The expression and relationship of ZEB2 and ACSL4 in breast cancer. A, Heat maps of the 38 up-regulated EMT-related genes detected by RNA-seq analysis in the paclitaxel-resistant MCF-7 cell line (TAXOL) and Epirubicin-resistant MCF-7 cell line (EPI) compared to wild-type MCF-7 cell line. B, The correlation between ACSL4 and ZEB2 mRNA expression in the TCGA cohort consisting of 1222 breast cancer patient samples. Spearman correlation and linear regression analysis were employed. C, The correlation between ACSL4 and ER mRNA expression in the TCGA cohort consisting of 666 breast cancer patient samples. Spearman correlation and linear regression analysis were employed. D, OS (Overall survival) was examined by Kaplan–Meier analysis to compare the survival rates in ACSL high and low expression of breast cancer patients. E, OS (Overall -progression survival) as examined by Kaplan–Meier analysis to compare the survival rates in ZEB2 high and low expression of breast cancer patients. F OS (Overall -progression survival) was examined by Kaplan–Meier analysis to compare the survival rates in the four groups of breast cancer patients. G, Expression of ACSL4 and ZEB2 were analyzed by western blot in a panel of 5 breast cancer cell lines, including two basal-like (MAD-231, BT549), two luminal (T47D, MCF-7) and a taxol-resistant cell lines. H, The correlation between ACSL4 and ZEB2 protein expression in the IHC cohort consisting of 45 breast cancer patient samples. I, The correlation between ACSL4 and ER protein expression in the IHC cohort consisting of 45 breast cancer patient samples. J, HE staining and IHC analysis of ACSL4, ZEB2, ER expression in representative basal-like and luminal subtype breast cancer tissues. Representative pictures were shown. Scale bar, 50 μm.

Over-expression of ACSL4 contributes to ZEB2-mediated breast cancer invasion. A, Cell metastatic capacity was analyzed by wound healing assay in control or ACSL4 over-expression MCF-7 cells (Left panel). Expression of ACSL4 was analyzed by western blot in control or ACSL4 over-expression MCF-7 cells. Quantification of relative migration distance (Right panel). Scale bar, 5mm. B, Cell invasive capacity was analyzed by transwell invasion assay in control or ACSL4 over-expression MCF-7 cells (Left panel). Quantification of relative invasive cells (Right panel). Scale bar, 1 mm. C, Cell metastatic capacity was analyzed by wound healing assays in control or ACSL4 silencing MDA-MD-231 cells (shACSL4) (Left panel). Expression of ACSL4 was analyzed by western blot in control or ACSL4 silencing MDA-MD-231 cells (shACSL4). Quantification of relative migration distance (Left panel). Scale bar, 5mm. D, Cell invasive capacity was analyzed by transwell invasion assay in control or ACSL4 silencing MDA-MD-231 cells (shACSL4) (Right panel). Quantification of relative invasive cells (Left panel). Scale bar, 1 mm. E, KEGG pathway enrichment analysis of differentially expressed genes by RNA-sequencing between control and ACSL4 knockdown MDA-MB-231 cells. The top 10 deferential pathways were listed. F, Expression of three EMT-related genes, E-cadherin, N-cadherin, and vimentin, was analyzed by western blotting in control, ACSL4, or ZEB2 silencing MDA-MB-231 cells. G, Cell invasive and metastatic capacity were analyzed by transwell invasion assay and wound healing assays (Scale bar, 1mm/5mm) in control and ZEB2 knockdown, or ZEB2 knockdown cells that over-express ACSL4. H, Protein expression was analyzed by western blot in control, ZEB2 knockdown (shZEB2), and ZEB2 knockdown with ACSL4 overexpression (shZEB2+ACSL4) cells. I. Quantification of relative invasive cells in G. J, Quantification of relative migration distance in G. Graphs indicated the Statistical analysis in G analyzed by Student’s t-test (mean ± SEM). ** P<0.01, *** P<0.001, **** P<0.0001. All results are from three or four independent experiments.

