Decreased RBM7 expression was associated with poor prognosis of breast cancer.

(A-B) The correlation between RBM7 mRNA expression and overall survival (OS) or disease-free-survival (DFS) of breast cancer patients (n=1980) was analyzed based on the METABRIC dataset. The samples were divided into four equal parts, containing lower quartile Q1, median quartile Q2, upper quartile Q3 and higher quartile Q4 according to the expression of RBM7. Data are presented as means ± SD and P values were obtained by Mantel-Cox log-rank test. (C) The expression of RBM7 in primary breast carcinoma (n=1097) compared to normal tissues (n=114) was analyzed through UALCAN dataset. (D) Analysis of RBM7 expression in breast cancer lymph node metastases in comparison with breast cancer tissues with no lymph node metastasis based on TCGA dataset. BRCA samples were classified into N0 (No regional lymph node involvement) (n=515), and metastases in lymph node (N1-N3) (n=565). (E) Representative IHC images of RBM7 staining for patients at high (n=32) or low (n=58) clinical stages on tissue microarray of breast cancer specimens. Scale bars = 300 μm (top) or 30 μm (bottom). (F) Representative images of the IHC staining of RBM7 in a tissue microarray containing triple-negative breast cancer (n=119) and para-carcinoma tissues (n=20). Scale bars = 300 μm (top) and 30 μm (bottom). (G) Quantitative analysis of RBM7 expression according to IHC staining scores in primary breast cancer (n=12) and distant metastases (breast cancer lung metastases, n=4; breast cancer liver metastases, n=5). (H-I) Representative IHC images and quantitative analyses for RBM7 staining in 3 paired primary breast cancer and lymphatic metastases are shown. Scale bar = 100 μm. Data are presented as means ± SD and P values were obtained by unpaired Student t test (C-I). **P < 0.01, ***P < 0.001, ****P < 0.0001.

RBM7 negatively regulated breast cancer metastatic potential. (A) Gene Ontology analysis showed the significantly affected biological process upon RBM7 knockdown in MDA-MB-231 cells. (B) Functional association network of RBM7-regulated targets. The genes were analyzed using the STRING database, and subgroups are marked according to their function. (C) A heatmap showing the qRT-PCR analysis of differentially expressed genes upon RBM7 knockdown in breast cancer cells from three biological replicates. (D-F) 4T1 cells without or with RBM7 knockdown were injected into tail vein of immunodeficient BALB/c mice to establish a lung metastasis model. Spontaneous pulmonary metastases were assessed after about 3 weeks, the white arrow indicates lung lesions macroscopically (D). H&E stained lung sections were quantified for the number of spontaneous metastatic lesions from BALB/c mice (n=6, 5 and 9). Data are presented as mean ± SD and P values were determined by one-way ANOVA with Dunnett’s multiple comparisons test (E). Representative images of H&E stained lung sections are shown. Top panel scale bars indicate 7 mm; bottom panel scale bars indicate 300 μm (F) . (G-J) The metastatic ability of breast cancer cells with RBM7 depletion (G) or ectopic expression of RBM7 (I) were evaluated by transwell assay. Scale bars: 100 μm. Data are mean ± SD from three random fields. P values were determined by one-way ANOVA with Dunnett’s multiple comparisons test (H) or unpaired Student’s T test (J). (K) Representative images of the tube formation of HUVEC cells treated with conditional medium from RBM7-KD cells or control cells for 12h. Scale bars: 100 μm. Quantification of number of junctions formed by HUVEC was calculated by Image J software. **P < 0.01, ***P < 0.001, ****P < 0.0001.

Global identification of alternative splicing events regulated by RBM7.

