graph representing the percentage of reduction in cell invasion after the respective miRNA transfection. *: Student’s T-test, p < 0.05; **: Student’s T-test <0.01.

Figure 3

Differential gene expression in 12 breast cancer cell lines. (A) A volcano plot shows that 2412 genes were differentially expressed (p < 0.05 and fold-change >1.5) between normal and less aggressive cell lines vs. aggressive cell lines. (B) The number of predicted miRNA target genes that were up or down-regulated in aggressive breast cancer cell lines. (C) The percentage of predicted miRNA target genes that were up and down-regulated in aggressive breast cancer cell lines. The number above each bar represents the number of predicted target genes of each miRNA by TargetScan 6.2. (D) Confirmation of 9 miRNA target genes using qRT-PCR in 12 breast cancer cell lines. Fold change was calculated using ΔΔCt method. MCF10A was used as the control and therefore the fold change for MCF10A is 1 in all plots.

Figure 4

Integrative analysis of mRNA and miRNA expression in breast cancer cells. (A) Heatmap representing the differentially expressed genes in miR-200c, miR-205, and miR-375 mimic transfected MDA-MB-231 cells. (B) Venn diagrams showing the intersection between: green circle, mRNA transcripts displaying at least a 1.4-fold decrease in expression in miR-200c, miR-205, and miR-375 mimic transfected MDA-MB-231 cells; blue circle, mRNA transcripts (p < 0.05 and fold change >1.5) that showed an increased expression in aggressive vs. normal and less aggressive cell lines. red circle, likely mRNA targets for miR-200c, miR-205, and miR-375 predicted using TargetScan6.2. The grey area indicates the intersection of all 3 circles. (C-E) qRT-PCR analysis of 11 candidate target genes of miR-200c (C), miR-205 (D), and miR-375 (E) in miR-200c, miR-205, and miR-375 transduced breast cancer cell lines, BT549, HS578T, MDA-MB-231 and SUM159, respectively.

Figure 5

3^′-UTR reporter assays confirming the interaction of individual miRNA with the 3^′-UTR of candidate target genes. (A) Schema of the pmirGLO dual luciferase vector carrying the 3^′-UTR regions of 6 selected genes. The entire 3^′-UTR of 6 genes: CDH11, CFL2, SEC23A, ZEB-1, PTPRM, and LDHB were cloned into the pmirGLO dual luciferase vector. (B) The alignments of miRNA and their predicted target genes. (C) Percent relative luciferase activity 48 hours post-transfection with the indicated reporter vector.

Figure 6

CFL2 and breast cancer cell migration. (A) Bar graph representing the normalized number of migratory or invaded cells after the respective siRNA transfection. (B) Overexpression of CFL2 can compensate the effect of exogenous miR-200c. Bar graph represents the normalized number of migratory cells after the respective transfection of miRNA mimic control, miR-200c mimic, and co-transfection of miR-200c and a plasmid encoding CFL2, respectively. *: Student’s T-test, p < 0.05. (C). Rhodamin-phalloidin staining of F-actin in the CFL2-siRNA and miR-200c transfected MDA-MB-231 cells. (D) Immunohistochemical analysis of TMA of primary breast tumors using antibodies against CFL2. Representative images of TMA cores are presented showing normal breast tissue, grades I, II, and III breast tumors with increasing staining intensities. (E) Bar graph representing the quantification of CFL2 immunostaining from blindly scored tissue microarray sections. Staining intensity of each sample was given a modified histochemical score (MH-score) that considers both the intensity and the percentage of cells stained at each intensity level. The intensity of each grade is the average of MH-score of all samples in that grade. Data were analyzed by one-way ANOVA with p-values as noted in the figure.