Article Figures & Data
Figures
- Figure 1.
Functional and molecular characterization of breast cancer cells derived from the long-term lung metastatic nodules in an orthotopic xenograft mouse model. A, The representative images of the cell tracking assay. The result showed that LC subline displayed reduced motility as compared with its isogenic 231-P and IV2 lines. B, Analysis of cell movement in three categories (distance of migration, distance to origin, and velocity) generated from the cell tracking assay showed that LC was inferior to its isogenic lines in all three criteria. Data are mean ± SEM (n = 5). *, P < 0.05. C, Transwell invasion assay indicated that LC cells displayed the poorest invasiveness among isogenic lines. Data are mean ± SEM (n = 3). *, P < 0.01. D, Soft-agar assay revealed that LC cells showed the strongest ability of AIG as compared with the parental MDA-MB-231 and IV2 lines. Data are mean ± SEM (n = 3). *, P < 0.01. E, EMT/MET- and BCIC-related genes were analyzed with qRT-PCR in LC sublines in comparison with other two isogenic lines, the parental MDA-MB-231 and IV2 lines. *, Changes over 2-fold and P < 0.01. F, Flow cytometry analysis of the surface CD44 and CD24 expression in three isogenic lines. G, EMT/MET-related proteins were analyzed for the three isogenic lines by Western blot. H, Expression and localization of claudin-1 among the parental MDA-MB-231, IV2-1, and LC-1 cells. Scale bars, 15 μm. I, Analysis of Rac1/Cdc42 activity in the three isogenic lines using GST-PAK pull-down assay. Each experiment was performed in triplicates and was repeated at least 3 times.
- Figure 2.
Depletion of CD24 greatly reduces tumor growth, LN, and lung metastatic abilities of IV2 and LC cells in a xenograft orthotopic mouse model. A, Primary tumor growth of mice injected with CD24-depleted IV2 cells or the control cells was examined in the orthotopic mouse model. B and C, CD24 KD IV2 cells grew slower and formed smaller tumors than the control cells. Data are mean ± SEM. *, P < 0.05. D, The qRT-PCR analysis of CD24 levels in the control, CD24 KD, and CD24 KD-2 tumors. Data are mean ± SEM. *, P < 0.01. E, Long exposure (60 seconds) of IVIS imaging revealed that CD24 KD tumor–bearing SCID mice showed no sign of LN metastasis compared with the control tumor-bearing SCID mice whose LN metastases were detected. F, The quantification of luminescent photon signals of axillary LN in the control tumor- and CD24 KD tumor–bearing mice. The photon signals were normalized to the respective primary tumor size. Data are mean ± SEM. *, P < 0.01. G, The representative IVIS image of ex vivo examination of LN metastasis of the control tumor- and CD24 KD tumor–bearing mice. H, Detection of metastatic cells in mouse lung and LNs with qRT-PCR using human-specific primers. Data are mean ± SEM. *, P < 0.01. I–K, CD24 KD LC cells (n = 5) grew slower and formed smaller tumors than the control LC cells (n = 5) in the orthotopic mouse model. L, The qRT-PCR analysis of CD24 levels in the control and CD24 KD LC tumors. M and N, The presence of metastatic cells in mouse lung and LNs was determined with qRT-PCR using human-specific primers. The photon signals were normalized to the respective primary tumor size. Data are mean ± SEM. *, P < 0.01.
- Figure 3.
Depletion of CD24 attenuates EGF/EGFR/Stat3/Src and HGF/Met/Stat3/Src signaling cascades by destabilizing EGFR and Met. A, GSEA was applied to analyze the potential enriched pathway(s) in LC cells. Two angiogenesis-related RTK signaling pathways, Met and EGFR, were shown to be enriched in LC cells. B, The effect of CD24 depletion on HGF- and EGF-driven VEGF signaling cascades in CD24-expressing TNBC lines. Western blot analysis showed that knockdown of CD24 led to the reduction of HGF- and EGF-induced Met/Stat3/Src/Akt and EGFR/Stat3/Src/Akt signaling cascades. Each experiment was performed in triplicates and was repeated at least 3 times. C, The effect of CD24 KD on protein stability of EGFR and Met. Cells transfected with the control siRNA or CD24 siRNA were both treated with CHX (20 μg/mL) over the course of 24 hours. Loss of CD24 greatly decreased protein stability of EGFR and Met. D, The effect of bafilomycin A1 (BA) treatment in CD24-depleted cells in the absence or presence of HGF (50 ng/mL) and EGF (100 μg/mL).
- Figure 4.
Depletion of CD24 facilitates EGFR/Met internalization to promote protein degradation via lysosome-dependent pathway. A and C, Representative images of immunofluorescence staining of EGFR/Met and LAMP-1 before EGF/HGF stimulation in the control and CD24-depleted LC cells. B and D, Representative images of immunofluorescence staining of EGFR/Met and lysosome after EGF/HGF stimulation in the control- and CD24-depleted LC cells. The internalized receptor and its colocalization with lysosome were measured by confocal microscopy. Z stack (y–z axis) and colocalization intensity analysis were shown at the right and beneath the immunofluorescence images, respectively. Cells treated with or without EGF/HGF for 2 hours were immunostained with antibodies against EGFR/Met and LAMP-1, a lysosome marker, and imaged by 3D confocal microscopy. The y–z cross-sectional images of the cell are shown. The overlapped region of red and green fluorescence signals indicated by black arrows was measured by LAS AF software. It showed the colocalization of EGFR/Met and LAMP-1. Quantification of colocalization of EGFR with LAMP-1 was measured using the Mander's coefficient (JACoP plugin, ImageJ). 4.8% ± 2.5% (n = 10 cells; mean ± SD) and 38.1% ± 2.7% (n = 10 cells; mean ± SD) of EGFR were observed to colocalize with LAMP-1 in the control- and CD24-depleted LC cells, respectively, without EGF stimulation. 49.9% ± 4.4% (n = 10 cells; mean ± SD) and 74.2% ± 4.0% (n = 10 cells; mean ± SD) of EGFR were observed to colocalize with LAMP-1 in the control- and CD24-depleted LC cell, respectively, with EGF stimulation for 2 hours. 5.1% ± 0.5% (n = 10 cells; mean ± SD) and 35.1% ± 3.7% (n = 10 cells; mean ± SD) of Met were observed to colocalize with LAMP-1 in the control- and CD24-depleted LC cells, respectively, without HGF stimulation. 53.11% ± 2.8% (n = 10 cells; mean ± SD) and 69.9% ± 1.7% (n = 10 cells; mean ± SD) of Met were observed to colocalize with LAMP-1 in the control- and CD24-depleted LC cells, respectively, with HGF stimulation for 2 hours. Scale bars, 15 μm.
