Table 1.
Score and criteria in the semi-quantitative evaluation of murine metastasis models.
Fig 1.
Morphological comparison of parental cell lines and FL sublines.
(A) A549 parental cells showed a polygonal, epithelial morphology in attachment, and formed somewhat irregularly rounded spheroids on low-binding dishes. After long-term low attachment cultures (at day 90), cells (now designated as A549-FL) formed smaller spheroids, with some cells floating as single cells. Upon replating on ordinary dishes, A549-FL cells readily reattached and spread. The A549-FL subline, when attached on ordinary plates, showed more spindle or rounded shapes than parental cells. (B) Similar morphological changes occurred in H441 cells during long-term 3D low-attachment cultures.
Fig 2.
FL sublines exhibit greater cell growth potential than parental cells on low-binding cultures.
(A) The A549 cell line and A549-FL subline grew at a similar rate in attachment cultures. The A549-FL subline grew at a markedly higher rate than the A549 cell line in low-attachment (detached) cultures (p < 0.05). The H441-FL subline showed a slightly higher cell growth rate than the H441 cell line in attachment cultures (p < 0.05). The H441-FL subline showed a markedly higher cell growth rate than the H441 cell line in detachment cultures (p < 0.05). Data are shown as the box-and-whisker plot (minimum, lower quartile, median, upper quartile and maximum) of 5 replicates of cultures. Statistical analysis was performed by the Mann-Whitney U test. (B) The A549-FL subline showed a markedly larger fold increase in DNA than the A549 cell line in detachment cultures (p < 0.05). The H441-FL subline also showed a markedly larger increase in DNA than the H441 cell line in detachment cultures (p < 0.05). Note that H441 cells actually showed a slight decrease in DNA during the 7 days in a detachment culture. Data are shown as the box-and-whisker plot of the ratio (day 7/day 0) of the amount of DNA extracted from 4 samples of replicates of cultures. Statistical analysis was performed by the Mann-Whitney U test.
Fig 3.
Transwell migration and invasion assays.
(A, B) Transwell migration assay. Representative microscopic images of cells that migrated through the transwell in the migration assay. (Giemsa stain, magnification ×200). No significant differences were observed between the FL subline and parental cell line for A549 or H441 cells. The quantitation of cells that migrated through the transwell in the migration assay. The data are shown as the box-and-whisker plot of the mean number of migrated cells per visual field (A) or 3 visual field (B) (magnification ×100) of 4 replicate wells. No significant differences were found between the FL subline and parental cell line for A549 cells or for H441 cells. (C, D) Transwell Matrigel invasion assay. Representative microscopic images of cells that invaded through the transwell in the Matrigel invasion assay. (Giemsa stain, magnification ×200). No significant differences were observed between the FL subline and parental cell line for A549 or H441 cells. Quantitation of cells that invaded through the transwell in the Matrigel invasion assay. The data are shown as the box-and-whisker plot of the mean number of cells per visual field (C) or 3 visual field (D) (magnification ×100) of 4 replicate wells. The results obtained were similar between A549 and A549-FL cells. Invasion was slightly weaker in H441-FL cells than in H441 cells. Statistical analysis was performed by the Mann-Whitney U test.
Fig 4.
Levels of apoptosis and cell cycle regulators in parental and FL cells in detachment.
(A) Parental and FL cells in low attachment cultures were lysed and subjected to a Western blot analysis with antibodies to poly ADP-ribose polymerase (PARP), cleaved PARP, and β-actin. Parental cells in attachment cultures were included for comparison. Under detached conditions, apoptosis was greater in FL cells than in parental cells. * in PARP indicates the cleaved fragment of PARP. (B) Parental and FL cells in low attachment cultures were lysed and subjected to a Western blot analysis with antibodies to cyclin A2, B1, D1, E1, p27, and β-actin. While the levels of cyclin A2, B1, D1, and E1 were similar between parental A549 and A549-FL cells, p27 levels were markedly lower in A549-FL cells than in parental A549 cells. Similarly, H441-FL cells showed reduced p27 levels. Number below gel images represent the relative protein levels of the indicated proteins after normalization to β-actin. Data shown are representative of 2 independent experiments.
Fig 5.
Metastatic tumor formation after an intracardiac injection of parental and FL cells.
(A) Metastatic tumors were formed in multiple organs in all mice injected with parental or FL cells. Representative macroscopic and microscopic pictures are shown of lung (white arrow), adrenal (white arrow head), and brain metastases of luc-A549 cell lines. (B) In vivo luciferase imaging of luc-A549 cell lines and luc-A549-FL sublines at 5 weeks post-implantation. Representative images from ventral view are shown. Quantification of tumor-derived photons at 5 weeks post-implantation. Data are shown as the mean ± standard deviation of the photon flux of 5 animals. The Luc-A549-FL subline showed markedly greater metastatic tumor growth than luc-A549 cells. (C) In vivo luciferase imaging of luc-H441 cell lines and luc-H441-FL sublines at 6 weeks post-implantation. Representative images from dorsal view are shown. Quantification of tumor-derived photons at 6 weeks post-implantation. No significant differences were observed in tumor-derived photons between luc-H441 and luc-H441-FL cells. Data are shown as the box-and-whisker plot of the photon flux of 5 animals. Statistical analysis was performed by the Mann-Whitney U test.
