MicroRNA-378-mediated suppression of Runx1 alleviates the aggressive phenotype of triple-negative MDA-MB-231 human breast cancer cells

Abstract

The Runx1 transcription factor, known for its essential role in normal hematopoiesis, was reported in limited studies to be mutated or associated with human breast tumor tissues. Runx1 increases concomitantly with disease progression in the MMTV-PyMT transgenic mouse model of breast cancer. Compelling questions relate to mechanisms that regulate Runx1 expression in breast cancer. Here, we tested the hypothesis that dysregulation of Runx1-targeting microRNAs (miRNAs) allows for pathologic increase of Runx1 during breast cancer progression. Microarray profiling of the MMTV-PyMT model revealed significant downregulation of numerous miRNAs predicted to target Runx1. One of these, miR-378, was inversely correlated with Runx1 expression during breast cancer progression in mice and in human breast cancer cell lines MCF7 and triple-negative MDA-MB-231 that represent early- and late-stage diseases, respectively. MiR-378 is nearly absent in MDA-MB-231 cells. Luciferase reporter assays revealed that miR-378 binds the Runx1 3' untranslated region (3'UTR) and inhibits Runx1 expression. Functionally, we demonstrated that ectopic expression of miR-378 in MDA-MB-231 cells inhibited Runx1 and suppressed migration and invasion, while inhibition of miR-378 in MCF7 cells increased Runx1 levels and cell migration. Depletion of Runx1 in late-stage breast cancer cells resulted in increased expression of both the miR-378 host gene PPARGC1B and pre-miR-378, suggesting a feedback loop. Taken together, our study identifies a novel and clinically relevant mechanism for regulation of Runx1 in breast cancer that is mediated by a PPARGC1B-miR-378-Runx1 regulatory pathway. Our results highlight the translational potential of miRNA replacement therapy for inhibiting Runx1 in breast cancer.

Keywords: Breast cancer; Invasion; MMTV-PyMT; MiR-378; Migration; Runx1.

Fig. 1. Correlation of the expression of Runx1 targeting miRNAs with Runx1 expression during breast cancer progression in MMTV-PyMT transgenic mice

(A) Heat map depicting statistically significant mature miRNA expression changes from miRNA microarrays performed on three biological replicates of mammary gland at three time points at early, mid and late breast cancer progression in MMTV-PyMT mice. Red represents high and white represents low expression with degree of change indicated by intensity of color. All 31 miRNAs showed a p-value ≤ 0.0025 for differential expression over time and all miRNAs passed a False Discovery Rate < 0.05. Bolded miRNA names represent miRNAs predicted to target Runx1. (B) Table showing specific microarray signal values (Log₂), fold change over time and p-values for indicated Runx1-targeting miRNAs; red highlighting denotes miR-378 as the most significantly down-regulated miRNA during breast cancer progression. (C) Trend in miR-378 expression similar over time as determined by either qPCR (left y-axis) or microarray (right y-axis). (D) Runx1 gene expression was evaluated by qPCR in the same samples that microarray was performed on. Values were normalized to GAPDH and data is presented as fold change in Runx1 expression from early to late disease. (E) Model summarizing the demonstrated reciprocal relationship between Runx1 and Runx1-targeting miRNA expression during disease progression in the MMTV-PyMT transgenic mouse model of breast cancer. Data (C, D) are presented as mean ± SEM of each group.

Fig. 2. Inverse expression of Runx1 and miR-378 expression in human breast cancer cells

(A) Endogenous Runx1 mRNA expression in human breast cancer cell models representative of early (MCF7) and late stage metastatic (MDA-231) disease assessed by qPCR and (A, inset) Runx1 protein expression as demonstrated by representative by western blot. (B) Endogenous expression of miR-378 in MCF7 and MDA-231 cells examined by qPCR using U6 as the internal control. (C, D) Adapted UCSC genome browser view of H3K4me3 enrichment tracks for MCF7 (green) and MDA-231 (orange) cells; blue boxes highlight H3K4me3 enrichment at genomic loci of the three Runx1 transcripts for Runx1 (C), and at the miR-378A locus (D). Gene names are consistent with NCBI Reference Sequence records. Gene annotation follows standard display conventions used by the UCSC genome browser (exons, solid boxes; introns, solid lines; direction of transcription, arrows). Data (A, B) are presented as mean ± SD of each group. *P<0.05 (student’s t-test).

