Role of actin filaments in correlating nuclear shape and cell spreading

Abstract

It is well known that substrate properties like stiffness and adhesivity influence stem cell morphology and differentiation. Recent experiments show that cell morphology influences nuclear geometry and hence gene expression profile. The mechanism by which surface properties regulate cell and nuclear properties is only beginning to be understood. Direct transmission of forces as well as chemical signalling are involved in this process. Here, we investigate the formal aspect by studying the correlation between cell spreading and nuclear deformation using Mesenchymal stem cells under a wide variety of conditions. It is observed that a robust quantitative relation holds between the cell and nuclear projected areas, irrespective of how the cell area is modified or when various cytoskeletal or nuclear components are perturbed. By studying the role of actin stress fibers in compressing the nucleus we propose that nuclear compression by stress fibers can lead to enhanced cell spreading due to an interplay between elastic and adhesion factors. The significance of myosin-II in regulating this process is also explored. We demonstrate this effect using a simple technique to apply external compressive loads on the nucleus.

Figure 1. Variation of cell spreading, nuclear projected area and nuclear volume studied using gels of different stiffnesses.

(a) Cell and nuclear projected area as a function of Young's modulus of the substrate. Each point is an average taken over 100 cells. (b) Dependence of nuclear projected area on cell spreading obtained after putting all the data from all rigidities together and then binning the data points for cell spreading. Note, the difference in maximum spread area between the two figures arises due to this pooling and binning of data according to cell spread area. Binning size used was 26 and is calculated using the curve fitting toolbox, MATLAB. (c) Scatter plot (raw data) of the two areas of individual cells obtained from different substrates (same data as in a and b). Note that the range of measured cell area increases with substrate stiffness. (d) Nuclear volume as a function of the elastic modulus of the substrate measured from confocal stacks as describes in the text (20 cells for each data point). Error bars in all the plots represent mean standard error (SE).

Figure 2. Variation of cell and nuclear projected areas studied as a function of dynamic cell spreading.

(a) Cell area and nuclear projected area as a function of time for cells grown on plastic surfaces for up to 24 hrs. For each time point around 50–60 cells were imaged and the average values of cell spreading and nuclear projected area is plotted as a function of time. (b) Relation between the two areas obtained from the data shown in Fig. 2a. This plot is obtained by binning the cell areas of all 540 cells. Binning size used was 27 and is calculated using the curve fitting toolbox, MATLAB. Note, the difference in maximum spread area between the two figures arises due to this pooling and binning of data according to cell spread area. The smooth line is the fit from Fig. 1 given for the sake of comparison. The error bars in (a) and (b) are mean SE.

Figure 3. Plot of the cell and nuclear areas for three different cell types obtained using substrates with different stiffness.

For each substrate stiffness, nearly 100 cells were imaged for each type of cell. The data from all the substrates for, each type of cell, is pooled together and plotted after binning as done previously for mMSCs. The error bars here represents SE. The correlated behavior between cell and nuclear projected area seem to be roughly intact even across these different cell types although the nuclear projected areas is reaching saturation at different values.

Figure 4. The relation between cell area and nuclear projected area during trypsin mediated deadhesion.

(a, b) Changes in cell spreading and nuclear projected area (normalized values) as a function of time, obtained from individual cells (different symbols). Normalization is done using the formula . (c) Variation in nuclear projected area as a function of cell spreading for the same cells. The line is the same fit as in Fig. 1b, and is plotted for comparison. In some cases nuclear area shows an undershoot where the area decreases below the final value as seen in (b). Moreover, in some cases cell shrinkage precedes nuclear shrinkage as can be seen in (c).

Figure 5. The robustness of the correlated behavior was tested against various cytoskeletal drug treatments.

(a) Changes in cell and nuclear projected areas after microtubule disruption using Nocodazole and actin depolymerisation using Latrunculin-A. (b, c) Variation in the two areas after treatment with blebbistatin to deactivate myosin-II. Note that substrate sensitivity is significantly diminished after mysosin-II inhibition. Further, for all drug treatments, a change in cell area causes a correlated change in nuclear area. All the values for the cell area and nuclear projected area are taken from over 80–100 cells in each case and the error bars in the plots represent standard error. Student t-test values are and the comparison is with control for (a) and with 65 kPa substrate for (b).