Table 2.
Comparison of murine metastasis models with a semi-quantitative evaluation.
A, A549 vs A549-FL; B, H441 vs H441-FL.
Fig 6.
Selected examples of a gene set enrichment analysis (GSEA).
Among the “hallmark gene sets” analyzed, EPITHELIAL MESENCHYMAL TRANSITION (EMT) was noted in A549-FL and H441-FL cells versus their parental counterparts. Enrichment plots for EMT signatures (A549-FL vs A549 and H441-FL vs H441) are shown.
Table 3.
A. Number of gene alternations in the gene expression array with limiting criteria in the FL subline (versus the parental cell line). B. Gene set enrichment analysis: significantly enriched in the A549-FL subline (versus the A549 cell line). C. Gene set enrichment analysis: significantly enriched in the NCI-H441-FL subline (versus the NCI-H441 cell line).
Fig 7.
Zinc finger E-box binding homeobox 1 (ZEB1) was induced in A549-FL cells and its knockdown (KD) led to apoptosis and the inhibition of cell growth in A549-FL cells.
(A) Parental cells under detached conditions (3 days after detachment) and FL cells in a continued detachment (low attachment) culture were lysed and subjected to a Western blot analysis with antibodies to E-cadherin, vimentin, ZEB1, and β-actin. A549 cells in an attachment culture were included for comparison. Long-term low attachment (detached) cultures caused the slightly weaker expression of E-cadherin and stronger expression of vimentin and ZEB1 in A549-FL cells than in A549 cells. (B-D) A549-FL cells were transduced with lentiviral vectors (shZEB1-A, shZEB1-B, and shScramble), and after a brief selection with puromycin, used for a Western blot analysis, colony formation assay, and cell count assay, without cloning (see the Materials and methods for details). (B) ZEB1 KD induced apoptosis in A549-FL cells, as indicated by the emergence of the cleaved PARP fragment. Number below gel images represent the relative protein levels of the indicated proteins after normalization to β-actin. Data shown are representative of 2 independent experiments. (C) Representative images of macroscopic colony formation assay of 4 replicates of cultures (Giemsa stain). Colony formation was inhibited more by shZEB1-A and shZEB1-B than by the control (shScramble). (D) Cell growth was inhibited more by shZEB1-A and shZEB1-B than by the control (shScramble) in attachment and detachment cultures. Data are shown as the box-and-whisker plot of 10 replicates. Statistical analysis was performed by the Kruskal-Wallis test followed by Tukey's test.
Fig 8.
c-Myc was amplified in H441-FL cells and its inhibition decreased the growth of these cells.
(A) A copy number analysis by CGH demonstrated genomic regions of amplification (red bar) and deletion (green bar) in H441-FL cells. The red arrow indicates amplification on chromosome 8q24, which harbors the c-Myc gene. c-Myc expression was increased in H441-FL sublines at the protein level. (B-D) H441-FL cells were transduced with lentiviral vectors (shc-Myc-A, shc-Myc-B, and shScramble), and after a brief selection with puromycin, used for a Western blot analysis, colony formation assay, and cell count assay, without cloning (see the Materials and methods for details). (B) C-Myc KD did not induce apoptosis, but caused a modest increase in p27 levels. (C) Representative images of macroscopic colony formation assay of 4 replicates of cultures (Giemsa stain). Colony formation was inhibited more by shc-Myc-A and shc-Myc-B than by the control (shScramble). (D) Cell growth was inhibited more by shc-Myc-A and shc-Myc-B than by the control (shScramble) in attachment and detachment cultures. Data are shown as the box-and-whisker plot of 10 replicates. Statistical analysis was performed by the Kruskal-Wallis test followed by Tukey's test. (E) The H441 FL subline was treated with the indicated doses of JQ1 for 24 or 48 hours, or with 0.1% DMSO as a negative control. Cell lysates were subjected to a Western blot analysis for c-Myc, p27, PARP, cleaved PARP, and β-actin. An analysis of c-Myc was performed after a 24-hour treatment, while that of the others was performed after a 48-hour treatment. The results obtained demonstrated that JQ1 caused the depletion of c-Myc, an increase in p27, and the induction of cleaved PARP in a dose-dependent manner. (F) The JQ1 treatment inhibited the growth of H441 and H441-FL cells in detachment cultures. The H441-FL subline was slightly more sensitive to the treatment with 0.1 μM JQ1 than the parental H441 cell line. Data are shown as the box-and-whisker plot of 5 replicates. Statistical analysis was performed by the Kruskal-Wallis test followed by Tukey's test. (A, B, E) Number below gel images represent the relative protein levels of the indicated proteins after normalization to β-actin. Data shown are representative of 2 independent experiments.
Table 4.
Significantly amplified gene lesions with gene names in a copy number analysis of the NCI-H441-FL subline (versus the NCI-H441 cell line).