Compressive stress inhibits proliferation in tumor spheroids through a volume limitation

Abstract

In most instances, the growth of solid tumors occurs in constrained environments and requires a competition for space. A mechanical crosstalk can arise from this competition. In this article, we dissect the biomechanical sequence caused by a controlled compressive stress on multicellular spheroids (MCSs) used as a tumor model system. On timescales of minutes, we show that a compressive stress causes a reduction of the MCS volume, linked to a reduction of the cell volume in the core of the MCS. On timescales of hours, we observe a reversible induction of the proliferation inhibitor, p27Kip1, from the center to the periphery of the spheroid. On timescales of days, we observe that cells are blocked in the cell cycle at the late G1 checkpoint, the restriction point. We show that the effect of pressure on the proliferation can be antagonized by silencing p27Kip1. Finally, we quantify a clear correlation between the pressure-induced volume change and the growth rate of the spheroid. The compression-induced proliferation arrest that we studied is conserved for five cell lines, and is completely reversible. It demonstrates a generic crosstalk between mechanical stresses and the key players of cell cycle regulation. Our results suggest a role of volume change in the sensitivity to pressure, and that p27Kip1 is strongly influenced by this change.

Figure 1

Multicellular spheroids and mechanical stress. (A) Growth of a CT26 MCS from Day 0 to Day 10 (scale bar, 200 μm). (B) Principle of the experiment. Dextran is added to the culture medium, and does not penetrate the MCS. This results in a moderate osmotic stress exerted only on the outermost layer of cells, which is mechanically transmitted to the inner cells, resulting in a compressive stress. (C–G) The volume normalized to the initial volume plot as a function of time. The initial volume is ∼4 × 10⁶μm³ for every MCS, and four MCSs are displayed per condition to show reproducibility. (C–F) 0 Pa (blue) and 5 kPa (red); (G) 0 Pa (blue) and 10 kPa (red). The error bars are fixed errors due to measurements, and we estimated that we measured a spheroid diameter with a precision of 10 μm.

Figure 2

Volume reduction through a compressive stress. Cell-to-cell distance plot as a function of the distance from the center for different time points: Day 0 (●), 5 min (▵), and Day 1 (□). The error bar obtention is described in the Supporting Material. The experiments have been repeated N ≥ 3. A two-tailed t-test yields a p value < 0.002 for the points at the center of the MCS, between t = 0 and t = 5 min. To see this figure in color, go online.

Figure 3

Proliferation arrest at the restriction point. (A) Results of the flow cytometry experiments. The percentage of cells in G1 is calculated using the software MODFIT, and the variation of this percentage is plot as a function of the applied stress, either for CT26 MCSs (blue), CT26 individual cells (green), or HT29 MCSs (red). (B) Level of total pRb and p27^Kip1 as a function of time. (C) Level of pRb (T373) plot as a function of time for different conditions of siRNA and mechanical stress. Each experiment was performed either under 0 kPa or 10 kPa. N ≥ 2.

Figure 4

Evolution of pRb (T373) and p27^Kip1. (Top) Averaged images of cryosections of HT29 MCSs for the staining of pRb (T373) and p27^Kip1. The color code indicated is a hot map, blue being the lowest value. The intensity of each image is normalized to its highest value, then averaged over the angle. (Bottom) Density profiles of these images. (A) t = 0 h (○) for P = 10 kPa; (B) 24 h for P = 10 kPa (▵); (C) t = 48 h for P = 10 kPa (□); (D) t = 72 h for P = 10 kPa (×) or 0 Pa (○). The densities are given in number of nuclei positive for the staining/μm². Each time point was repeated N ≥ 3. We show the control in the uncompressed scenario only for Day 0 and Day 3, inasmuch as the level does not change.

Figure 5

Biomechanical sequence and spatial correlation. (A) Sequence from compressive stress (B) to pRb (T373) reduction (E), through mean diameter reduction (C) and p27^Kip1 overexpression (D). (Plain arrow) Causality between events; (dashed arrow) spatial and temporal correlation between events. Note the log scale in time in panels B–E. Error bar obtention is described in the corresponding paragraphs of the Materials and Methods. To see this figure in color, go online.

Figure 6

Macroscopic correlation between CT26 MCS volume reduction obtained in 5 min and the growth rate obtained by the growth curve fitting over 15 days. A linear regression gives a correlation of 0.98. To see this figure in color, go online.