Integrins as mechanochemical transducers.

A recent resurgence of interest in mechanical forces and cell shape as biological regulators has revealed extracellular matrix as the site at which forces are transmitted both to and from cells. at the same time, great advances have been made in terms of defining cell-surface integrin receptors as transmembrane molecules that mediate cell attachment and physically interlink extracellular matrix with the intracellular cytoskeleton. Convergence of these two lines of research has begun to elucidate the molecular mechanism by which cells sense physical forces and transduce mechanical signals into a biochemical response.

Cellular control lies in the balance of forces.

Mechanical tension generated within the cytoskeleton of living cells is emerging as a critical regulator of biological function in diverse situations ranging from the control of chromosome movement to the morphogenesis of the vertebrate brain. In this article, we review recent advances that have been made in terms of understanding how cells generate, transmit and sense mechanical tension, as well as how they use these forces to control their shape and behavior. An integrated view of cell regulation that incorporates mechanics and structure as well as chemistry is beginning to emerge.

The structural and mechanical complexity of cell-growth control.

Tight control of cell proliferation is required to ensure normal tissue patterning and prevent cancer formation. The analysis of cultured cells has led to an explosion in our understanding of the molecules that trigger growth and mediate cell-cycle progression. However, the mechanism by which the local growth differentials that drive morphogenesis are established and maintained still remains unknown. Here we review recent work that reveals the importance of cell binding to the extracellular matrix, and associated changes in cell shape and cytoskeletal tension, to the spatial control of cell-cycle progression. These findings change the paradigm of cell-growth control, by placing our understanding of molecular signalling cascades in the context of the structural and mechanical complexity of living tissues.

Mechanical control of cyclic AMP signalling and gene transcription through integrins.

This study was carried out to discriminate between two alternative hypotheses as to how cells sense mechanical forces and transduce them into changes in gene transcription. Do cells sense mechanical signals through generalized membrane distortion or through specific transmembrane receptors, such as integrins? Here we show that mechanical stresses applied to the cell surface alter the cyclic AMP signalling cascade and downstream gene transcription by modulating local release of signals generated by activated integrin receptors in a G-protein-dependent manner, whereas distortion of integrins in the absence of receptor occupancy has no effect.

Mechanochemical switching between growth and differentiation during fibroblast growth factor-stimulated angiogenesis in vitro: role of extracellular matrix.

The angiogenic factor, basic fibroblast growth factor (FGF), either stimulates endothelial cell growth or promotes capillary differentiation depending upon the microenvironment in which it acts. Analysis of various in vitro models of spontaneous angiogenesis, in combination with time-lapse cinematography, demonstrated that capillary tube formation was greatly facilitated by promoting multicellular retraction and cell elevation above the surface of the rigid culture dish or by culturing endothelial cells on malleable extracellular matrix (ECM) substrata. These observations suggested to us that mechanical (i.e., tension-dependent) interactions between endothelial cells and ECM may serve to regulate capillary development. To test this hypothesis, FGF-stimulated endothelial cells were grown in chemically defined medium on bacteriological (nonadhesive) dishes that were precoated with different densities of fibronectin. Extensive cell spreading and growth were promoted by fibronectin coating densities that were highly adhesive (greater than 500 ng/cm2), whereas cell rounding, detachment, and loss of viability were observed on dishes coated with low fibronectin concentrations (less than 100 ng/cm2). Intermediate fibronectin coating densities (100-500 ng/cm2) promoted cell extension, but they could not completely resist cell tractional forces. Partial retraction of multicellular aggregates resulted in cell shortening, cessation of growth, and formation of branching tubular networks within 24-48 h. Multicellular retraction and subsequent tube formation also could be elicited on highly adhesive dishes by overcoming the mechanical resistance of the substratum using higher cell plating numbers. Dishes coated with varying concentrations of type IV collagen or gelatin produced similar results. These results suggest that ECM components may act locally to regulate the growth and pattern-regulating actions of soluble FGF based upon their ability to resist cell-generated mechanical loads. Thus, we propose that FGF-stimulated endothelial cells may be "switched" between growth, differentiation, and involution modes during angiogenesis by altering the adhesivity or mechanical integrity of their ECM.

Adhesion to fibronectin stimulates inositol lipid synthesis and enhances PDGF-induced inositol lipid breakdown.

The aim of these experiments was to investigate whether inositol lipids might mediate some of the effects of extracellular matrix (ECM) on cellular form and functions. The lipid phosphatidylinositol bisphosphate (PIP2) plays a role in cytoskeletal regulation while its hydrolysis products, diacylglycerol and inositol triphosphate, serve as second messengers. We therefore measured the effect of adhesion to fibronectin (FN) on PIP2 and its hydrolysis products, in the presence and absence of the soluble mitogen PDGF. PDGF induced a threefold increase in release of water-soluble inositol phosphates in C3H 10T1/2 fibroblasts when cells were attached to FN, but had little effect in suspended cells. Suppression of inositol phosphate release in unattached cells was not due to dysfunction of the PDGF receptor or failure to activate phospholipase C-gamma; PDGF induced similar tyrosine phosphorylation of PLC-gamma under both conditions. By contrast, the total mass of phosphatidylinositol bisphosphate (PIP2), the substrate for PLC-gamma, was found to decrease by approximately 80% when cells were detached from their ECM attachments and placed in suspension in the absence of PDGF. PIP2 levels were restored when suspended cells were replated on FN, demonstrating that the effect was reversible. Furthermore, a dramatic increase in synthesis of PIP2 could be measured in cells within 2 min after reattachment to FN in the absence of PDGF. These results show that FN acts directly to stimulate PIP2 synthesis, and that it also enhances PIP2 hydrolysis in response to PDGF. The increase in PIP2 induced by adhesion may mediate some of the known effects of FN on cell shape and cytoskeletal organization, while regulation of inositol lipid hydrolysis may provide a means for integrating hormone- and ECM-dependent signaling pathways.