(A-E) RNA-seq analysis was performed on RBM7-knockdown MDA-MB-231 cells or control cells, and the changes in splicing events were analyzed. (A) Volcano plot illustrating up-regulated/down-regulated alternative splicing events upon RBM7 depletion. (B) Comparison of the differential splicing events in five types of AS events affected by RBM7 depletion. (C) The relative fraction of splicing events undergoing percent spliced in (PSI) increase or decrease induced by RBM7 depletion. (D) Alternative splicing of MFGE8 was chosen to represent a decrease of PSI, and numbers of exon junction reads were indicated. (E) The hit top enriched motifs identified in the differential spliced genes regulated by RBM7. (F) RBM7-regulated exon skipping events were identified by semiquantitative RT-PCR using RBM7-depleted or control MDA-MB-231 cells. The mean ± SD of PSI from 3 independent experiments was plotted, with P values determined by one-way ANOVA with Dunnett’s multiple comparisons test. (G) Gene Ontology analysis showed the significantly enriched biological processes of changed splicing events affected upon RBM7 knockdown in MDA-MB-231 cells. (H) KEGG pathway enrichment analysis of functional association network of the RBM7-controlled AS targets.

RBM7 knockdown promoted exon 7 skipping of MFGE8.

(A) Comparison of RBM7-binding candidates from the RNA immunoprecipitation (RIP) sequencing (GSE144075 dataset) and two gene sets containing top 43 differentially AS genes in RNA-seq data of breast cancer cells expressing RBM7 shRNA1/2 as presented by venn diagram. The gene lists on the right were shown according to the FDR P value of top genes with significantly changed AS in RNA-seq. (B) Upper: schematics of human MFGE8 pre-mRNA (NM_005928.4). MFGE8-long isoform included the exon 7 (hereafter referred to as MFGE8-L), whereas it is skipped in MFGE8-short isoform (hereafter referred to as MFGE8-S). The arrow indicates the direction of transcription. Lower: primers were designed on exon 6/8, and RT-PCR was performed to identify PSI changes using RBM7-knockdown or control MDA-MB-231 cells. The base peak diagram of sanger sequencing of RT-PCR results showed the splicing junction sites (filled with gray). (C) Various breast cancer cells lines with stable depletion of RBM7 or control were constructed. Alternative splicing events of MFGE8 regulated by RBM7 were examined by semi-quantitative RT-PCR. The knockdown efficiency of RBM7 expression was detected by Western Blotting. The mean ± SD of PSIs from three independent repeated experiments were plotted with P value calculated by one-way ANOVA with Dunnett’s multiple comparisons test. (D) Left, Schematics of MFGE8 mini-splicing reporter including exons 6-8 and two flanking introns in the 600bp region upstream and downstream of 5’/3’ splice sites; middle, RBM7 stable knockdown or control cells were transfected with splicing reporter and collected for RNA extraction after 48h. The splicing changes of exogenous MFGE8 were detected by RT-PCR with specific primers on the reporter gene. The level of RBM7 protein was detected in MCF7 cells by western blotting. P value was calculated by one-way ANOVA with Dunnett’s multiple comparisons test from three repeated experiments. (E) As shown in schematic diagram, orange box indicates cassette exon 7 and primers were designed in the putative binding regions P1 and P2. Binding of MFGE8 pre-mRNA with RBM7 was examined by RIP assay in HEK293T cells expressing Flag-RBM7. (F) Upper: the red line in diagram indicates ASOs targeting region which contain UUUCUU residues; down: MCF7 and BT-549 cells were transfected with ASOs targeting MFGE8 pre-mRNA for 48h and then applied for RT-PCR identification. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

MFGE8-L inhibited breast cancer cell migration and invasion.