Figure 6. Changes in nuclear projected area plotted against cell spreading follow a Master Curve irrespective of the way cell spreading is altered or after perturbing different cytoskeletal components.

(a) Nuclear projected area vs. cell spreading obtained after different perturbation experiments. The solid line is the fit obtained from Fig. 1b. (b) and 5(c). Data points from nuclear compression experiments show the change in the two areas with varying applied load as discussed in the text (30 cells were imaged outside the lens for zero load, 15 for 0.97 g lens, and 20 for 1.07 g lens). Error bars are mean SE. (b) Residue plot for the data in (a). The residue plot shows the extend of deviations of the data obtained after perturbations from the fit obtained for the unperturbed cells. The agreement is remarkable considering the fact that perturbation experiments, although performed using specific drugs, lead to global reorganization of cellular components.

Figure 7. Confocal images showing the actin distribution and nuclear shape in cells spread on gels of two different stiffnesses.

(a, b, c) Confocal images of a cell on a 70 kPa stiffness substrate. The images show actin filaments close to the plane of the substrate, in an approximate mid plane, and just above the nucleus respectively. (d, e) The nucleus of the same cell projected in the plane of the substrate and in a perpendicular plane respectively. (f–j) Similar observations of a cell spread on a soft substrate (3 kPa). Note the difference in the nuclear shape compared to the upper set. (k, l) 3D reconstruction of confocal images showing perinuclear stress fibers running over the nucleus in the case of the first cell (stiff substrate) and a predominantly cortical actin mesh in the case of the second cell (soft substrate). Images in k and l are 3D reconstructions of the cells shown in a–e and f–j respectively.

Figure 8. Changes in the perinuclear actin stress fibers and nuclear geometry for different rigidities.

(a–d) The observed decrease in stress fiber density as function of substrate rigidity for substrates with elastic modulus 65 kPa, 23 kPa, 5 kPa and 3 kPa respectively (also see panel in Fig. S11). No stress fibers were observed in cell cultured on substrate with elastic modulus 3 kPa. (e–h) The transverse view (3D projections) of the nucleus under different compressive loading for the cells grown on substrates with different rigidities.

Figure 9. Confocal slices of cells after nuclear perturbations (control, TSA treatment and lamin a/c knock down).

All the cells were cultured on fibronectin coated coverslips. Actin filaments, nucleus and focal adhesions are labelled in red, blue and green color respectively. It can be seen that perinuclear stress fibers are lost and the nucleus bulges out in treated cells.

Figure 10. Diagrams showing the scheme used for the estimation of the compressive loading of the nucleus by stress fibers.

(a) Schematic showing nuclear deformation under uniform loading. The undeformed nucleus has a radius . After loading the contact area with the plate is and the nuclear height is . (b) Schematic showing how normal stresses arise due to a perinuclear stress fiber.

Figure 11. Nuclear compression experiment.

(a) Schematic of the arrangement used to apply an external compressive load on the nucleus. The pressure exerted on the cells is maximum at the central point determined by observing the interference pattern in reflected light (Newton's rings) as shown in (b). A 4x objective and 200 mm focal length lens were used for this image for better illustration of the pattern. (c–e) Phase contrast images showing the effect of a compressive load on cells applied using a convex lens (c) Cells outside the lens and (d,e) cells under lenses of two different weights, 0.97 g and 1.07 g respectively. The slight reduction in image quality is due to inclusion of the lens. An increase in cell spreading and nuclear projected area was observed with increasing load, shown quantitatively in master plot Fig. 6.

Figure 12. Schematic showing how nuclear compression helps in cell spreading.

Flattening the nucleus by perinuclear stress fibers (also see [12]), or an external load as in the case of lens experiment, allows the cell to spread to a greater extent. In the absence of such a compressive loading of the nucleus, the nucleus exerts an upwards force on the cell cortex which constraints cell spreading due to balance of adhesive and elastic forces as elaborated in the text.