Figure 1.
Cell scattering assays for monitoring cell invasive ability.
A) a. Single-cell scattering assay: Single, isolated LNCaP cells were mixed with collagen and examined daily. MT1-GFP-expressing cells gradually displayed a scattering growth pattern after 6 days in culture as compared to the aggregated growth of GFP-expressing cells. b. Aggregate scattering assay: SK-3rd cells were cultured for 12 days to form mammospheres, transferred and embedded in collagen, and then cultured under normoxia or hypoxia. A scattered pattern was seen after 3 days under hypoxia. c. Microcarrier bead scattering assay: MT1-GFP expressing LNCaP cells were cultured with beads for 24 hours, then transferred and embedded into collagen. Invaded cells were seen as a scattered pattern at day 3. Scale bars = 20 μm. B) Assessment of cancer cell invasiveness using novel 3-D invasion assay: LNCaP cells expressing vector control (a) or MT1-MMP (b), human prostate cancer PC3 cells (c), and Du145 cells (d) were suspended in collagen, dotted into the well of a 96-well plate, and covered with collagen followed by media. After 18-h incubation, PC3 and Du145, as well as MT1-MMP expressing LNCaP cells, showed increased numbers of invaded cells as compared to LNCaP vector control cells. Scale bar = 50 μm.
Figure 2.
Evaluation of novel 3-D invasion assay.
A & B) Comparison of the novel 3-D invasion assay and 3-D cell scattering assay: The effect of control IgG (Rabbit, 10 μg/ml), tissue inhibitor of metalloproteinase-2 (TIMP-2) (10 nM), or anti-MT1-MMP-hemopexin domain antibody (Anti-PEX Ab) (10 μg/ml) on MT1-GFP-induced LNCaP cell invasion was evaluated via the 3-D invasion assay (A) and 3-D cell scattering assay (B). An invasive growth pattern was photographed on day 1 (A) and day 6 (B) respectively. TIMP-2 and the anti-PEX Ab decreased the cell invasive/scattering ability of MT1-GFP expressing cells. Scale bar = 20 μm. C) Dose-dependent inhibition of human breast cancer MDA-MB-231 cells by anti-β1 integrin antibody: MDA-MB-231 cell invasion was examined using the 3-D invasion assay in the presence of the anti-β1 integrin antibody at different concentrations vs. control IgG. Invaded cells were microscopically examined (left panel) and counted (right panel). Bars represent the mean + SD.
Figure 3.
Standardization and automation of the 3-D invasion assay.
A) Schematic diagram (top panel) of the 3-D invasion assay with tooled plate: (a) The cell-collagen mixture is dotted into the 2 mm in diameter pits in the center of each well of the tooled 96-well plate, creating spheres that occupy a volume of 4.18 mm3; (b) protruding cell-collagen hemispheres are covered with collagen; (c) medium containing compounds is added and plate is incubated for 18 hours; and (d) invaded cells beyond pit boundary are counted. Black dots represent cells. Diagram not to scale. An image of a section of a tooled plate with a pit in the center of each well is shown (bottom panel). B) Inhibition of collagen contraction via dialdehyde dextran: HT1080 cells were mixed with collagen with or without the addition of dialdehyde dextran before overlaying with collagen. Phage contrast images were taken at time 0 and 18 hours (left panel, Scale bar = 250 μm) and 18 hours in the tooled plate (right panel, Scale bar = 100 μm). Red arrows indicate the edge of the pits. Red arrowheads indicated the edge of the cell-collagen spheres. C) Phase contrast and fluorescent images of invaded HT1080 cells in the tooled plate stained with Hoechst dye. Scale bar = 100 μm. D) Automated quantification: Both phase contrast and fluorescent images (a & b) of HT1080 cells following invasion were acquired using a Nikon TE-2000s controlled by the NIS Elements imaging software. A threshold of the phase contrast image of the first well was adjusted based on the contrast to create two binary layers, one inside the pit (highlighted in black) and one outside of pit (highlighted in pink) (c). The binary layer outside the pit (pink area) was copied to form a region of interest (ROI), which was then applied to the Hoechst image (d). The invaded cells within the ROI were then automatically counted with the counted cells appearing purple in color (e).
Figure 4.
Simultaneous determination of invasion and cytotoxicity.
HT1080 cells expressing GFP were assembled in the 3-D invasion assay in the presence of (A) DMSO, (B) anti-β1 integrin antibody (5 μg/ml), (C) staurosporine (STS) (500 nM), or (D) paclitaxel (Taxol, 1 μM) for 18 hours. The cells were stained with Hoechst and propidium iodide (PI) for 30 minutes followed by microscopic examination. Decreased cell invasion is observed upon treatment with anti-β1 integrin antibody, STS, and taxol. However, STS and taxol show higher levels of PI staining, indicative of increased cell death. Numbers in the top right corners represent the total number of invasive cells (Hoechst images) or cells that died after invading (PI images) within the field. Scale bar = 100 μm.
Figure 5.
Primary screening using the developed 3-D invasion assay.
A) The NCI Diversity Set II Compound Library was screened using the 3-D invasion assay and MDA-MB-231 cells, resulting in 24 positive hits, 9 of which were inhibitory but not cytotoxic. B) A quinocarmycin analog, DX-52-1 was positively identified as an anti-invasive compound. Representative images are shown for DMSO control and DX-52-1 (10 μM) after 18 hours incubation. Abrogated cell invasion and low cytotoxicity were observed in cells treated with DX-52-1. Numbers in the top right corners represent the total number of invasive cells within the field. Scale bar = 100 μm. C) Transwell chamber migration was performed with MDA-MB-231 cells treated with DMSO or DX-52-1 for 18 hours. Representative images of membranes (8 μm pore size) are shown in left panel. Quantification of cell migration is shown in right panel. DX-52-1 treated cells had decreased migratory ability. Bars represent the mean + SD. D) Fluorescent Protease Detection assay (Sigma) performed with cell lysates from MDA-MB-231 cells treated with DMSO or DX-52-1 for 18 hours. No effect on proteolytic activity was observed upon treatment with DX-52-1. Bars represent the mean + SD. ** refers to p<.01.