Establishment of a zebrafish xenograft model for analyzing BCBM.

(A) Confocal image illustrates the visualization of the circulatory system in Tg(kdrl:mCherry) zebrafish at 4 dpf and schematic representation illustrates the zebrafish xenograft model using MDA-MB-231 (GFP⁺) cells. Scale bar: 100 μm. (B) Top: Following the implantation of MDA-MB-231 (GFP⁺) cells into zebrafish, representative confocal microscopy images show the localization of MDA-MB-231 cells within various blood vessels. MDA-MB-231 cells are displayed in green, and the vasculature is shown in red. Scale bar: 100 μm. Bottom: A pie chart illustrates the distribution ratio of MDA-MB-231 cells across distinct blood vessels, based on an analysis of n = 45 zebrafish. (C, D) Three-dimensional imaging (scale bar: 100 μm) and reconstruction (scale bar: 50 μm) were performed to visualize the interaction between transplanted MDA-MB-231 cells and the PCeV in zebrafish. The MDA-MB-231 cells are shown in green, and the vasculature is depicted in red. Scale bar: 100 μm. (E) Representative confocal microscopy images of larvae injected with MDA-MB-231 cells at 24, 36, 48, 60, and 72 hpi. MDA-MB-231 cells are displayed in green, and the vasculature is shown in red. The white arrow heads indicated the disseminated cancer cells. Scale bar: 100 μm. PHBC primordial hindbrain channel, PMBC primordial midbrain channel, DCV dorsal ciliary vein, MsV mesencephalic vein. Source data are available online for this figure.

LRP8 is highly expressed in TNBC and predicts a high risk of brain metastasis.

(A) A Venn diagram illustrates the co-expressed genes identified in two datasets (GSE100534 and GSE52604, P < 0.01, logFC ≥ 1.5). (B, C) Heatmaps visualize 26 co-expressed genes from datasets GSE100534 (BTT vs. BCBMT) and GSE52604 (NNBT vs. BCBMT), organized according to the HR of DMFS derived from Kaplan–Meier analyses. (D, E) Kaplan–Meier survival curves depict DMFS and OS probabilities of BC patients, stratified by high and low expression levels of LRP8. (F) Box plots illustrate the expression levels of LRP8 across breast cancer patients and normal individuals. Normal (minimum: 0.127, Q1: 0.858, median: 1.085, Q3: 1.354, maximum: 2.207), BC (minimum: 0.126, Q1: 1.209, median: 2.267, Q3: 4.518, maximum: 11.792). P = 1 × 10⁻¹². (G) Box plots illustrate the expression levels of LRP8 across various breast cancer subtypes and normal individuals. Normal (minimum: 0.127, Q1: 0.858, median: 1.085, Q3: 1.354, maximum: 2.207), luminal (minimum: 0.126, Q1: 1.045, median: 1.765, Q3: 3.269, maximum: 7.787), HER2 + (minimum: 0.497, Q1: 3.551, median: 5.129, Q3: 10.331, maximum: 20.231), triple negative (minimum: 0.404, Q1: 4.661, median: 7.863, Q3: 10.133, maximum: 19.513). P = 2.3 × 10⁻¹⁶ (normal vs. Luminal), P = 6 × 10⁻⁸ (normal vs^. HER2 + ), P = 1.63 × 10⁻¹² (normal vs. triple negative). (H) Immunohistochemistry staining for LRP8 was performed on human TNBC tissues and paracancerous tissues. Scale bar: 100 μm. (I) IHC scores of (H) were presented. n = 10. (J) Western blot analysis was performed on MCF-10A, MCF-7, BT549, and MDA-MB-231 cell lines to evaluate LRP8 protein expression levels. (K) Quantitative analysis revealed the relative LRP8 protein level normalized to β-tubulin. n = 3. BTT breast tumor tissue, NNBT non-neoplastic breast tissue, BCBMT breast cancer brain metastasis tissue, DMFS distant metastasis-free survival, OS overall survival, HR hazard ratio. Data information: data are shown as mean ± SD. P values were analyzed with unpaired Student’s t test (F), Mann–Whitney test (I) and one-way ANOVA test (G, K). ns non-significant. Source data are available online for this figure.

LRP8 knockdown inhibits proliferation, migration and invasion of MDA-MB-231 cells.