Fig. 3. Regulation of Runx1 by miR-378 in human breast cancer cells

(A) Schematic showing potential binding site for the human miR-378 in the 3′UTR of Runx1. The miR-378 seed sequence is highlighted in bold. (B) To determine if miR-378 could directly target the Runx1 3′UTR, miR-C or miR-378 mimic and the Runx1 3′UTR luciferase construct were transfected into HeLa cells and luciferase activity was normalized to co-transfected renilla luciferase, with data presented as relative light units produced. (C) Ectopic expression of miR-378 after transfection of MDA-231 cells with miR-378 mimic (or miR-C non-targeting control) confirmed by qPCR. (D) Runx1 mRNA expression in MDA-231 cells administered miR-C or miR-378 mimic. (E) Representative western blot (upper) showing endogenous Runx1 protein expression in MDA-231 cells after transfection with miR-C or miR-378 mimic and (lower) Runx1 protein quantitation (relative to GAPDH control). All data are presented as mean ± SD of each group. *P<0.05, **P<0.01 (student’s t-test).

Fig. 4. Effect of inhibition of Runx1 on expression of the miR-378 host gene PPARGC1B and pre-mir-378

(A) Adapted UCSC Genome Browser view of the three PPARGC1B transcripts, with promoter regions (+2kb from the transcription start site) highlighted by red boxes. Gene names are in line with NCBI Reference Sequence records. Gene annotation follows standard display conventions used by the UCSC genome browser (exons, solid boxes; introns, solid lines; direction of transcription, arrows). Zoomed insets highlight positions (solid bar, black) of Runx1 core binding motif sites (TGTGGT, red) within the promoter region of PPARGC1B. (B, C) Endogenous expression of each of the three isoforms of PPARGC1B (B) and of precursor miR-378 (pre-miR-378) (C) determined by qPCR in both MCF7 and MDA-231 cells. (D–F) Gene expression of Runx1 (D), PPARGC1B (E) and pre-miR-378 (F), as determined by qPCR in siNS- or siRunx1-transfected MDA-231 cells. (G) Representative western blot showing demonstrating changes in expression of Runx1 and PPARGC1B protein expression upon transfection of MDA-231 cells with siNS or siRunx1. GAPDH was used as the loading control. In all cases, data are shown relative to GAPDH and are presented as ± SD of each group. ND, not determined, *P<0.05, **P<0.01 (student’s t-test).

Fig. 5. Changes in human breast cancer cell phenotype upon modulation of miR-378-Runx1 regulatory pathway

MDA-231 cells treated with non-targeting control miRNA (+miR-C) or miR-378 mimic (+miR-378) were assayed as follows: (A) Representative phase contrast images (mag. 10×) of MDA-231 cells treated as above were subjected to a scratch assay for times indicated. The area of the scratch was plotted as a percentage of total area for three independent experiments carried out in duplicate. (B) Light microscopy images (mag.12×) of stained cells from a representative trans-well migration assay experiment with MDA-231 cells treated as above; inset images show an overview of all cells stained in the well at lower magnification (2.5×) (left). Quantitation of migrated cells assessed by measurement of the absorbance of solubilized crystal violet stain retained by migrated cells from two independent experiments carried out in triplicate (right). (C) Light microscopy images (mag.12×) of stained cells from a representative matrigel invasion assay experiment with MDA-231 cells treated as above; inset images show an overview of all cells stained in the well at lower magnification (2.5×) (left). Quantitation of invaded cells assessed by measurement of the absorbance of solubilized crystal violet stain retained by invaded cells from two independent experiments carried out in triplicate (right). (D) Cell proliferation assessed in MDA-231 cells treated as above for time points indicated. All quantitative data are depicted as mean ± SD per group. NS, no significant difference, *P<0.05, **P<0.01 (student’s t-test).

Fig. 6. Inhibition of miR-378 increases Runx1 and migration of MCF7 cells

MCF7 cells treated with control inhibitor (+miR-C) or hsa-miR-378a-3p hairpin inhibitor (+miR-378 inhibitor) were assayed as follows: (A) Representative phase contrast images (mag. 20×) of MCF7 cells subjected to a scratch assay for time points indicated. The area of the scratch was quantified using the MiToBo plug-in for ImageJ software and plotted as a percentage of total area for three independent experiments performed in triplicate. (B) Runx1 mRNA expression in MCF7 cells treated with miR-Control or miR-378 inhibitor. Data are shown relative to GAPDH and are presented as ± SD per group. (C) Schematic summarizing demonstrated miR-378 – Runx1 regulatory pathway in human breast cancer cells. Schematic key: Pointed arrowhead, promotes expression; Blunt end arrowhead, inhibits expression; Dashed line, assumed – not supported by experimental evidence in this system; Solid lines, proven – supported by experimental evidence in this system. *P<0.05, **P<0.01 (student’s t-test).