Dysregulation of the BRCA1/long non-coding RNA NEAT1 signaling axis contributes to breast tumorigenesis

Abstract

Dysregulation of long non-codng RNA (lncRNA) expression has been found to contribute to tumorigenesis. However, the roles of lncRNAs in BRCA1-related breast cancer remain largely unknown. In this study, we delineate the role of the novel BRCA1/lncRNA NEAT1 signaling axis in breast tumorigenesis. BRCA1 inhibits NEAT1 expression potentially through binding to its genomic binding site upstream of the NEAT1 gene. BRCA1 deficiency in human normal/cancerous breast cells and mouse mammary glands leads to NEAT1 overexpression. Our studies show that NEAT1 upregulation resulting from BRCA1 deficiency stimulates in vitro and in vivo breast tumorigenicity. We have further identified molecular mediators downstream of the BRCA1/NEAT1 axis. NEAT1 epigenetically silences miR-129-5p expression by promoting the DNA methylation of the CpG island in the miR-129 gene. Silencing of miR-129-5p expression by NEAT1 results in upregulation of WNT4 expression, a target of miR-129-5p, which leads to activation of oncogenic WNT signaling. Our functional studies indicate that this NEAT1/miR-129-5p/WNT4 axis contributes to the tumorigenic effects of BRCA1 deficiency. Finally our in silico expression correlation analysis suggests the existence of the BRCA1/NEAT1/miR-129-5p axis in breast cancer. Our findings, taken together, suggest that the dysregulation of the BRCA1/NEAT1/miR-129-5p/WNT4 signaling axis is involved in promoting breast tumorigenesis.

Keywords: BRCA1; NEAT1; WNT4; breast cancer stem cells; miR-129.

Figure 1. BRCA1 functions as an upstream regulator to inhibit the expression of the NEAT1 gene

(A) Expression analysis of NEAT1 in MCF10A, MCF10DCIS and HCC1937 cells. qRT-PCR results are shown in the left panel. Western blot analysis of BRCA1 protein levels in MCF10A and MCF10DCIS cells is shown in the right panel. β-Actin protein levels were used as a protein loading control. (B) BRCA1 overexpression downregulates NEAT1 expression. qRT-PCR analysis of NEAT1 expression was performed on MCF10A, MCF10DCIS and HCC1937 cells transfected with the empty vector or BRCA1 expression plasmid DNA. (C, D) BRCA1 knockdown leads to induction of NEAT1 expression. Western blot analysis of BRCA1 and β-Actin (left panels) and qRT-PCR analysis of NEAT1 expression (right panels) were performed on MCF10A (C) and MCF7 (D) cells transfected with either the control or BRCA1 siRNA. Two different BRCA1 siRNAs (siBRCA1-1 and siBRCA1-2) were used in the knockdown experiment. (E) Expression analysis of Neat1 RNA levels in wild-type and Brca1-deficient mammary glands. qRT-PCR analysis of Neat1 expression was performed on mammary glands from wild-type and Brca1-deficient (Brca1^Co/Co) mice. (F) Brca1 deficiency gives rise to elevated Neat1 RNA levels in ductal epithelial cells of mammary glands. In situ hybridization (ISH) analysis of Neat1 RNA expression was performed on mammary gland tissue sections from wild-type and Brca1^Co/Co mice. The scale bar indicates 50 μm. (G) BRCA1 protein binds its cognate binding site in vivo located in the upstream genomic region of the human NEAT1 gene. ChIP assays using the BRCA1 antibody or the mouse IgG control were performed on MCF10A cells transfected with the control empty vector or BRCA1 expression plasmid DNA. qPCR assays were performed on ChIP samples to quantitate five distinct immunoprecipitated genomic DNA regions (R1 to R5 as indicated in the map) upstream of the NEAT1 gene by using five different pairs of primers. The genomic map for the 5′-end of the NEAT1 gene and its upstream DNA region is shown in the top panel. The putative transcription factor binding sites are depicted in the map. The result of quantitative ChIP analysis is presented as a bar graph shown in the bottom panel. Ectopically expressed BRCA1 protein predominantly bound to the R2 region. An approximately two-fold increase in the binding to the R2 region in the empty-vector-transfected cell sample was derived from the binding of endogenous BRCA1 protein. For all of bar graphs presented here, the error bar represents the standard deviation (SD) of the dataset (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 2. NEAT1 is an oncogenic factor required for invasiveness, anchorage-independent growth and stemness of MCF10DCIS tumor cells