(A) MDA-MB-231 cells transfected with sh-ctrl, sh-LRP8-1^# and sh-LRP8-2^# were subjected to western blot assay to analyze the protein level of LRP8. (B) Quantitative analysis showed the relative LRP8 protein level normalized to β-tubulin, n = 3. (C) Cell viability of MDA-MB-231 cells was evaluated by the CCK-8 assays. n = 3. P = 2.5 × 10⁻⁷ (sh-ctrl vs. sh-LRP8-1^#), P = 5 × 10⁻¹¹ (sh-ctrl vs. sh-LRP8-2^#). (D) Images of the number of colonies formed from MDA-MB-231 cells in three cell lines. Scale bar: 500 μm. (E) Images of GFP cancer cell proliferation in zebrafish after transplantation. Scale bar: 200 μm. (F) Following injection of MDA-MB-231 (GFP⁺) cells into zebrafish, whole-mount immunofluorescence staining for KI67 (red signal) and GFP was conducted on the embryos. Scale bar: 50 μm. (G) Transwell migration assay and Matrigel transwell invasion assay were performed in MDA-MB-231 cells transfected with sh-ctrl, sh-LRP8-1^# and sh-LRP8-2^#. Scale bar: 200 μm. (H) Quantification of the number of colonies formed from MDA-MB-231 cells in three cell lines. n = 3. (I) Quantitative analysis of E showed the relative fluorescence intensity of GFP in vivo. The measurement of fluorescence of GFP at 24 hpi was used as the baseline. n = 8. (J) Quantitative analysis of F showed the KI67 positive cell rate in two cell lines in vivo. n = 5. P = 7.3 × 10⁻⁵. (K) Quantitative analysis of G showed the migratory abilities in three cell lines of MDA-MB-231 cells. n = 3. P = 8.2 × 10⁻⁵ (sh-ctrl vs. sh-LRP8-1^#), P = 6.5 × 10⁻⁵ (sh-ctrl vs. sh-LRP8-2^#). (L) Quantitative analysis of G showed the invasive abilities in three cell lines of MDA-MB-231 cells. n = 3. P = 8.8 × 10⁻⁵ (sh-ctrl vs. sh-LRP8-1^#), P = 6 × 10⁻⁵ (sh-ctrl vs. sh-LRP8-2^#). (M) Real-time in vivo imaging was conducted at 24 hpi, 36 hpi, 48 hpi, 60 hpi and 72 hpi to monitor the dynamic behavior of MDA-MB-231 cells transfected with sh-ctrl and sh-LRP8-2^#. The white arrow heads indicated the disseminated cancer cells. n = 3. Scale bar: 100 μm. Data information: data are shown as mean ± SD, P values were analyzed with one-way ANOVA test (B, H, K, L), unpaired Student’s t test (I, J) and two-way ANOVA test (C). Source data are available online for this figure.

LRP8 knockdown decreases the GTP-CDC42 expression to inhibit filopodia formation.

(A) Top 20 of Reactome enrichment of downregulated genes in MDA-MB-231 cells transfected with sh-ctrl and sh-LRP8-2^#. FC >｜2｜and FDR < 0.05. (B) The heatmap visualization analyses demonstrating the significantly differentially expressed GEFs in MDA-MB-231 cells transfected with sh-ctrl and sh-LRP8-2^#. (C) The PPI network analysis conducted by STRING revealed the interaction of CDC42 and GEFs. (D) The classical Rho GTPases cycle involves both an inactive GDP-bound form and an active GTP-bound form. (E) Quantitative real-time PCR (qRT-PCR) assays were used to measure the mRNA levels of GEFs in MDA-MB-231 cells transfected with sh-ctrl and sh-LRP8-2^#. n = 3. P = 2 × 10⁻¹³(ARHGEF38), P = 1^.5 × 10⁻⁹ (ARHGEF37), P = 4.5 × 10⁻⁹ (ARHGEF16), P = 8.7 × 10⁻⁷ (FGD4). (F, G) MDA-MB-231 cells transfected with sh-ctrl and sh-LRP8-2^# were subjected to western blot assay to analyze the protein level of activated GTP-bound CDC42. n = 4. P = 1.2 × 10⁻⁵. (H) MDA-MB-231 cells transfected with sh-ctrl and sh-LRP8-2^# were stained with TRITC-labelled phalloidin (red) and DAPI (blue) to detect the formation of filopodia. The white arrow heads indicated the filopodia of MDA-MB-231 cells. Scale bar: 10 μm. (I, J) Quantification analysis of the number of filopodia per cell (n = 12, P = 4.3 × 10⁻⁷) and average filopodia length (n = 14, P = 8.8 × 10⁻¹⁰) of MDA-MB-231 cells transfected with sh-ctrl and sh-LRP8-2^#. FC: fold change. FDR^: false discovery rate. Data information: data are shown as mean ± SD, P values were analyzed with unpaired Student’s t test (G, I, J) and two-way ANOVA test (E). ns non-significant. Source data are available online for this figure.

Reelin-LRP8 signaling pathway may activate CDC42 to regulate migration and invasion of MDA-MB-231 cells.