- Figure 5.
CD24 mAb treatment reduces the lung metastatic burden and significantly prolongs the survival time in a lung metastasis mouse model. A, The effect of CD24 mAb on anchorage-independent growth of LC cells. Incubation of LC cells with 2 μg/mL CD24 mAb ALB9 30 minutes prior to the soft-agar assay impaired the colony-forming ability of LC cells as compared with the cells treated with the control antibody mouse IgG1. B, The schematics of CD24 mAb treatment. Note that 0.5 million luciferase-labeled LC cells were injected intravenously via tail vein into SCID mice. Seven days later, CD24 mAb ALB9 (n = 5) or mouse IgG1 control (n = 5) was administered through tail vein injection (10 μg in 100 μL PBS; 0.03 mg/kg) every 2 days for 6 injections in total. The IVIS images of mice after intravenous injection after 1 week and 4 weeks later were shown. The histogram is the quantification result of mouse lung ROI signals 10 minutes, 10 days, and 30 days after intravenous injection. *, P < 0.05. C, Ten mice were intravenously injected with 2 × 105 LC cells and randomized into two groups prior to CD24 mAb administration. Mice were injected with CD24 mAb (n = 5) or mouse IgG1 Ab (n = 5) at a dose of 40 μg (0.12 mg/kg). Body weight of mice during the course of treatment. P < 0.05. D, The Kaplan–Meier survival plot of mice received mouse IgG1 Ab (n = 5) or CD24 mAb ALB9. Log-rank: P = 0.03.
- Figure 6.
CD24high and METhigh “double-positive” signature predicts poorer overall survival and shorter time of metastatic recurrence in TNBC patients. A, Kaplan–Meier survival analysis of breast cancer patients in Curtis dataset. The median value of CD24 for the cohort of total BC, TNBC, HER2, luminal A, and luminal B in Curtis dataset is 7.572, 7.891, 8.059, 7.393, and 7.233, respectively. B, Kaplan–Meier survival analysis of metastasis recurrence in Ma dataset. The medium value of CD24 in Ma dataset is 3.678. C, Kaplan–Meier survival analysis of breast cancer patients and patients with or without LN metastasis according to CD24 level in TCGA RNA-seq dataset. D, The representative images of CD24 IHC staining from paired primary tumor and metastatic LN (left). CD24 IHC scores in 15 patients (right). E, Kaplan–Meier survival analysis of TNBC patients with “CD24 high MET high,” “CD24 high MET low,” “CD24 low MET high,” or “CD24 low MET low.” The median values of CD24 and MET are 7.981 and 3.31. F, Kaplan–Meier survival analysis on TNBC patients with “CD24 high EGFR high,” “CD24 high EGFR low,” “CD24 low EGFR high,” or “CD24 low EGFR low.” The median values of CD24 and EGFR are 7.981 and 2.155. G, The schematic of CD24′s role in TNBC progression.
Additional Files
Supplementary Data
- Figure S1-S11 and Table S1-S4 - Establishment of breast cancer cells derived from the long-term lung metastatic nodules in a xenograft mouse model (Figure S1). LC cells display strong lymph node-tropic metastasis and tumorigenicity in an orthotopic mouse model (Figure S2). Inhibition of CD24 impairs the ability of in vitro anchorage-independent growth and cell proliferation in CD24-expressing TNBC cell lines (Figure S3). Effect of inhibition of CD24 on luminal breast cancer cell line MCF-7 (Figure S4). CD24-depleted IV2 cells failed to colonize in the lung of SCID mice (Figure S5). CD24 depletion did not alter initial lung targeting of IV2 cells (Figure S6). Ectopic expression of CD24 in 231 cells promotes primary tumor growth and lymphangiogensis/angiogenesis (Figure S7). Depletion of CD24 impairs lymphangiogenesis and angiogenesis via downregulation of VEGF-C and -A (Figure S8). HGF- and EGF-induced VEGF-A and -C expression in TNBC cell lines (Figure S9). Treatment of CD24 mAb reduced lung colonization of MDA-MB-468 in a tail vein injection mouse model (Figure S10). Analysis of clinical correlation between CD24 and EGFR/Met protein expression in 133 TNBC specimens using IHC staining (Figure S11) Cox proportional regression analysis of the influence of CD24 level on 5-year survival of breast cancer patients with different subtypes (Table S1). Cox proportional regression analysis of the influence of CD24 level on recurrence of BC patients in Ma dataset (Table S2). All antibodies and reagents used in this study (Table S3). Primers, oligos and plasmids used in this study (Table S4).
Volume 18, Issue 1