(A) The knockdown efficiency of two NEAT1 siRNAs. qRT-PCR analysis of NEAT1 expression was performed on MCF10DCIS cells transfected with the control or two distinct NEAT1 siRNAs. (B) NEAT1 knockdown suppresses the migratory ability of MCF10DCIS tumor cells. Transwell migration assays were performed on MCF10DCIS cells transfected with the control or two different NEAT1 siRNAs for 48 hours. The stained pictures are shown in the top panel and the quantitative bar graph data (n = 3) are shown in the bottom panel. The scale bar indicates 100 μm. (C) NEAT1 knockdown inhibits the invasive ability of MCF10DCIS tumor cells. Invasion assays were performed on MCF10DCIS cells transfected with the control or two different NEAT1 siRNAs for 48 hours. The stained pictures are shown in the top panel and the quantitative bar graph data (n = 3) are shown in the bottom panel. The scale bar indicates 100 μm. (D) NEAT1 knockdown inhibits the anchorage-independent growth of MCF10DCIS tumor cells. Soft agar assays were performed on MCF10DCIS cells transfected with the control or two different NEAT1 siRNAs for 48 hours. MCF10A cells were also included in assays to serve as a non-malignant cell control. The stained pictures are shown in the top panel and the quantitative bar graph data (n = 3) are shown in the bottom panel. The scale bar indicates 200 μm. (E) NEAT1 knockdown results in the decreased self-renewal and proliferation of BCSCs in MCF10DCIS cells. Stem-cell sphere formation assays were performed on MCF10DCIS cells transfected with the control or two NEAT1 siRNAs for 48 hours. Pictures of BCSC spheres are shown in the top panel and sphere formation efficiency data (n = 3) are shown in the bottom panel. The scale bar indicates 100 μm. (F) NEAT1 overexpression enhances the self-renewal of breast stem cells. Stem-cell sphere formation assays (the bottom-right panel) and qRT-PCR analysis of NEAT1 expression (the bottom-left panel) were performed on MCF10A cells transfected with the empty vector or NEAT1 expression plasmid DNA for 24 hours. Pictures of breast stem-cell spheres are shown in the top panel and the scale bar indicates 100 μm. The error bar in bar graphs represents the SD of the dataset (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 3. NEAT1 upregulation induced by BRCA1 deficiency promotes in vitro malignancies and in vivo tumorigenicity of breast tumor cells