ACSL4 promotes lipid droplets accumulation and lipogenesis. A, BODIPY 493/503 staining of lipid droplets in MCF-7, Taxol-resistant MCF-7 cells (TAXOL-7), or MDA-MB-231 cells (Left panel). Quantification of normalized lipid contents (Right panel). Scale bar, 5μm. B, BODIPY 493/503 staining of lipid droplets in control or ACSL4 knockdown MDA-MB-231 cells (Upper panel). Quantification of normalized fluorescence intensity and percentage of LD-containing cell number (Lower panel). Scale bar, 50μm. C, BODIPY 493/503 staining of lipid droplets in control or ZEB2 knockdown MDA-MB-231 cells (Upper panel). Quantification of normalized fluorescence intensity and percentage of LD-containing cell number (Lower panel). Scale bar, 50μm. D, The heatmap of representative down-regulated lipid species (TG, PE, PC) with hierarchical clustering in the control cells and ACSL4 knockdown cells. Each species was normalized to the corresponding mean value, as determined by two-way ANOVA. E, Quantification of different Fatty Acids containing TG, PC, and PE species in control or ACSL4 knockdown MDA-MD-231 cells. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001. Error bars, SD.

Exogenous lipids promote LD accumulation and fuel cell migration. A, BODIPY 493/503 staining of lipid droplets in control cells or oleic acid (OA) loaded (200 µM) MDA-MB-231 cells. Scale bar, 10 μm. Quantification of normalized lipid contents from the conditions in (A). B, Cell invasive capacity was analyzed by transwell invasion assay in control or oleic acid (OA) loaded (200 µM) MDA-MB-231 cells. Cells invaded for 16 hours through a Matrigel-coated filter toward high-serum media (Left panel). Quantification of relative invasive cells (Right panel). Scale bar, 1 mm. C, Cell metastatic capacity was analyzed by wound healing assays in control or oleic acid (OA) loaded (200 µM) MDA-MB-231 cells. Scale bar, 5 mm. D, Cell invasive capacity and metastatic capacity were analyzed by transwell invasion assay and wound healing assays in control, shACSL4 and oleic acid (OA) loaded ACSL4 knockdown MDA-MB-231 cells. Quantification of relative invasive cells and migrated cells. Scale bar, 1 mm /5 mm. E, BODIPY 493/503 staining of lipid droplets in the cells at the leading edge of the scratch and the cells that away from the edge. The 2.5 D figure of the cell was shown. Quantitation of total LD area per cell. Scale bar, 10 µm. F, Cells were seeded in a transwell chamber. After 24 hours, cells migrated to the lower side of the chamber, and the fluorescence intensity per cell was calculated (Left panel). Quantitation of total LD area per cell before and after cell migration (Right panel). Scale bar, 10 µm (small) and 50 µm (big). * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001.

ACSL4 and ZEB2 increase fatty acid oxygen consumption and promote ATP generation in BLBC Cells. A, Reactome pathway analysis of differentially expressed genes by RNA-sequencing between control and ACSL4 knockdown MDA-MB-231 cells. The representative pathways were shown. B, Heatmap showing the differentially expressed genes between control cells and ACSL4 knockdown cells. The representative FAO-related genes were shown. C, The mRNA levels of CPT1A, CPT1B, and CPT1C were analyzed by quantitative PCR in control and ACSL4 knockdown cells. D, Quantitation of the normalized oxygen consumption rate (OCR) for long-chain fatty acids was monitored by Agilent XF Substrate Oxidation Stress Test in control or ACSL4 knockdown cells. Specific inhibitors were added as indicated. E, Quantitation of the normalized oxygen consumption rate (OCR) for long-chain fatty acids was monitored by Agilent XF Substrate Oxidation Stress Test in control or ZEB2 knockdown cells. Specific inhibitors were added as indicated. F and G, ATP production was quantified in control or ACSL4 knockdown cells (Left) or ZEB2 knockdown cells (Right). H and I, Lactate production was examined in control and ACSL4 knockdown (shACSL4) or ZEB2 knockdown (shZEB2) cells. Data are represented as mean SEM of three independent experiments, analyzed by Student’s t-test, * P<0.05, **** P<0.0001.

