3D full-field quantification of cell-induced large deformations in fibrillar biomaterials by combining non-rigid image registration with label-free second harmonic generation

Abstract

To advance our current understanding of cell-matrix mechanics and its importance for biomaterials development, advanced three-dimensional (3D) measurement techniques are necessary. Cell-induced deformations of the surrounding matrix are commonly derived from the displacement of embedded fiducial markers, as part of traction force microscopy (TFM) procedures. However, these fluorescent markers may alter the mechanical properties of the matrix or can be taken up by the embedded cells, and therefore influence cellular behavior and fate. In addition, the currently developed methods for calculating cell-induced deformations are generally limited to relatively small deformations, with displacement magnitudes and strains typically of the order of a few microns and less than 10% respectively. Yet, large, complex deformation fields can be expected from cells exerting tractions in fibrillar biomaterials, like collagen. To circumvent these hurdles, we present a technique for the 3D full-field quantification of large cell-generated deformations in collagen, without the need of fiducial markers. We applied non-rigid, Free Form Deformation (FFD)-based image registration to compute full-field displacements induced by MRC-5 human lung fibroblasts in a collagen type I hydrogel by solely relying on second harmonic generation (SHG) from the collagen fibrils. By executing comparative experiments, we show that comparable displacement fields can be derived from both fibrils and fluorescent beads. SHG-based fibril imaging can circumvent all described disadvantages of using fiducial markers. This approach allows measuring 3D full-field deformations under large displacement (of the order of 10 μm) and strain regimes (up to 40%). As such, it holds great promise for the study of large cell-induced deformations as an inherent component of cell-biomaterial interactions and cell-mediated biomaterial remodeling.

Fast quantitative time lapse displacement imaging of endothelial cell invasion

“…Finally, we tailor displacement microscopy algorithms to calculate both absolute and incremental displacements ( Fig 1C and 1D) of the collagen matrix. Due to its capability to deal with large distortions in complex materials [51], Free Form Deformation-based non-rigid image registration is used for both approaches to compare images of the fluorescent beads embedded in the collagen and extract the corresponding matrix displacement fields. Absolute displacements (t ref = t fix ) result from the registration and comparison of the image stack of fluorescent beads acquired under cell activity at each time point t i against the (fixed reference) image stack acquired in a stress-free state of the hydrogel at timepoint t fix .…”

Combustion-derived particles inhibit in vitro human lung fibroblast-mediated matrix remodeling

“…The effect of CB particles on the displacements generated by cells was evaluated using cell-induced displacement microscopy according to a method described by us previously [ 30 ].…”

“…However, results from this type of assay lack crucial information such as changes in the interplay at ECM—cellular level and, moreover, the general result is biased by the proliferation rate of the embedded cells. Here, we studied cell-mediated matrix remodeling at the cellular level under normal and exposed conditions using 3D CDM [ 30 ]. As far as we know, CDM is employed for the first time to study toxicological effects of (nano)particles.…”

TFMLAB: a MATLAB toolbox for 4D traction force microscopy

“…For flexibility, TFMLAB uses the Bio-Formats toolbox which can read multiple microscopy file formats [33]. From the input file, the software reads the channels corresponding to the cells (e.g., imaged after transduction with LifeAct for filamentous actin staining) and the fiduciary markers such as fluorescent beads and/or fibers (e.g., in case of a fibrillar collagen hydrogel imaged by means of second harmonic generation [27]). Any other channels (e.g.…”

“…Current research includes both beadbased TFM [20], [46], [47] and fiber-based TFM (with fiber networks e.g. being imaged by means of label-free techniques such as second harmonic generation or confocal reflection microscopy) [23], [27], [48]. Our software includes specific image filters for both markers and the FFD algorithm handles both of them [27], [30].…”