(A) NEAT1 siRNA abolishes the upregulation of NEAT1 by BRCA1 knockdown. qRT-PCR analysis of NEAT1 expression was performed on MCF10DCIS cells transfected with the control siRNA (siControl), the siRNA targeting BRCA1 (siBRCA1-1) or NEAT1 (siNEAT1-2), or the siRNA combination targeting both genes for 48 hours. (B) Co-knockdown of NEAT1 partially abolishes the enhanced proliferation of BRCA1-knockdown MCF10DCIS cells. At 24 hrs after transfection, 10,000 siRNA-transfected MCF10DCIS cells shown in (A) were seeded for cell proliferation assays. Live cell counting was performed by trypan blue dye exclusion assays. (C) Co-knockdown of NEAT1 attenuates the expansion of BCSCs in BRCA1-knockdown MCF10DCIS cells. Flow cytometry analysis of surface antigen markers CD44, CD49f and CD24 for BCSC identification was performed on siRNA-transfected MCF10DCIS cells shown in (A). The 2D dot plots that profile CD44+CD49f+ and CD44+CD49f+CD24– cell subsets are shown in Supplementary Figure S7A. The percentages of BCSC-enriched CD44+CD49f+CD24– cell subsets from three independent flow cytometry experiments were used to make the bar graph. Error bars indicate standard deviation (SD). (D) BRCA1 knockdown significantly expands the EpCAM+BCSC population and co-knockdown of NEAT1 attenuates its expansion. Flow cytometry analysis of BCSC-related protein antigens CD44, CD49f, CD24 and EpCAM was performed on siRNA-transfected MCF10DCIS cells shown in (A). The gated CD44+CD49f+ cell subsets shown in Supplementary Figure S7A were subjected to the 2D dot plot analyses that profile CD24 and EpCAM as shown in Supplementary Figure S7B. The percentages of EpCAM+BCSC cell subsets (CD44+CD49f+CD24–EpCAM+) in siRNA-transfected MCF10DCIS cells from three independent flow cytometry experiments were used to make the bar graph. (E) Co-knockdown of NEAT1 impairs BRCA1-knockdown-induced increases in the size and formation efficiency of BCSC spheres developed from MCF10DCIS cells. At 48 hrs after transfection, siRNA-transfected MCF10DCIS cells shown in (A) were seeded in six-well plates for stem-cell sphere formation assays. Pictures of formed BCSC spheres are shown in the left panel and the bar graph of sphere formation efficiency data is shown in the right panel. The scale bar indicates 100 μm. (F) NEAT1 knockdown abolishes a BRCA1-deficiency-induced increase in invasiveness of MCF10DCIS tumor cells. 48 hours posttransfection, siRNA-transfected MCF10DCIS cells shown in (A) were subjected to invasion assays. The stained pictures are shown in the top panel and the quantitative bar graph of invasion data is shown in the bottom panel. The scale bar indicates 100 μm. (G) Co-knockdown of NEAT1 attenuates the increased anchorage-independent growth of MCF10DCIS tumor cells induced by BRCA1 knockdown. 48 hours posttransfection, soft agar assays were performed on siRNA-transfected MCF10DCIS cells shown in (A). The stained pictures are shown in the top panel and the quantitative bar graph of colony formation efficiency data is shown in the bottom panel. The scale bar indicates 200 μm. (H) Neat1 knockdown inhibits the in vivo development of Brca1-deficient xenograft mammary tumors. 1 × 10⁶ stable scramble shRNA-expressing or shNeat1-expressing tumor cells that were derived from Brca1-deficient mammary tumors developed in MMTV-cre;Brca1^co/co mice were injected into the fourth mammary fat pad of syngeneic female mice. The development of xenograft tumors was monitored for eight weeks and tumor size was measured weekly. Tumor growth curves (n = 6) were plotted and are shown in the right panel. The dissected tumors were photographed and are shown in the left panel. The knockdown efficiency of these two Neat1 shRNAs (shNeat1-1 and shNeat1-2) is shown in Supplementary Figure S6A. The error bar shown in all bar graphs and the growth rate plot represents the SD of the dataset (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 4. Epigenetic regulation of miR-129-5p expression by the BRCA1/NEAT1 axis

(A) Expression profiling of miRNAs in MCF10A cells with NEAT1 knockdown by PCR array assays. The common logarithms of miRNA expression values from control siRNA-transfected cells were plotted against those from NEAT1 siRNA (siNEAT1-2)-transfected cells to make the scatter plot. miRNAs that were upregulated and downregulated at least 2-fold in NEAT1-knockdown MCF10A cells are indicated by the red and green colors, respectively. These identified NEAT1-regulated miRNAs were confirmed by another NEAT1 siRNA (siNEAT1-1) (data not shown). (B) Downregulation of miR-129-5p expression by BRCA1 knockdown is abolished by co-knockdown of NEAT1. qRT-PCR analysis of miR-129-5p expression was performed on MCF10A and MCF10DCIS cells transfected with the control siRNA, the siRNA targeting BRCA1 or NEAT1, or the siRNA combination targeting both genes for 48 hours. (C) Ectopic expression of BRCA1 upregulates miR- 129- 5p expression. qRT-PCR analysis of miR-129-5p expression was performed on MCF10A and MCF10DCIS cells transfected with the empty vector or BRCA1 expression plasmid DNA. (D) NEAT1 knockdown leads to partial demethylation of the CpG island in the miR-129 gene. Bisulfite sequencing analysis of the DNA region (containing 32 CpG dinucleotides) within the CpG island of the miR-129 gene was performed on genomic DNA samples isolated from control and NEAT1 siRNA-transfected MCF10DCIS cells. Filled and unfilled circles represent methylated and unmethylated CpG dinucleotides, respectively. The methylation level bar graph with error bars was plotted based on sequencing results from five randomly selected clones for each sample. The error bar shown in sub-figures (B, C) represents the SD of the dataset (n = 3). **p < 0.01, ***p < 0.001.

