Cells interact with structurally and biochemically distinct types of extracellular matrix in different tissues, at different stages of embryonic development, and during adult wound repair. This project focuses on addressing the following major questions concerning the mechanisms of these cell-extracellular matrix interactions: 1. What unique mechanisms do mammalian cells use to migrate through three-dimensional (3D) extracellular matrix environments rather than on flat culture surfaces? 2. What distinct signal transduction mechanisms control cell behavior in different 3D microenvironments? We are exploring whether classical models of cell motility and signaling established using regular 2D cell culture are valid in the structurally complex 3D environments found in tissues. We discovered a unique mode of 3D migration using high-resolution live-cell imaging to visualize intracellular signaling using different in vitro models of 3D extracellular matrix. Primary dermal fibroblasts migrating in dermal tissue explants and cell-derived matrix were found to use blunt, cylindrical protrusions termed lobopodia. In contrast, cells migrating in 3D collagen gels exhibited lamellipodia-based migration similar to 2D cell culture, with small, fan-shaped protrusions enriched in F-actin at the leading edge. The leading-edge lobopodia of primary human fibroblasts in 3D matrix lack both classical lamellipodial markers and polarized signaling, and instead require actomyosin contractility and a 3D matrix microenvironment that is linearly elastic. We now find that instead of the classical polymerization of actin to extend the leading edge of migrating cells, it is intracellular pressure that provides the motive force to push the leading edge forward in a physiological 3D matrix. Unexpectedly, actomyosin (via myosin II) contractility pulls the nucleus forward like a piston to elevate intracellular pressure in the forward cytoplasmic compartment to drive lobopodial protrusion. We have identified a contractility-dependent protein complex between the intermediate filament protein vimentin, actin, myosin II, and the nucleoskeleton-cytoskeleton linker protein nesprin 3. Locally inhibiting myosin II in front of the nucleus results in rearward recoil of the nucleus with substantially reduced anterior pressure, and the leading edge retracts. Applying the same drug (blebbistatin) behind the nucleus has no such effects. Knocking-down nesprin 3 detaches the nucleus from the actomyosin pulling machinery and results in reduced and equalized intracellular pressure. Importantly, both global inhibition of myosin II and knock-down of the nesprin 3 connector protein reduce the overall speed of cell migration. Primary human intestinal myofibroblasts and chondrocytes also both exhibit this novel phenomenon of compartmentalized pressure in lobopodia when placed in 3D matrix, suggesting that the nuclear piston mechanism may be a widespread, central feature of cell movement in vivo. We conclude that this new mechanism of cell migration uses actomyosin machinery in front of the nucleus to pull the nucleus forward, pressurizing the forward cytoplasmic compartment and extending lobopodial protrusions at the leading edge in a process necessary for efficient migration of primary human cells in a physiological 3D matrix. We are currently collaborating with Michael Davidson to develop a new suite of FRET-based biosensors to visualize the connections linking the extracellular matrix to the nucleus that are predicted by our nuclear piston model. The biochemical composition of the extracellular matrix can also regulate 3D cell migration. Although Rho-family GTPases (Cdc42, Rac1, and RhoA) regulate cell migration, it was not known how they are differentially regulated when cells interact with different extracellular matrix ligands. We hypothesized that adhesion to different matrix molecules, such as collagen and fibronectin, would trigger differential regulation of guanine nucleotide exchange factors (GEFs), molecules that activate the Rho GTPases to regulate migration. We first used an affinity precipitation-based, mass spectrometry screen to identify a central role for collagen-specific regulation of the GEF beta-Pix. Depletion of beta-Pix dramatically inhibited cell motility in 3D or 2D fibrillar collagen matrices, but had no effect on migration on 2D fibronectin or in 3D cell-derived matrix, thereby establishing the first evidence for matrix-specific regulation of Rho GTPase activity during 3D cell migration. Further mechanistic analyses unexpectedly revealed beta-Pix regulation of crosstalk between Cdc42 and RhoA through the GTPase activating protein (GAP) srGAP1. As summarized below, we have validated this Cdc42/RhoA crosstalk during collagen migration using FRET microscopy, identified a key phosphorylation site on beta-Pix needed for its collagen-specific function, discovered that beta-Pix function downstream of adhesion to fibrillar collagen is regulated by alpha5-beta1 integrin and the phosphatase PP2A, and demonstrated that this beta-Pix/Cdc42/srGAP1 pathway is conserved across diverse cell types. FRET biosensors for Cdc42 or RhoA activity permitted direct visualization of spatial changes in GTPase activity during migration. Knockdown of beta-Pix during migration on fibrillar collagen led to a loss of leading-edge polarization and decreased overall Cdc42 activity, yet had no effects when the cells were on fibronectin. Moreover, polarization of RhoA activity toward the rear of the cell was lost with global elevation of RhoA activity on fibrillar collagen, but not on fibronectin. Phospho-proteomics on beta-Pix isolated from fibroblasts migrating on fibronectin versus fibrillar collagen revealed differential de-phosphorylation of the amino acid residue T526 on collagen. To test the role of this residue, we generated beta-Pix knockdown/rescue fibroblasts that expressed either a phospho-mimetic (T526E) or a phospho-null (T526A) variant of beta-Pix. Phosphorylation at T526 blocked morphological and migratory rescue in 3D collagen. Mechanistically, phosphorylation at T526 blocked the association between beta-Pix and srGAP1, but not Cdc42. We discovered a collagen-specific association between beta-Pix and a regulatory subunit of protein phosphatase 2 (PPP2R1A) using mass spectrometry. PP2A was found to mediate the loss of threonine phosphorylation on beta-Pix at T526 in response to collagen, and knockdown of PPP2R1A phenocopied loss of beta-Pix/Cdc42/srGAP1, identifying a critical regulatory role for integrin regulation of PP2A in this migratory signaling cascade. This matrix-specific beta-Pix migratory pathway was found to be conserved in multiple cell types, including fibroblast lines, primary human osteoblasts, smooth muscle and endothelial cells, and tumor cells. Adenocarcinoma cell migration was inhibited by beta-Pix knockdown and was not rescued by the beta-Pix T526E mutant, suggesting a possible specific therapeutic target. These studies represent the first detailed mechanistic characterization of ECM-specific regulation of Rho GTPase activity during migration used by diverse cell types, and it involves a novel interaction between a GEF and a GAP protein to mediate signaling crosstalk between Cdc42 and RhoA.
Showing the most recent 10 out of 23 publications
