Fig 1.
Device design, operation and applied pressure profiles.
(a) 3D optical profilometer image of a micro-piston device showing the compartment including total height of the channel where the micro-piston is suspended, height of the micro-piston and the gap between the piston and surface of the channel. (b) Device operation steps from hydrogel coating, to piston-retracted loading of cells and cyclic compression on cells. The device was kept in static state during hydrogel loading and put into piston-retracted state for cell loading. In this state the membrane, and thus micro-piston, are retracted from the static position towards the top control channel by the applied negative pressure. After loading and settlement of the cells, flexible components were actuated down to their initial position. Cyclic compression of cells is illustrated by the membrane deflection and micro-piston brought onto cells by pressure applied through the control channel and retracted back after compression (repeated cyclically). Representative pressure profiles for piston-retracted loading, and sequential cyclic compression process are shown in (c) and (d), respectively.
Fig 2.
Computational and experimental validation of the micro-piston contact pressures across different membrane thicknesses.
(a) Simulated piston displacement and maximum piston contact pressures at different membrane thicknesses as function of applied pressure. The left pattern of curves refers to the left y-axis (piston displacement (μm)), while the curve pattern on the right refers to the right y-axis (max piston contact pressure (N/m2)), resulted at each externally applied pressure shown in x-axis. At a certain applied pressure (x-axis), micro-piston moves towards the bottom substrate and thus shows a displacement (left y-axis) depending on membrane thickness and applies a corresponding contact pressure (right y-axis). Colors of the curves refers to the membrane thicknesses indicated on the graph. (b) Correlation among independent experiments with the pressure sensor read for the externally applied pressures applied based on simulations in (a) run for the corresponding membrane thicknesses, resulting in a similar piston contact pressure.
Table 1.
Compression applied to cancer cells via fabricated different membrane thicknesses, showing externally applied pressures with sensor readings (in parentheses) from experiments and the predicted resulting internal maximum piston contact pressures.
Fig 3.
Representative fluorescent microscopy images of live imaging of GFP-tagged actin cytoskeleton of cells.
(a) Schematic of the on-chip BacMam transduction into cancer cells for the expression of actin-GFP fusion protein. (b) Phase contrast (right) and fluorescent (left) images of the GFP-tagged actin expressing SKOV-3 ovarian cancer cells in static state (0 kPa) or retracted back to static state (0 kPa) after compression for the given pressure range. Yellow arrows indicate example cells exhibiting rearrangement of actin structures with formation of new extensions, blue arrows for actin deformation with rearrangement of existing actin structures, green arrows for actin disruption with raptures and pink arrows for complete actin disruption. (c) Phase contrast (left) and fluorescent (right) images showing the change in GFP-tagged actin in the cells as a result of the applied pressure in the compressed state. Pressure was applied sequentially in ascending order.
Fig 4.
Corrected total cell fluorescence (CTCF) calculated from images of GFP-tagged actin cytoskeleton of cells during sequential cyclic compression at the indicated pressures.
Student’s t-test was used to determine the statistical significance of live cell actin deformation across increasing pressure ranges. Continuous horizontal bars show significant differences between compressed cell groups under micro-pistons for images taken at the time of compression at the indicated pressure ranges. Dashed horizontal bars show significant differences between the cell groups at rest under micro-piston for images taken after compression stages at the corresponding pressure ranges of the cycles. Results represent at least three independent experiments.
Fig 5.
End point assay staining and analysis for actin and nuclei of cancer cells fixed at zero time or after 24 h-recovery, following 1 hour-long cyclic compression between 0 and 15.6 kPa.
(a) Representative images of control and compressed cell groups stained for actin (green) and nuclei (blue) for their shape at zero time and 24 h-recovery after cyclic compression. Dashed areas are under piston, while the surrounding is control region (CR). Merged: merged form of the phalloidin (stain for actin) and Hoechst (stain for nuclei) epi-fluorescence images; merged with PH: merged form of the epi-fluorescence images with the corresponding phase-contrast (PH) image; magnified: two-hundred-fold magnification images of the region under micro-piston. The unevenness of brightness in phase contrast images is due to that imaging focal plane was on cells cultured on glass while the micro-piston, brought back to static state after compression, was suspended on average 108 μm above the glass. Representative arrows (white) show the distinct actin deformation by the increased fluorescence signals at the edges of the cells in the compressed groups under the micro-piston at zero time after compression. Nuclei deformation in control cells at control regions (CR) versus in compressed cells under piston at zero time or 24 h-recovery after cyclic compression, was measured for area (b), circularity (c) and aspect ratio (d) of the nuclei (mean ± s.e.m. n = 588, 230, 358, 174 cell nuclei from at least two or three independent experiments per each group). Student’s t-test was used to determine the statistical significance of nuclear deformations in compressed and recovered cells after compression. Dashed horizontal bars show significant differences between the control region (CR) group and compressed group under micro-piston. Continuous horizontal bars show the comparison pairs between the cell groups under the micro-piston at zero time and 24-recovery after compression. The statistical significance levels among all groups are given in S2 File.
Fig 6.
End point assay staining and analysis for actin and nuclei of cancer cells fixed at zero time or after 24 h-recovery, following 1 hour-long cyclic compression between 0 and 20.8 kPa.
(a) Control and compressed cell groups stained for actin (green) and nuclei (blue) for their form at zero time and 24 h-recovery after cyclic compression. Dashed areas are under piston, while the surrounding is control region (CR). Merged: merged form of the phalloidin (stain for actin) and Hoechst (stain for nuclei) epi-fluorescence images; merged with PH: merged form of the epi-fluorescence images with the corresponding phase-contrast (PH) image; magnified: two-hundred-fold magnification images of the region under micro-piston. Representative white arrows show distinct actin deformation in form of increased fluorescence signals at the edges of the cells in the compressed groups under the micro-piston at zero time after compression, while yellow arrows indicate no full recovery and thus an altered actin profile of cells at 24 h-recovery after compression. Several cells can be observed to have their actin disrupted over the nuclei and cytoplasm, with highly increased stress fibers present at the membrane. Nuclei deformation in control cells at control regions (CR) versus in compressed cells under piston at zero time or 24 h-recovery after cyclic compression, measured for area (b), circularity (c) and aspect ratio (d) of the nuclei (mean ± s.e.m. n = 406, 162, 342, 143 cell nuclei from at least two or three independent experiments per each group). Student’s t-test was used to determine the statistical significance of nuclear deformations in compressed and recovered cells after compression. Dashed horizontal bars show significant differences between the control region (CR) group and compressed group under micro-piston. Continuous horizontal bars show the comparison pairs between the cell groups under the micro-piston at zero time and 24-recovery after compression. The statistical significance levels among all groups are given in S2 File.