Figure 5. The NEAT1/miR-129-5p axis mediates the effect of BRCA1 deficiency to enhance malignancies and stemness of breast tumor cells

(A) The miR-129-5p mimic attenuates enhanced proliferation of BRCA1-knockdown MCF10DCIS cells. MCF10DCIS cells were transfected with the control siRNA, BRCA1 siRNA, miR-129-5p mimic or BRCA1 siRNA plus miR- 129- 5p mimic. At 24 hrs after transfection, 10,000 transfected MCF10DCIS cells were seeded for cell proliferation assays as described in Figure 3B. (B) The miR-129-5p mimic abolishes the BRCA1-knockdown-induced increase of the BCSC cell population in MCF10DCIS cells. Flow cytometry analysis of BCSC protein antigens CD44, CD49f and CD24 was performed on transfected MCF10DCIS cells 72 hours posttransfection as indicated. The gating and analysis were performed as described in Figure 3C. The quantitative bar graph was plotted based on percentages of CD44+CD49f+CD24– cell subsets from three independent flow cytometry experiments. (C) The miR-129-5p mimic reduces the BRCA1-deficiency-induced expansion of the EpCAM+BCSC population in MCF10DCIS cells. Flow cytometry analysis of BCSC protein antigens CD44, CD49f, CD24 and EpCAM was performed on transfected MCF10DCIS cells 72 hours posttransfection as indicated. The gating and analysis were performed as described in Figure 3D. Quantitative analysis of the EpCAM+BCSC subset (CD44+CD49f+CD24–EpCAM+) for a representative flow cytometry experiment is also shown in Supplementary Figure S9. The quantitative bar graph for the measurement of the CD44+CD49f+CD24–EpCAM+ cell subset was plotted based on three independent experiments. (D) The miR-129-5p mimic impairs BRCA1-knockdown-induced increases in the size and formation efficiency of BCSC spheres developed from MCF10DCIS cells. At 48 hrs after transfection, 10,000 transfected MCF10DCIS cells as indicated were seeded for stem-cell sphere formation assays. The pictures of formed BCSC spheres are shown in the left panel and the BCSC sphere formation efficiency data are shown in the right panel. The scale bar in sphere pictures indicates 100 μm. (E) Inhibition of miR-129-5p enhances the stemness of MCF10A cells. Stem-cell sphere formation assays were performed on MCF10A cells transfected with either the control scramble or miR-129-5p inhibitor RNA. The mammosphere pictures are shown in the top panel and the quantitative bar graph of the mammosphere formation efficiency data is shown in the bottom panel. The scale bar indicates 100 μm. (F) The miR-129-5p mimic impairs the BRCA1-knockdown-induced enhancement of invasiveness of MCF10DCIS tumor cells. Transfected MCF10DCIS cells as indicated were subjected to invasion assays 48 hours posttransfection. The stained pictures are shown in the top panel and the quantitative bar graph of invasion data is shown in the bottom panel. The scale bar indicates 100 μm. (G) The miR-129-5p mimic attenuates a BRCA1-knockdown-induced increase in anchorage-independent growth of MCF10DCIS tumor cells. Transfected MCF10DCIS cells as indicated were subjected to soft agar assays 48 hours posttransfection. Colony formation efficiency data from three independent experiments were plotted into a bar graph. The error bar in bar graphs represents the SD of the dataset (n = 3). *p < 0.05, **p < 0.01.

Figure 6. WNT4 is a miR-129-5p target gene that is regulated by the BRCA1/NEAT1/miR-129-5p axis

