Collective dynamics of gastric cancer cells in fluid

Gastric cancer (GC) is the most common digestive system malignant cancer, and gastric cancer cells (GCC) can migrate in normal solid tissue and lymphatic fluid. Previously, much research has focused on the migration process when the cells are in the solid condition, such as migration through tissue, adhesion, and invasion processes, while little is known about the migration process of GCC in lymphatic fluid. In the current study, we investigate the migration of GCC in a fluid condition in an in vitro environment. We find that the cells diffuse mainly because of their cell viability. Therefore, despite the fact that lymph fluid is almost quiescent, GCCs can migrate around easily. The dynamics of cells also demonstrate a collective glassy dynamic similar to ordinary inactive glassy materials. As density of the cells increases, the movement of the cells becomes slower, and the collective dynamic becomes heterogeneous, which is similar to the dynamically heterogeneous behavior in glassy materials. The results will help us gain a better knowledge of the characteristics of GCC dynamics in the liquid phase which is crucial for the understanding of the mechanism for lymphatic metastasis. This can also potentially help early diagnosis of lymph node metastasis in GC and provide new insights for future clinical treatment.

Figure 1

Images by fluorescence inverted microscope of AGS cells through (a) bright field channel and (b) GFP fluorescence channel (488nm). Images of (c) normal AGS cells without cisplatin treatment and (d) the low viability AGS cells incubated with cisplatin (50 μg/ml) for 12 h. (e) The relative cell survival rate of AGS cells incubated with different concentration cisplatin for 12 h.

Figure 2

(a) Three-dimensional reconstruction image of suspended cells by confocal microscope. (b) The 2D cross section with the maximum $\varphi =0.748$ of the 3D image stacks. (c) Binarized picture using adaptive local thresholds. (d) The color image of different cells segmented by a marker-based watershed algorithm. (e) The overlaid blue circles are cells optimized by a generalized Hough transform.

Figure 3

The 2D cross section of the cell packing by confocal microscope of the NAC groups with $\varphi =\left(\mathrm{a}\right)0.527$, (b) 0.596, (c) 0.634, (d) 0.675, (e) 0.715, and (f) 0.748, and (g) the MIXED group with $\varphi =0.672$.

Figure 4

(a) The displacement maps of GCCs of $\varphi =0.675$ within 20 min. The displacements of similar direction are marked by the same color. (b) The probability distribution of cell displacements in $X$ direction of the fourth experiment with the area fraction $\varphi =0.675$ at time interval $\mathrm{\Delta }t=1.3$ and 32.5 min respectively.

Figure 5

Dynamics of GCCs. (a) MSD curves of GCCs for NAC groups whose area fraction $\varphi =0.527$, 0.596, 0.634, 0.675, 0.715, 0.720, and 0.748, LVAC and MIXED groups whose area fraction $\varphi =0.719$ and 0.672, respectively. (b) MSD curves for LVAC and NAC cells in the MIXED groups of $\varphi =0.672$. (c), (d) The auto-correlation overlap coefficient $Q\left(\mathrm{\Delta }t\right)$ and the four point susceptibility ${\chi }_4$ curves of GCCs as a function of time interval $\mathrm{\Delta }t$.

Figure 6

(a) Self-diffusion coefficient $D$ calculated by a linear fit of the MSD curves. (b) The corresponding time $\tau$ and length scales $\xi$ associated with the collective dynamics as function of $1/|\varphi /{\varphi }_0-1|$. Lines are power-law fitting results.