Cells Under Stress: An Inertial-Shear Microfluidic Determination of Cell Behavior

Abstract

The deformability of a cell is the direct result of a complex interplay between the different constituent elements at the subcellular level, coupling a wide range of mechanical responses at different length scales. Changes to the structure of these components can also alter cell phenotype, which points to the critical importance of cell mechanoresponse for diagnostic applications. The response to mechanical stress depends strongly on the forces experienced by the cell. Here, we use cell deformability in both shear-dominant and inertia-dominant microfluidic flow regimes to probe different aspects of the cell structure. In the inertial regime, we follow cellular response from (visco-)elastic through plastic deformation to cell structural failure and show a significant drop in cell viability for shear stresses >11.8 kN/m². Comparatively, a shear-dominant regime requires lower applied stresses to achieve higher cell strains. From this regime, deformation traces as a function of time contain a rich source of information including maximal strain, elastic modulus, and cell relaxation times and thus provide a number of markers for distinguishing cell types and potential disease progression. These results emphasize the benefit of multiple parameter determination for improving detection and will ultimately lead to improved accuracy for diagnosis. We present results for leukemia cells (HL60) as a model circulatory cell as well as for a colorectal cancer cell line, SW480, derived from primary adenocarcinoma (Dukes stage B). SW480 were also treated with the actin-disrupting drug latrunculin A to test the sensitivity of flow regimes to the cytoskeleton. We show that the shear regime is more sensitive to cytoskeletal changes and that large strains in the inertial regime cannot resolve changes to the actin cytoskeleton.

Figure 1

(a) Schematic of the cross-flow region. (b) Parameters extracted from high-speed videos of cell deformation are shown: A is the initial diameter of the cell before it deforms, H is the height of the cell, W is the width of the cell, and l is the perimeter of the cell. To see this figure in color, go online.

Figure 2

Deformation index, DI ± standard error of HL60 cells versus flow rate, at μ = 1 cP. DI ± standard error was averaged from multiple cell events combined from N = 3 repeats; each data point includes 30 > n > 500 cell events. For $Q<400\mathit{\mu }\text{L}/\text{min}$, deformation can be fitted by an exponential, which tends toward a maximal deformation of DI_max. To see this figure in color, go online.

Figure 3

(a) Deformation index versus Q for HL60 cells in four different media with viscosity changing between 1 and 33 cP. DI ± standard error was averaged from multiple cell events combined from N = 3 repeats; each data point includes 30 > n > 500 cell events. The data is fitted with an exponential. (b) Images of a cell deformation for each flow condition where $DI\cong D{I}_{\max }$are shown. They are accompanied by superimposed color contour plots that show how the deformation changes as a function of time. To see this figure in color, go online.

Figure 4

(a) Strain, ε, as a function of time, averaged over 50 cells, with the standard error shown in gray. Q was fixed at 5 μL/min, and the suspension medium viscosity was 33 cP. The exponential fits shown in red were used to quantify the deformation and relaxation of the cells. (b) A superimposed brightfield image of a cell as it deforms and relaxes at 5 μL/min (μ=33 cP) is shown. Scale bars, 30 μm. The arrows indicate the direction of cell motion. To see this figure in color, go online.

Figure 5

(a) The DI as a function of flow rate, Q, of HL60 cells, SW480 cells, and SW480 cells treated with 1 μM of LatA. The flow regime was shear dominant, and the viscosity of the cell suspension buffer was 33 cP. (b) DI versus flow rate Q for HL60 cells, SW480 cells, and SW480 cells treated with 1 μM of LatA. The flow regime was inertia dominant, the viscosity of the cell suspension buffer was ∼1 cP. To see this figure in color, go online.

Figure 6

Strain ε was tracked for SW480 (N = 56) and SW480 treated with LatA (N = 30) as a function of time, with the standard error shown. Q was fixed at 5 μL/min, and the suspension medium viscosity was 33 cP. The dashed lines represent the extrapolated final strain ${\epsilon }_{\infty }$ for both samples. To see this figure in color, go online.