(A) A map for the predicted miR-129-5p targeting site in the 3′-untranslated region of the WNT4 mRNA. A DNA fragment with mutations in the seeding site of WNT4 3′-UTR was used to construct the mutant reporter and its RNA sequence is shown under the map with its wild-type and miR-129- 5p sequences. (B) The miR-129-5p mimic inhibits the luciferase expression of the wild-type, but not the mutated WNT4 3′- UTR reporter. HEK-293T cells were transfected with the wild-type or mutated WNT4 3′-UTR reporter plasmid DNA along with either the control scramble dsRNA or the miR-129-5p mimic. All cell samples were also co-transfected with Renilla plasmid DNA, which was used as a transfection efficiency control. Dual Luciferase assays were performed on transfected cells 24 hrs posttransfection. The measured Luciferase activity values were normalized by Renilla activity values. The error bar represents the SD of the dataset (n = 3). **p < 0.01. (C) The miR-129-5p mimic downregulates WNT4 expression and WNT signaling activity. Western blot analysis of WNT4, β-catenin and β-actin was performed on scramble dsRNA-tansfected and miR-129-5p-transfected MCF10A and MCF10DCIS cells. (D) Inhibition of miR-129-5p in MCF10A cells leads to WNT4 upregulation and activation of WNT signaling. 48 hours posttransfection, Western blot analysis of WNT4, β-catenin and β-actin was performed on MCF10A cells transfected with either the scramble or miR-129-5p inhibitor RNA. (E) WNT4 is functionally required for endogenous WNT signaling activity. Western blot analysis of WNT4, β-catenin and β-actin was performed on MCF10A and MCF10DCIS cells transfected with the control siRNA or the WNT4 siRNA for 48 hrs. (F) Upregulation of WNT4 expression and activation of WNT signaling by BRCA1 knockdown are NEAT1-dependent. Western blot analysis of WNT4, β-catenin and β-actin was performed on MCF10A and MCF10DCIS cells transfected with the control siRNA, the siRNA targeting BRCA1 or NEAT1, or the siRNA combination targeting both genes for 48 hours. (G) Ectopic expression of BRCA1 downregulates WNT4 expression and suppresses WNT signaling activity. Western blot analysis of BRCA1, WNT4, β-catenin and β-actin was performed on MCF10A and MCF10DCIS cells transfected with the empty vector or BRCA1 expression plasmid DNA. (H) Ectopic expression of NEAT1 upregulates WNT4 expression and activates WNT signaling activity. Western blot analysis of WNT4, β-catenin and β-actin was performed on MCF10A and MCF10DCIS cells transfected with the empty vector or NEAT1 expression plasmid DNA. (I) Upregulation of WNT4 expression and activation of WNT signaling by BRCA1 knockdown are abolished by the miR-129-5p mimic. Western blot analysis of WNT4, β-catenin and β-actin was performed on MCF10A and MCF10DCIS cells transfected with the control siRNA, the BRCA1 siRNA or BRCA1 siRNA plus miR-129-5p mimic for 48 hours. β-actin was used as a protein loading control for all Western blot analyses shown here.

Figure 7. Upregulation of WNT4 by the NEAT1-miR129 axis is functionally implicated in promoting malignant phenotypes and stemness of BRCA1-deficient breast tumor cells

