Driving Cells with Light-Controlled Topographies

Abstract

Cell-substrate interactions can modulate cellular behaviors in a variety of biological contexts, including development and disease. Light-responsive materials have been recently proposed to engineer active substrates with programmable topographies directing cell adhesion, migration, and differentiation. However, current approaches are affected by either fabrication complexity, limitations in the extent of mechanical stimuli, lack of full spatio-temporal control, or ease of use. Here, a platform exploiting light to plastically deform micropatterned polymeric substrates is presented. Topographic changes with remarkable relief depths in the micron range are induced in parallel, by illuminating the sample at once, without using raster scanners. In few tens of seconds, complex topographies are instructed on demand, with arbitrary spatial distributions over a wide range of spatial and temporal scales. Proof-of-concept data on breast cancer cells and normal kidney epithelial cells are presented. Both cell types adhere and proliferate on substrates without appreciable cell damage upon light-induced substrate deformations. User-provided mechanical stimulation aligns and guides cancer cells along the local deformation direction and constrains epithelial colony growth by biasing cell division orientation. This approach is easy to implement on general-purpose optical microscopy systems and suitable for use in cell biology in a wide variety of applications.

Keywords: cell migration; cell orientation; cell‐instructive substrates; light‐responsive polymers; optical manipulation.

Figure 1

Platform concept. a) Laser projection system coupled to a commercial inverted microscope for life sciences. Two computer‐controlled elements are herein introduced: a phase‐only liquid crystal‐based SLM to dynamically generate arbitrary laser patterns to be projected onto the micropillar arrays, and a half‐wave plate mounted on a motorized rotational stage for rotating the laser linear polarization according to any desired orientation. Proper opto‐mechanics is employed for alignment purposes. b–d) Representative deformation patterns of an array of azopolymeric micropillars obtained upon irradiation with a laser pattern having hyperbolic polarization, b) with a 50 s exposure time, c) doughnut‐shaped azimuthal polarization projected, with a 90 s exposure time, and d) spatially dependent linear polarization states projected, with a 90 s exposure time. In the latter case, the pattern is obtained by superposing a time‐sequence of two complementary spatial patterns, each one having a defined linear polarization state as synchronously set by the motorized half‐wave plate. Objective magnification is 20×. e–h) Sequence of operations illustrating the conceptual work‐flow of a prototypical use of the proposed platform. e) First, an initial image of the sample is captured, wherein one or more target cells are included. Here, a single cell is imaged in bright field with a 20× objective. Then, a processing stage lets the user defining a proper mask to be used as an illumination pattern. f) In alternative, an automatic segmentation algorithm can identify the object(s) outline(s) and produce a corresponding binary mask. g) The binary mask is fed into the SLM, in order to produce a corresponding laser pattern (e.g., conformal to the target cell in the case shown here) with a desired polarization state which is ruling the deformation direction(s) on the pillars. Here, the polarization is uniform over the whole irradiated area and oriented as indicated by the black arrows. h) Finally, the deformation is produced on the micropillars and directly observed on the microscope camera, in live capturing (see Movie M5, Supporting Information). The laser exposure time can be either fixed a priori or defined by the user, depending on the specific application case. This work‐flow can be iterated at any time, and made adaptive to the specific evolution of the observed target cells distribution.

Figure 2

Adhesion and migration of cancer cells on deformed substrates. a) Snapshots of fixed MDA‐MB‐231 seeded on deformed substrates, stained for polymerized actin (phalloidin, green), nuclei (dapi, blue or DRAQ5, red), and pillars (auto‐fluorescence, red). Cells are oriented along the direction of deformation of the substrate and show aligned stress fibers. b) Sketch of the conventional directions used in the figure. The angle φ between the local direction of the deformed pattern and the direction of the major axis of the cell is used as a measure of alignment. c) Quantification of cell polarization and orientation on deformed substrates corresponding to panel (a) (with local angle indicated by blue arrows). Each cell is depicted here as an arrow which length is proportional to the ratio between major and minor axis lengths, and which orientation is that of the major axis. Green arrows indicate cells sitting outside the deformed region, i.e., undeformed pattern, while arrows colored from blue to magenta indicate cells sitting onto the deformed region. The color‐map indicates the absolute value of the cosine of the angle φ. Alignment of cells is qualitatively clear. d) Quantification of cell alignment over deformed versus undeformed substrate shown as a violin plot. Continuous lines represent the probability distribution of the angle. Kolmogorov–Smirnov 49 statistical test yields rejection of the null hypothesis with a p‐value of 5.910⁻¹⁴. e) Tracks of cells migrating on deformed substrates are shown to illustrate directed migration. The orientation of cell displacements in each trajectories (corresponding to segments of 1 h) was projected onto the local orientation of the pattern by calculating the cosine of the angle φ. The color‐map indicates the degree of alignment. The deformation pattern is shown by blue arrows while each line is a cell trajectory. f) Statistical distribution of displacement orientation on deformed versus undeformed patterns shown as a violin plot. Continuous lines represent the distribution of the angle. Kolmogorov–Smirnov statistical test yields rejection of the null hypothesis with a p‐value of 6.910⁻¹⁸⁵.

Figure 3

Substrate topography can bias the growth of epithelial colonies. a) Microscopy image showing LifeAct‐Ruby/H2B‐GFP MDCK cells cultured on patterned substrates. b) Quantification of cell division positioning with respect to deformation of the pattern. The top panel represents a dividing cell, and the definition of the deviation between the orientation of cell division (green arrow) and the orientation of the deformation (red arrow). The lower panel compares the distributions of angles on deformed versus undeformed patterns. Statistical significance was assessed by means of a Kolmogorov–Smirnov test, yielding rejection of the null hypothesis with a p‐value of 310⁻³. c,d) Colony outlines as a function of time showing time progression of the colony, alongside with corresponding quantification of the elongation and orientation on undeformed pattern. The outline was extracted by segmenting the LifeAct‐Ruby signal. e–h) Same as previous panels, except with linear deformation. For all these four panels, the orientation of the pattern is indicated by the dashed black line.