Assembly and use of microfluidic devices to study nuclear mechanobiology in confined migration

Introduction:: During in vivo migration, metastatic tumor cells invade through extracellular environments and interstitial spaces about 1-2 μm in size, which is far smaller than the cell diameter and requires substantial deformation of the large and stiff nucleus (5-10 µm). The associated physical stress can lead to nuclear envelope rupture, DNA damage, and heterochromatin rearrangement [1, 2]. To date, most in vitro studies employ transwell assays, which pose difficulties in imaging and reproducibility, or microfluidic devices featuring continuously confining “channels” that do not mimic the intermittent and heterogeneous pores encountered in vivo. Here, we present two complementary microfluidic devices to study confined migration that precisely control pore size and geometry, mimic the heterogeneity of interstitial spaces, and are compatible with time-lapse imaging, with one of the devices additionally enabling the collection of thousands of cells after confined migration for downstream analysis such as RNA-seq or ATAC-seq.


Materials and Methods:: Our devices are replica molded from a silicon master mold fabricated using soft photolithography and/or reactive ion etching. The PDMS devices are then bonded to glass coverslips to create a 3D confined microenvironment and functionalized with collagen. Pillars at varying densities provide different degrees of confinement (Figure 1A) in the 5 µm tall migration area. Cells introduced into the seeding channel (Figure 1B, blue) migrate through the “field of pillars” (Figure 1B, red). Device A features a relatively large distance of pillars, such that migratory fitness can be assessed by quantifying distance traveled from seeding port into constriction area of fluorescently labeled cells. Device B features a shorter constriction distance such that cells exit into a second port, from which they can be collected for other experiments or characterization (Figure 1B, Device B, blue).


Results, Conclusions, and Discussions:: We designed two microfluidic devices to study confined migration with significant advantages over existing technologies. Device A enables the rapid assessment of migratory fitness of various cell lines or experimental conditions by live cell imaging in conditions closely mimicking physiological in vivo environments, whereas Device B additionally allows for the collection of cells that have completed migration through precisely controlled degrees of confinement for further functional, biochemical, or genomic analyses (Figure 1B). To validate the principle of our physiological “field of pillars” constriction design, we utilized lamin A/C-deficient (Lmna^–/–) mouse embryonic fibroblasts (MEFs), which have increased nuclear deformability and migrate more rapidly than wild-type (Lmna^+/+) MEFs [3]. Device A was used to assess distance traveled in confined conditions over several days (Figure 1C) and normalized to distance traveled in control conditions to account for differences in unconfined migration speed (Figure 1D). Lmna^–/– MEFs exhibited significantly faster migration than Lmna^+/+ cells, as measured by the increased distance traveled in device A, thus validating our constriction design and method. Comparison of the original cell population to cells that have completed migration through either confined conditions or unconfined conditions using RNA-seq analysis will reveal changes in gene expression in response to confined migration. This will provide insights into the short- and long-term changes induced by confined migration, and could provide clarity in the longstanding debate between selection and/or adaptation of cells to confined migration. Overall, these collective findings will not only be relevant to such as tumor cell infiltration and cancer metastasis, but also inform various physiological and pathological processes involving cell migration.

References (Optional): : [1] Shah et. al. 2021 Curr Biol. 31(4):753-765; [2] Golloshi et. al. 2022 EMBO Rep. 23(10):e52149; [3] Davidson et. al. 2014 CMBE 7, 293-306