(A) The function of MFGE8-L/S on breast cancer cell migration and invasion were evaluated by transwell assays. (B) Gelatin degradation assay was performed to test the effect of RBM7 knockdown on invadopodia function. 10000 cells were plated onto FITC-gelatin substrates (Green) and cultured for 48 h. Representative images are shown (red, Cy3-phalloidin; blue, DAPI) and the degraded areas were quantified by Image J software. Scar bars= 50 μm. P values were determined by one-way ANOVA with Tukey’s multiple comparisons test (n = 3). (C) The subcellular localizations of MFGE8 isoforms. MDA-MB-231 cells were transfected with Flag-tagged MFGE8-L or MFGE8-S, and visualized with immunofluorescence assay. Scale bar = 10 μm. The percent of cells with foci were quantified and plotted (about 80 cells were captured and quantified in both samples). (D-E) The KEGG pathway and Gene ontology analysis of genes up-regulated by MFGE-S compared with MFGE8-L. (F) The expression levels of p-STAT1(Tyr701) in HCC1937 cells overexpressing MFGE8-L/S was detected by western blotting. P values were determined by one-way ANOVA with Tukey’s multiple comparisons test. **P < 0.01, ****P < 0.0001.

RBM7 knockdown enhanced aggressiveness of breast cancer relying on MFGE8 splicing switch to the short variant and NF-κB pathway activation. (A) Breast cancer cells MDA-MB-231 and MCF7 were cotransfected with shRNA targeting RBM7 or control and shRNA-resistant Flag-MFGE8-L vectors and then subjected to transwell migration assay or invasion assay. Migrated or invaded cells were counted from random sites of the transwell with Image J. P values were determined by one-way ANOVA with Tukey’s multiple comparison test. (B) Representative transwell analysis of migrative/invasive capability of breast cancer cells transfected with 500 nM ASO directed against RBM7-binding region in MFGE8 pre-mRNA. Cells migrating through transwell membrane with matrigel or not were imaged and representative images are shown (upper) along with quantification (down). P values were determined by one-way ANOVA with Tukey’s multiple comparison test. (C) The expression of metastasis-related factors in HCC1937 and BT-549 cells stably expressing MFGE8-L, MFGE8-S or control vector was measured by real-time PCR assay. (D) NF-κB signaling pathway was enrichment via KEGG analysis in RBM7 knockdown breast cancer cells. (E) Western blotting showed the expression of p65, p-p65, IκBα and p-IκBα in NF-κB pathway in RBM7-depleted or control MDA-MB-231 and BT-549 breast cancer cells. (F) Western blotting was performed to test the expression of RBM7, p65 and p-p65 in RBM7-depleted MDA-MB-231 cells treated with or without NF-κB inhibitor PDTC 10 μm for 48h. (G) Representative images of the tube formation of HUVEC cells treated with conditional medium from RBM7-depleted MDA-MB-231 in the presence or absence of 10 μm NF-κB inhibitor PDTC. Scale bars: 100 μm. Quantification of junctions of endothelial network was conducted by ImageJ software. P values were determined by one-way ANOVA with Tukey’s multiple comparison test (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Splicing shift of MFGE8 toward exon7 exclusion was negatively correlated with RBM7 expression in patients with breast cancer.

(A) MFGE8 exon7 expression level was analyzed in breast carcinoma in comparison with normal tissues based on the TCGA dataset. Plotted are the mean ± SD from 114 normal tissues and 1094 tumor tissues, with **** P<0.0001 as determined by unpaired Student’ t test. (B) The splicing alteration of MFGE8 in 6 pairs of primary breast cancer tissues and adjacent normal tissues was examined using RT-PCR.The quantification of PSI vales was based on relative band intensities using Image J software. (C) The splicing alteration of MFGE8 in primary breast cancer tissues and corresponding lymph node metastases was identified by RT-PCR assays. The quantification of PSI vales was determined by Image J software. (D) The correlation between the expression of MFGE8 exon7 and overall survival (OS) of breast cancer patients was analyzed based on the TCGA dataset. P value was assessed with Mantel-Cox log-rank test. (E) Correlation of RBM7 expression with MFGE8 exon7 levels was analyzed using Pearson’s correlation coefficient method. (F) The mechanistic model of how RBM7 regulated metastasis of breast cancer through regulating MFGE8 splicing switch.