Fig 1.
Diagrams and representative pictures of cancer-endothelial cell adhesion measurement.
A single cancer cell is attached on the end of the Con A-coated tipless cantilever (A) and then brought into contact with the endothelial (HBMEC-60) cell-cell junctions (B). After a defined contact time, the cantilever is retracted (C) until the cancer cell is entirely separated from endothelial cells (D). The phase contrast image (E) shows an MB435 cell (white arrow) attached to the cantilever and in contact with a HBMEC-60 cell-cell junction. Force curves (F) are acquired during the cantilever approach (red) and retraction (blue), and the retraction curve is used to calculate adhesion forces. Scale bar in E = 10 µm.
Fig 2.
Typical adhesion force curves of MB435 (A) and MB231 (B) to HBMEC-60 as a function of increasing contact time.
Representative curves for AFM cantilever retraction that were recorded in the experiments were plotted for increasing lengths of cell-cell contact time in seconds (s). The total force required to separate the tumor cell from the endothelial cells increased as a function of increasing contact time. The peak force in the aggregate curve for both MB435 (A) and MB231 (B) cells shifts temporally to the right with increasing duration of contact. The vertical and horizontal black bars stand for adhesion force (pN) and cantilever retraction distance (µm), respectively.
Fig 3.
A relationship between the total cancer-HBME adhesion and duration of cell-cell contact.
Adhesion strength between MB231 and MB435 to HBMEC-60 cells was measured for different contact times: 0.5, 1, 2, 5, 10, 30, 60, 120, and 300 sec. The overall total adhesion forces in MB231 (asterisks) cells at all time points were lower than those in MB435 (diamonds). In both cell lines, from 0.5 to 300 sec cell-cell contact time the total adhesion forces increased progressively 23.79-fold (from 23.27 ± 5.73 to 553.64 ± 86.10 pN) and 9.29-fold (68.81 ± 5.91 to 639.37 ± 62.50 pN) for MB231 and MB435 respectively. Note similar time courses of adhesion dynamics to HBMEC-60 displayed by both cancer cell lines, whereas the adhesion force increased dramatically within the first 10 sec (MB231) or 30 sec (MB435) of contact with HBMEC, and then more gradually As a negative control, the MB435 (triangles) and MB231(circles) cells were brought into contact with a hydrogel-coated dish instead of an endothelial monolayer. *p< 0.001 by ANOVA analysis.
Fig 4.
Distributions of rupture forces detected during cancer-HBME adhesion with cell-cell contact times of 0.5, 1, 2, 5, 10, 30, 60, 120 and 300 sec.
Note 3.89-fold increase (from 903 to 3516 counts) in the frequency counts of the detectable rupture events for MB435 and 15.39-fold increase (from 156 to 2401 counts) for MB231 from 0.5 to 300 sec of cell-cell contact. Also note a significant shift of the rupture force frequency distribution histograms to the right for both cancer cell lines. Similar changes in histograms of rupture force distribution for both MB435 (A) and MB231 (B) cells indicate that both the increase in frequencies of individual adhesion events and involvement of stronger ligand-receptor interactions contributed to the change in total adhesion force over time.
Fig 5.
Changes in cancer-HBME total adhesion force with antibody treatment.
In adhesion between MB435 and HBMEC-60 (A), from 0.5 to 30 to 60 sec contact times, total adhesion forces were significantly reduced by antibodies to TF-Ag and β1 showing descending inhibition dynamics with the highest inhibition at 0.5 sec, compared to the baseline; while anti-α3 and anti-Gal-3 showed ascending inhibition dynamics with the highest inhibition at 60 sec. As positive control, activating β1 antibody (Activ β1) induced the greatest adhesion (64.55%) at 0.5 sec, which is in agreement with what was observed with inhibiting anti-β1 antibody. In adhesion between MB231 and HBMEC-60 (B), significant inhibition was observed at the 60 sec contact time with antibodies against TF-Ag (-23.26%), β1 (-23.81%) and Gal-3 (-41.71%), but not statistically significant with anti-α3, compared to the baseline. *p< 0.05; **p< 0.01, vs the baseline. § p< 0.05 vs 0.5 sec contact. n> 6.
Fig 6.
Distribution frequencies of rupture forces collected from MB435-HBMEC-60 adhesion for cell-cell contacts of 0.5 (A,D,G,J,M,P, and S), 30 (B,E,H,K,N,Q, and T) and 60 (C,F,I,L,O,R, and U) seconds before and after antibody (10 µg/ml) treatments.
There were no changes seen in rupture distribution in CO2IM treated control (A, B and C). Function-inhibiting antibodies to TF-Ag (D, E and F), Gal-3 (G, H and I), α3 (J, K and L) and β1 (M, N and O) decreased the number of rupture events compared to the respective baselines. Note reduced number of the weakest (30–50 pN) and midrange (50–80 pN) adhesions at the earliest 0.5 sec time point caused by anti-TF-Ag (D), which apparently prevented the formation of stronger adhesions at 30 and 60 sec time points (E and F). Anti-Gal-3 antibody caused rather limited inhibitory effect on the weakest (30–50 pN) and midrange (50–80 pN) adhesions at 0.5 sec time point (G) It demonstrated, however, a pronounced inhibition of adhesions across the entire spectrum at 30 and 60 sec (H and I), indicating that Gal-3 involvement requires additional time to unfold, and that stronger, integrin-mediated adhesive events, may depend on preceding Gal-3 mobilization. Limited inhibitory effect was detected with function-blocking antibodies against integrins α3 and β1 at 0.5 sec (J and M respectively). At 30 and 60 sec time points, however, both anti-α3 (K and L) and anti-β1 (N and O) caused more significant inhibition of adhesions across the entire spectrum including the stronger (>80 pN) adhesions. A combination of the four function-blocking antibodies (P-R) showed a greater inhibitory effect than any of the antibodies alone (D to O). In the positive control (S-U), the function-activating antibody to β1 increased the number of detectable ruptures.
Fig 7.
Distribution frequencies of rupture forces collected from MB231-HBMEC-60 adhesion experiments for cell-cell contact times of 0.5 (A,D,G,J,M,P, and S), 30 (B,E,H,K,N,Q, and T) and 60 (C,F,I,L,O,R, and U) sec before and after antibody (10 µg/ml) treatment.
No obvious changes were detected in rupture distribution in CO2IM treated controls (A-C) compared to the baseline. Noticeable reduction in the number of the weakest (30–50 pN) and midrange (50–80 pN) adhesions at the earliest (0.5 sec) time point was caused by anti-TF-Ag JAA-F11 antibody (D), which also prevented the formation of stronger adhesions at 30 and 60 sec time points (E and F). Only minimal inhibitory effect on midrange (50–80 pN) adhesions was detected at the earliest (0.5 sec) time point with anti-Gal-3 antibody (G). However, this same antibody caused very pronounced inhibition across the entire spectrum of adhesions at 30 and 60 sec (H and I). The inhibitory effect of function blocking anti-α3 and anti-β1 antibodies was rather limited at 0.5 sec time point (J and M respectively). They did cause, however, noticeable inhibition of adhesions across the entire spectrum at 30 and 60 sec time points (K, L, N and O). The inhibitory effect of the mixture of all four antibodies was significantly more pronounced across the entire spectrum of adhesions at all three time points (P-R) than any of the function blocking antibodies alone. In the positive control (S-U), the function-activating antibody to β1 increased the number of detectable ruptures.