UC Berkeley

The recognition that cells can sense physical cues has inspired numerous investigations into the roles that mechanical forces play in both healthy and diseased cells. This rapidly growing area of research, often called cellular mechanobiology, has shown that physical interactions between cells and the surrounding extracellular matrix can regulate a number of fundamental cell behaviors. This insight has prompted the development of new methods to systematically engineer the mechanical properties of extracellular matrices (e.g., rigidity and geometry), both as a way to study the mechanisms behind cellular mechanosensing and as a way to directly control cell behavior in tissue engineering applications. In contrast, an equally powerful approach for manipulating cell-matrix interactions could be to directly engineer the mechanical properties of the cells themselves by modifying the intracellular signaling pathways that regulate how cells sense and respond to physical cues. This type of "inside-out" strategy would be useful for controlling cell behavior independently from extracellular matrix properties and would allow investigation into how changes in cellular mechanics such as cell shape, stiffness, and contractility can directly alter cell behavior.

In these dissertation studies, a genetic strategy was developed to precisely control the mechanical properties and motility of cells by manipulating the activity of cytoskeletal signaling proteins. Genetic mutants of the signaling proteins RhoA GTPase, Rac1 GTPase, or myosin light chain kinase (MLCK) were introduced into human glioblastoma cells under the control of conditional promoters, thereby enabling graded and dynamic control over their expression through addition and withdrawal of the transcriptional inducers. Increasing the activity of RhoA or MLCK by inducibly expressing constitutively active (CA) mutants increased both the stiffness and the contractility of cells in a graded manner, which had an inhibitory effect on cell migration. Interestingly, decreasing RhoA activity through expression of a dominant negative (DN) mutant also produced a graded decrease in cell migration speed, indicating that cell motility varies biphasically with RhoA activity levels. A similar biphasic dependence was discovered upon varying the activity of Rac1. These results demonstrate the importance of using quantitative methods to reveal potentially nonlinear relationships between protein activity and cell behavior.

Expanding upon this strategy, two orthogonal promoter systems were combined to provide simultaneous control over the activity of two proteins in the same cell. RhoA and Rac1 are known to suppress each other through crosstalk between their signaling pathways, suggesting that cells can normally have high activity of only one protein or the other. To investigate the effects of forcing high activation of both proteins, CA RhoA and CA Rac1 were introduced into the same cells under different conditional promoters (either doxycycline-inducible or cumate-inducible). Expression of CA RhoA did not alter Rac1 activity and vice versa, demonstrating that the activity levels of RhoA and Rac1 can be independently varied with this strategy. Notably, expressing both CA mutants had a greater inhibitory effect on cell migration than either mutant alone, indicating that the effects of these mutants were additive rather than suppressive.

Finally, these orthogonal promoters were used to dynamically control the motility of multiple cell populations in a three-dimensional matrix, providing a new way to spatially pattern cells. When cells were cultured as multicellular spheroids within a collagen matrix, CA Rac1 expression stimulated cell migration, while DN Rac1 expression strongly inhibited it. Thus the ability to switch these two phenotypes on and off by adding and withdrawing the transcriptional inducers provided a way to control both the timing and the extent of cell migration. To exploit this as a patterning method, cells expressing DN Rac1 from either the doxycycline-inducible promoter or the cumate-inducible promoter were mixed together as multicellular spheroids and then subjected to alternating administration of the two inducers. When the inducer was switched (e.g., doxycycline removed and replaced with cumate), one population was stimulated to migrate while the other was inhibited, and this created radially symmetric patterns of cells over time.

The strategies developed in these dissertation studies represent a novel method for tuning cellular mechanical properties and behaviors, which we expect will be useful in a number of tissue engineering applications. In addition, by enabling graded control over the activity of multiple proteins, these methods provide a unique opportunity to investigate the quantitative relationships describing how protein activity levels influence cell behavior. Given that cell and tissue mechanics have been discovered to play critical roles in a number of human diseases, a better understanding of these relationships may lead to new therapeutic targets for disease treatments.