Assistant Professor National University of Singapore, United States
Introduction:: Cells in the body are continually exposed to differential levels of confinement as a function of their tissue niche, their migratory activities, or the application of external force. As the extracellular niche has been shown to become dysregulated as a function of disease or aging, it follows that the mechanical confinement and compression cells are exposed to will also be altered. Thus, understanding the cellular responses associated with different levels of confinement, and how to modulate them, is a key goal for mechanobiologists. Recent attention has been given to the biophysics of LLPS in the formation of cellular biomolecular condensates, especially given the context that many condensates form in response to cellular stress. This suggests that LLPS plays a key role in both maintaining homeostasis and cellular dysregulation. While biomolecular condensates have been stimulated in vitro via cell stressors (UV light, chemical damage) or the direct alteration of cell volume (osmotic shock), the role of mechanobiology (especially in physiological force environments) in LLPS and condensation has not been investigated. To address this, we have examined a variety of biomolecular condensates in the context of confined cell migration, including nuclear paraspeckles, Cajal bodies, and DNA damage repair complexes as well as cytoplasmic stress granules.
Materials and Methods:: Two-step photolithography and replica molding were used to produce pumpless microfluidic chips containing microchannels with widths between 3-20 µm, heights of 10 µm, and lengths between 150-500 µm. MDA-MB-231 breast cancer cells and HT1080 fibrosarcoma cells labelled with live-cell markers for a variety of biomolecular condensates including paraspeckles (SFPQ, NONO), DNA damage repair complexes (53BP1), stress granules (G3BP1), and Cajal bodies (coilin) were produced and introduced into the microchannel confinement assays. Live cell microscopy was performed to visualize the formation and dissolution dynamics of these biomolecular condensates as a function of confinement state. In the case of paraspeckles, fluorescence in situ hybridization (FISH) was conducted to observe the localization of long non-coding RNA (lncRNA) paraspeckle scaffolds. Changes in nuclear volume were quantified with both confocal Z-stacks and the development of a novel method of double fluorescence exclusion microscopy specific to microchannel environments.
Results, Conclusions, and Discussions:: We first visualized the dynamics of paraspeckles, stress-responsive subnuclear bodies that form by phase separation around the long non-coding RNA NEAT1. As cells entered moderate confinement, a significant increase in paraspeckle number and size was observed compared to unconfined cells. This increase was found to be quickly reversible upon cellular exit from microchannels, suggesting that the stimulus for condensate formation is the nuclear adaptation to specific levels of confinement. Paraspeckle polarization bias towards the leading edge was also observed in confinement, correlating with regions of euchromatin. Increasing paraspeckle abundance resulted in increases in confined migration likelihood, speed, and directionality, as well as an upregulation in invasion-related genes and an enhancement of paraspeckle polarization towards the leading edge. This polarization of paraspeckle condensates may play a key role in regulating confined migration and invasion in cancer cells, and illustrates the utility of microchannel-based assays for identifying phenomena not observed on 2D or 3D bulk substrates.
We next went beyond paraspeckles to characterize the formation and dynamics of a wider variety of biomolecular condensates, including stress granules, Cajal bodies, and DNA damage repair complexes. Intriguingly, by comparing the colocalization of 53BP1 puncta (a biomolecular condensate associated with DNA damage) with γH2AX puncta (a protein found to directly interact with double-stranded DNA breaks), we found that in confinement a mismatch between 53BP1 and γH2AX localization occurs. This suggests that 53BP1 condensates, which are often used as a specific marker for DNA damage, may be forming independently of DNA damage in specific confining microenvironments.
By identifying mechanical inputs that drive LLPS, we aim to open a new mechanobiology paradigm aimed at understanding, altering, and leveraging biophysical control of diverse cellular condensates.