ZEB2 activates ACSL4 expression by directly binding to its promoter. A, Relative mRNA levels of ZEB2 and ACSL4 in control or ZEB2 knockdown MDA-MB-231 cells. B, Protein levels of ZEB2 and ACSL4 in control or ZEB2 knockdown MDA-MB-231 cells. C and D, Truncated ACSL4 promoter segment activity analyzed by Luciferase reporter assay in control or ZEB2 over-expressed 293T cells. E, The specific primers designed for ACSL4 promoter according to the E-box position shown in C (set1 to −287bp, set2 to −995, set3 to −1038, set4 to −1116). F, ChIP assay analysis of the occupation of ZEB2 on ACSL4 promoter by using four primers as indicated in E. G, Quantitative PCR analysis of ZEB2-binding abundance of specific ACSL4 promoter region by using set1 primer indicated in F. Genomic DNA was purified after chromatin immunoprecipitation and analyzed by Quantitative PCR. H, Luciferase reporter analysis of the activity of wild-type ACSL4 promoter or its mutants in control or ZEB2 over-expression 293T cells. Data are represented as mean SD of three independent experiments, analyzed by Student’s t-test, ** P<0.01, *** P<0.001, **** P<0.0001, ns: no significance.

ACSL4 regulates ZEB2 mRNA expression and protein stability. A, Relative mRNA levels of ZEB2 and ACSL4 in control or ACSL4 silencing MDA-MB-231 cells. B, Protein levels of ZEB2 and ACSL4 in control or ACSL4 silencing MDA-MB-231 cells as indicated. C, Ubiquitylation of ZEB2 was examined by in vitro ubiquitin assay. 293T cells were co-transfected with indicated constructs. Cells were treated with MG132 for 6 hours before IP. Anti-MYC was used to pull down the ZEB2 protein. The polyubiquitinated ZEB2 protein was detected by an anti-HA antibody. D, Ubiquitylation of ZEB2 was examined in MDA-MB-231 cells. The indicated antibody was used to pull down the protein in control or ACSL4 knockdown MDA-MB-231 cells. The polyubiquitinated ZEB2 protein was detected by anti-ubiquitin antibody. E, The interaction between ACSL4 and ZEB2 was detected by Co-IP assay. 293T cells were co-transfected with ZEB2 and ACSL4 expressing construct.The indicated antibody was used to pull down the protein. F, The interaction between ACSL4 and ZEB2 was detected by IP assay in MDA-MB-231 cells. Anti-ZEB2 antibody was used to pull down the protein. G, The stability of ZEB2 protein was detected by CHX treatment assay in control or ACSL4 silencing MDA-MB-231 cells. Cells were treated with 100 µg/mL cycloheximide (CHX) and were harvested at the indicated times after the addition of CHX. GAPDH was used as the internal loading control. H, Quantification of stability assays shown in G. I, The stability of ZEB2 protein was detected by CHX treatment assay in control or ACSL4 over-expressed MCF-7 cells. GAPDH was used as the internal loading control. J, Quantification of stability assays shown in I. *** P<0.001.

ACSL4 knockdown attenuates lung metastasis of breast cancer. A, Cell metastatic capacity was determined by xenograft experiment in vivo. Control cells and ACSL4 knockdown MDA-MB-231 cells were mixed with Matrigel in a 1:1 ratio and injected into NSG mice. Lung metastatic burden in mice was quantified by bioluminescence imaging at the experimental endpoint (Left panel). Quantification of normalized gray area (Right panel). B, Representative xenograft tumor images and HE staining pictures of metastatic nodules in lungs (pointed by black arrows) were shown. The right panel shows the quantification of metastatic lung nodules in two groups as indicated. C, The image of xenograft tumors developed from the control and ACSL4 knockdown cells. D, The tumor volumes were measured and calculated at the indicated time. E, Representative IHC images of lung metastatic lesions (Left panel) and quantification is shown (Right panel). Scale. Bar, 50μm. F, The representative images of fluorescence assay for ACSL4 and lipid droplets (Left panel). The fluorescence intensity of lipid droplets and ACSL4 were calculated in the right panel. Scale. Bar, 100μm. G, A proposed mechanism to illustrate the positive feedback loop of ZEB2 and ACSL4 regulates lipid metabolism, which results in enhanced breast cancer metastasis. Data are represented as mean SEM of three independent experiments and analyzed by Student’s t-test, * P<0.05, *** P<0.001, **** P<0.0001.