(A) Transwell migration assay and Matrigel transwell invasion assay were performed in MDA-MB-231 cells transfected with sh-ctrl and sh-LRP8-2^# following Reelin treatment. Scale bar: 200 μm. (B, C) Quantity of migrated and invasive cells. n = 5. (D) MDA-MB-231 cells transfected with sh-ctrl and sh-LRP8-2^# were subjected to western blot assay to analyze the protein level of activated GTP-bound CDC42, following Reelin treatment. (E) Quantitative analysis revealed the relative GTP-bound CDC42 level normalized to total CDC42. n = 3. P = 4.9 × 10^-5 (sh-ctrl). (F, G) Images and quantification analysis of the migrated distance of brain metastatic cells along PceV in zebrafish models at 72 hpi after the anti-Reelin monoclonal antibody treatment. n = 9. Scale bar: 100 μm. (H) The schematic representation illustrates the isolation of brain metastatic cells after the transplantation of MDA-MB-231 (GFP⁺) cells into zebrafish. (I) The mRNA expression of LRP8 in various brain metastatic cell populations. n = 3. P = 0.0091 (BM0 vs. BM4), P = 0.0027 (BM0 vs. BM5), P = 3 × 10⁻⁵ (BM0 vs. BM6). (J) The expression of cell migration related genes from GO analysis. n = 86. BM0 (minimum: 0, Q1: 0.097, median: 1.377, Q3: 14.297, maximum: 871.831), BM6 (minimum: 0.161, Q1: 1.319, median: 5.963, Q3: 38.809, maximum: 2572.556). FC >｜2｜and FDR < 0.05. (K) The expression of cytoskeleton organization related genes from GO analysis. n = 87. BM0 (minimum: 0, Q1: 0.2095, median: 2.637, Q3: 14.2505, maximum: 282.875), BM6 (minimum: 0.327, Q1: 1.415, median: 7.716, Q3: 34.227, maximum: 1053.12). FC >｜2｜and FDR < 0.05. (L) Scatter plot of significant differentially expressed genes between BM0 vs. BM6. (M) Transwell migration assay and Matrigel transwell invasion assay were performed in BM0 cells and BM6 cells, Scale bar: 200 μm. (N, O) Quantification analysis of migrated and invaded BM0 or BM6 cells. n = 3. (P) Images of BM0 and BM6 cells migrating along PceV in zebrafish at 24 hpi. The white arrow heads indicated the disseminated cancer cells. Scale bar: 100 μm. (Q) Quantification analysis of migrated distance of brain metastatic cells along PceV in zebrafish. n = 9. P = 9.1 × 10⁻⁶. FC fold change, FDR false discovery rate. Data information: data are shown as mean ± SD, P values were analyzed with paired Student’s t test (J, K), unpaired Student’s t test (N, O, Q), two-way ANOVA test (B, C, E), Mann–Whitney test (G) and one-way ANOVA test (I). ns non-significant.  Source data are available online for this figure.

Onjisaponin B and MEN 10207 treatment inhibit the migration of MDA-MB-231 cells in vivo.

(A) Three-dimensional binding pattern diagram of two candidate compounds and LRP8 protein. (B, C) The root mean square deviation fluctuations and potential energy fluctuations of compounds with LRP8 protein. (D) Representative confocal microscopy images of zebrafish xenograft models treated with vehicle, OB and MEN 10207. The white arrow heads indicated the disseminated cancer cells. Scale bar: 100 μm. (E) Quantitative analysis of D showed the relative fluorescence intensity of GFP in vivo. The measurement of fluorescence of GFP at 24 hpi was used as the baseline. n = 10. P = 0.0012 (Vehicle vs. OB), P = 0.0003 (Vehicle vs. MEN 10207). (F) Quantification analysis of the migrated distance of brain metastatic cells along PceV in zebrafish at 72 hpi was conducted. n = 10. P = 1.1 × 10⁻⁶ (Vehicle vs. OB), P = 8.6 × 10⁻⁷ (Vehicle vs. MEN 10207). (G) Schematic representation illustrates the nude mice xenograft model using MDA-MB-231 (GFP⁺) cells. (H) Representative light-sheet fluorescence microscopy images of whole brains of nude mice xenograft models treated with vehicle, and MEN 10207. Scale bar: 1000 μm. (I) Quantitative analysis of H showed the brain metastatic tumor volume in vivo. n = 3. (J) Immunofluorescence staining (scale bar: 100 μm) for CD31 (red signal) and DAPI (blue signal) was conducted on mice brain section and three-dimensional imaging (scale bar: 50 μm) was performed to visualize the interaction between MDA-MB-231 cells and the blood vessels in mouse brain. MDA-MB-231 cells are displayed in green. The white arrow heads indicated the disseminated cancer cells in the mice brain. (K) MDA-MB-231 cells treated with vehicle and MEN 10207 were subjected to western blot assay to analyze the protein level of activated GTP-bound CDC42, following Reelin treatment. (L) Quantitative analysis revealed the relative GTP-bound CDC42 level normalized to total CDC42 protein. n = 5. Data information: data are shown as mean ± SD, P values were analyzed with unpaired Student’s t test (I, vehicle-treated group vs. MEN 10207-treated group) and one-way ANOVA test (E, F, L). ns: non-significant. Source data are available online for this figure.