A Critical Pull to Polarize the Cell

Cell polarization in response to chemical or mechanical cues is crucial to diverse important biological processes such as gastrulation, directed cell migration, epithelial tissue function, and neuronal growth-cone migration. Considering the complex chemical and mechanical landscapes in which cells are often embedded, determining the minimal cues required to polarize cells is an important conceptual step toward a broader mechanistic understanding. Although some details of cell polarization in response to chemical cues have been revealed in recent years, little is known about how the cell polarizes in response to mechanical cues. We do know that focal adhesions (FA) are integrin-based adhesions through which the cell binds and senses extracellular matrix (ECM) rigidity, allowing the cell to migrate toward stiffer ECMs in a process known as “durotaxis” (1).

Careful in vitro analysis in two-dimensional cell culture systems has demonstrated that external force on a single integrin-ECM complex can promote local FA maturation through recruitment of vinculin, which connects the contractile actomyosin cytoskeleton to the FA complex for mechanosensation (2). Furthermore, high-resolution traction-force microscopy has recently revealed how cells sample local ECM stiffness during durotaxis by tugging on the ECM via individual FAs, resulting in differential focal adhesion kinase (FAK)/phosphopaxillin/vinculin signaling dependent on the adhesion-ECM stability (3). These findings suggest that cells may sense even very small, localized forces and exhibit larger-scale responses such as directed migration.

Although these studies have begun to reveal underlying mechanisms for cellular mechanosensation in durotaxis, they have not addressed how the cell uses mechanosensation to spatially polarize and migrate during durotaxis. Cell polarization in response to rigidity gradient has been difficult to understand partly because of the nature of the cellular interaction with the ECM, which involves numerous FAs throughout its dorsal-ventral surface for two- and three-dimensional interactions with the ECM, thus making it unclear whether an ensemble of FAs or a single FA interaction with ECM is sufficient for a cell to sense and polarize toward a mechanical cue. An elegant study by Bun et al. (4) in this issue of the Biophysical Journal reports a key finding toward answering this important biological question. The authors have used optical trapping of ECM-coated beads and 3T3 fibroblasts in nonadherent conditions to test whether a single FA interaction with ECM is sufficient for a cell to polarize. The authors show that nonadhering cells exhibit spontaneous shape oscillations, but activation of a single FA by a fibronectin (FN)-coated bead coupled with a mechanical load by optical trap is sufficient to abolish shape oscillation and induce morphological polarization toward the trapped bead. A key feature of this polarity system is that once triggered above a critical force, the cell could sustain its morphological polarity even after removal of the force.

Furthermore, this mechanical cue is sufficient to turn-on a series of intracellular reorganization throughout this polarization process, including myosin-II-dependent redistribution of cortical actomyosin to the posterior-end of the cell with respect to the trapped bead, centrosome polarization in front of the nucleus toward the bead, and subsequent microtubule-dependent cell protrusion from underneath the trapped bead (Fig. 1). Cell polarization events observed here share the features of early embryonic polarization and cell migration on a two-dimensional culture. Cortical redistribution of actomyosin closely resembles Caenorhabditis elegans embryo fertilization where cortical actomyosin flows toward the opposite end of the sperm entry point in a PAR-protein-dependent manner (5), while centrosome polarization toward the leading edge in front of the nucleus is typical of polarized fibroblast cells (6).

Figure 1.

Cell polarization under constant mechanical load above the critical limit. (f) Mechanical load; (arrow) direction of force. The cell undergoes rapid morphological polarization with concomitant cortical actomyosin redistribution at the posterior-end and centrosome polarization toward the anterior-end. At the later stage of polarization, microtubule-dependent membrane protrusion pushes the bead away from the optical trap. To see this figure in color, go online.

The only stimulus required to induce this cell polarization process is the mechanical pulling on a single FN-coated bead, which mimics, basically, a single focal adhesion of a two-dimensional crawling cell. The existence of such a mechanical trigger leads to the possibility of having a signaling switch that could only be turned on above a minimum, localized mechanical load. One exciting possibility is the selective activation of the integrin receptors associated with the FN-bead, and concomitant activation of downstream events such as activation of FAK/phosphopaxillin/vinculin signaling (3). In support of this notion, it was also shown that localized activation of integrin could selectively activate FAK, leading to microtubule (MT) stabilization at the leading edge of a polarized fibroblast cell (7).

Downstream from integrin activation, the study by Bun et al. (4) raises several interesting questions for future analysis: Do similar localized mechanical checkpoints operate in other forms of cell polarization, such as apical-basolateral polarization in epithelia? How does pulling on a fibronectin bead create local reduction of actomyosin contractility, such that the actomyosin cortex redistributes away from the localized load? What downstream signaling pathway regulates centrosome polarization? And how do the cell protrusions emanate from the pulled cortex in the later stages of polarization?

The authors suggest that centrosome polarization could be driven by myosin-II-mediated pulling on MT plus-ends at the cortex (4). However, because actomyosin redistribution occurs toward the posterior-end of the cell, it is unclear how pulling MT plus-ends toward the posterior-end could polarize a centrosome toward the anterior cortex (toward the bead).

One may envision that at least two alternative mechanisms could cause centrosome polarization:

• 1.Pushing the nucleus back by myosin-mediated actin retrograde flow, passively leaving the centrosome polarized toward the leading edge (6); and
• 2.Asymmetric localization of the MT minus-end-directed motor dynein at the bead cortex and successive pulling on MT plus-ends to polarize centrosome toward the bead, which has been previously observed in centrosome polarization events (8).

Finally, how the cell protrudes toward the applied load in the later stages of polarization may be explained by localized actin polymerization induced by actin nucleation factors. Arp2/3 complex-mediated actin polymerization pushes on plasma membrane for the leading-edge protrusion and cell migration (9), and one mechanism for its activation is through signals downstream of Rac1 and integrin engagement (10). It seems reasonable that Arp2/3 complex is also a key player for the observed cell protrusion, given the gradual actin accumulation at the optically trapped bead cortex over time (4). However, formin-type actin nucleators have also been shown to be activated by mechanical force (11), and may also contribute locally by polymerizing actin to initiate protrusion. In sum, the study by Bun et al. provides an elegant illustration for polarized activation of myriad signaling and cytoskeletal rearrangements from just a single, localized pull on the cell.