(A) WNT4 knockdown attenuates enhanced proliferation of BRCA1-knockdown MCF10DCIS. MCF10DCIS cells were transfected with the control siRNA, BRCA1 siRNA, WNT4 siRNA, or combined siRNAs targeting both BRCA1 and WNT4. At 24 hrs after transfection, 10,000 siRNA-transfected MCF10DCIS cells were seeded for cell proliferation assays as described in Figure 3B. (B) Co-knockdown of WNT4 abolishes the BRCA1-knockdown-induced activation of WNT signaling in MCF10DCIS cells. Western blot analysis of BRCA1, WNT4, β-catenin and β-actin was performed on siRNA-transfected MCF10DCIS cells 48 hours posttransfection as indicated. (C) Co-knockdown of WNT4 abolishes the BRCA1-knockdown-induced increase of the BCSC cell population in MCF10DCIS cells. Flow cytometry analysis of three BCSC protein markers (CD44, CD49f and CD24) was performed on siRNA-transfected MCF10DCIS cells 72 hours posttransfection as indicated. The quantitative bar graph for the measurement of the CD44+CD49f+CD24– cell subset in siRNA-transfected MCF10DCIS cells was plotted based on three independent experiments as described in Figure 3C. (D) Co-knockdown of WNT4 attenuates the BRCA1-knockdown-induced expansion of the EpCAM+BCSC population in MCF10DCIS cells. Flow cytometry analysis of four BCSC protein markers (CD44, CD49f, CD24 and EpCAM) was performed on siRNA-transfected MCF10DCIS cells 72 hours posttransfection as indicated. The quantitative bar graph for the measurement of the CD44+CD49f+CD24–EpCAM+ cell subset in siRNA-transfected MCF10DCIS cells was plotted based on three independent experiments as described in Figure 3D. Quantitative analysis of the EpCAM+BCSC subset (CD44+CD49f+CD24–EpCAM+) for a representative flow cytometry experiment is also shown in Supplementary Figure S10. (E) Co-knockdown of WNT4 impairs BRCA1-knockdown-induced increases in the size and formation efficiency of BCSC spheres generated from MCF10DCIS cells. At 48 hrs after transfection, 10,000 siRNA-transfected MCF10DCIS cells as indicated were seeded for stem-cell sphere formation assays. The pictures of formed spheres are shown in the left panel and the sphere formation efficiency data are shown in the right panel. (F) Treatment with the recombinant WNT4 cytokine induces the increase of BCSCs in MCF10DCIS cells. After treatment with recombinant WNT4 (50 nM) for 72 hrs, treated MCF10DCIS cells were subjected to the same flow cytometry analysis as described in (C). (G) Treatment with the recombinant WNT4 cytokine enhances the expansion of EpCAM+BCSCs in MCF10DCIS cells. Flow cytometry analysis of EpCAM+BCSCs in rWNT4-treated MCF10DCIS cells was performed as described in (D). (H) Treatment with the recombinant WNT4 cytokine enhances self-renewal of BCSCs in MCF10DCIS cells. After treatment with recombinant WNT4 (rWNT4, 50 nM) for 72 hrs, 10,000 treated MCF10DCIS cells were subjected to stem-cell sphere formation assays. During sphere formation, the sphere culture medium was supplemented with rWNT4 (50 nM). The pictures of formed BCSC spheres are shown in the top panel and the sphere formation efficiency data are shown in the bottom panel. (I) Co-knockdown of WNT4 abolishes the BRCA1-knockdown-enhanced invasiveness of MCF10DCIS tumor cells. SiRNA-transfected MCF10DCIS cells as indicated were subjected to invasion assays 48 hours posttransfection. The stained pictures are shown in the left panel and the quantitative bar graph of invasion data is shown in the right panel. (J) Co-knockdown of WNT4 moderately impairs the BRCA1-knockdown-promoted anchorage-independent growth of MCF10DCIS tumor cells. SiRNA-transfected MCF10DCIS cells as indicated were subjected to soft agar assays 48 hours posttransfection. Colony formation efficiency data from three independent experiments were plotted into a bar graph. The scale bar shown in sphere and invasion pictures (E, H, I) indicates 100 μm. The error bar represents the SD of the dataset (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 8. In silico expression correlation analysis of BRCA1, NEAT1 and miR-129-5p in a cohort of human breast cancer cell lines

(A) Regression analysis of the expression correlation between BRCA1 and NEAT1 in breast cancer cell lines. The left regression analysis plot was made according to BRCA1 and NEAT1 expression datasets from 38 BC lines. The right regression analysis plot was made according to the datasets of 35 BC lines after the data of three cell lines (SUM-190PT, SUM-225CWN, and T47D; indicated by red dots shown in the left plot) with a poor correlation were excluded from analysis. (B) Regression analysis of the expression correlation between NEAT1 and miR-129-5p in breast cancer cell lines. The left regression analysis plot was made according to NEAT1 and miR- 129-5p expression datasets from 35 BC lines that were narrowed down from analysis shown in the right panel of (A). The right regression analysis plot was made according to the datasets of 31 BC lines after the data of four cell lines (HCC70, MDA-MB-231, SUM-185PE, and SUM-52PE; indicated by red dots shown in the left plot) with a poor correlation were excluded from analysis. (C) Regression analysis of the expression correlation between BRCA1 and miR-129-5p in breast cancer cell lines. The left regression analysis plot was made according to BRCA1 and miR-129-5p expression datasets from 31 BC lines that were narrowed down from analysis shown in the right panel of (B). The right regression analysis plot was made according to the datasets of 29 BC lines after the data of two cell lines (BT- 483 and HCC1143; indicated by red dots shown in the left plot) with a poor correlation were excluded from analysis. All of expression correlation analyses shown here were based on expression data values of Supplementary Table S1.

Figure 9. The model for the role of the BRCA1/NEAT1/miR-129-5p/WNT4 signaling axis in BRCA1-deficiency-driven breast tumorigenesis