Control of intracellular pH and growth by fibronectin in capillary endothelial cells.

The aim of this work was to analyze the mechanism by which fibronectin (FN) regulates capillary endothelial cell proliferation. Endothelial cell growth can be controlled in chemically-defined medium by varying the density of FN coated on the substratum (Ingber, D. E., and J. Folkman. J. Cell Biol. 1989. 109:317-330). In this system, DNA synthetic rates are stimulated by FN in direct proportion to its effect on cell extension (projected cell areas) both in the presence and absence of saturating amounts of basic FGF. To investigate direct growth signaling by FN, we carried out microfluorometric measurements of intracellular pH (pHi), a cytoplasmic signal that is commonly influenced by soluble mitogens. pHi increased 0.18 pH units as FN coating densities were raised and cells progressed from round to spread. Intracellular alkalinization induced by attachment to FN was rapid and followed the time course of cell spreading. When measured in the presence and absence of FGF, the effects of FN and FGF on pHi were found to be independent and additive. Furthermore, DNA synthesis correlated with pHi for all combinations of FGF and FN. Ethylisopropylamiloride, a specific inhibitor of the plasma membrane Na+/H+ antiporter, completely suppressed the effects of FN on both pHi and DNA synthesis. However, cytoplasmic pH per se did not appear to be a critical determinant of growth since DNA synthesis was not significantly inhibited when pHi was lowered over the physiological range by varying the pH of the medium. We conclude that FN and FGF exert their growth-modulating effects in part through activation of the Na+/H+ exchanger, although they appear to trigger this system via separate pathways.

Leukocyte adhesion to vascular endothelium induces E-selectin linkage to the actin cytoskeleton.

We have examined functions of the cytoplasmic domain of E-selectin, an inducible endothelial transmembrane protein, especially its ability to associate with the cytoskeleton during leukocyte adhesion. Confocal microscopy of interleukin-1 beta (IL-1 beta)-activated human umbilical vein endothelial cells (HUVEC) visualized clustering of E-selectin molecules in the vicinity of leukocyte-endothelial cell attachment sites. A detergent based extraction and Western blotting procedure demonstrated an association of E-selectin with the insoluble (cytoskeletal) fraction of endothelial monolayers that correlated with adhesion of leukocytes via an E-selectin-dependent mechanism. A mutant form of E-selectin lacking the cytoplasmic domain (tailless E-selectin) was expressed in COS-7 cell and supported leukocyte attachment (in a nonstatic adhesion assay) in a fashion similar to the native E-selectin molecule, but failed to become associated with the cytoskeletal fraction. To identify the cytoskeletal components that associate with the cytoplasmic domain of E-selectin, paramagnetic beads coated with the adhesion-blocking anti-E-selectin monoclonal antibody H18/7 were incubated with IL-1 beta-activated HUVEC, and then subjected to detergent extraction and magnetic separation. Certain actin-associated proteins, including alpha-actinin, vinculin, filamin, paxillin, as well as focal adhesion kinase (FAK), were copurified by this procedure, however talin was not. When a mechanical stress was applied to H18/7-coated ferromagnetic beads bound to the surface of IL-1 beta-activated HUVEC, using a magnetical twisting cytometer, the observed resistance to the applied stress was inhibited by cytochalasin D, thus demonstrating transmembrane cytoskeletal mechanical linkage. COS-7 cells transfected with the tailless E-selectin failed to show resistance to the twisting stress. Taken together, these data indicate that leukocyte adhesion to cytokine-activated HUVEC induces transmembrane cytoskeletal linkage of E-selectin through its cytoplasmic domain, a process which may have important implications for cell-cell signaling as well as mechanical anchoring during leukocyte-endothelial adhesive interactions.

Mechanotransduction across the cell surface and through the cytoskeleton.

Mechanical stresses were applied directly to cell surface receptors with a magnetic twisting device. The extracellular matrix receptor, integrin beta 1, induced focal adhesion formation and supported a force-dependent stiffening response, whereas nonadhesion receptors did not. The cytoskeletal stiffness (ratio of stress to strain) increased in direct proportion to the applied stress and required intact microtubules and intermediate filaments as well as microfilaments. Tensegrity models that incorporate mechanically interdependent struts and strings that reorient globally in response to a localized stress mimicked this response. These results suggest that integrins act as mechanoreceptors and transmit mechanical signals to the cytoskeleton. Mechanotransduction, in turn, may be mediated simultaneously at multiple locations inside the cell through force-induced rearrangements within a tensionally integrated cytoskeleton.

Geometric control of cell life and death.

Human and bovine capillary endothelial cells were switched from growth to apoptosis by using micropatterned substrates that contained extracellular matrix-coated adhesive islands of decreasing size to progressively restrict cell extension. Cell spreading also was varied while maintaining the total cell-matrix contact area constant by changing the spacing between multiple focal adhesion-sized islands. Cell shape was found to govern whether individual cells grow or die, regardless of the type of matrix protein or antibody to integrin used to mediate adhesion. Local geometric control of cell growth and viability may therefore represent a fundamental mechanism for developmental regulation within the tissue microenvironment.

Engineering cell shape and function.

An elastomeric stamp, containing defined features on the micrometer scale, was used to imprint gold surfaces with specific patterns of self-assembled monolayers of alkanethiols and, thereby, to create islands of defined shape and size that support extracellular matrix protein adsorption and cell attachment. Through this technique, it was possible to place cells in predetermined locations and arrays, separated by defined distances, and to dictate their shape. Limiting the degree of cell extension provided control over cell growth and protein secretion. This method is experimentally simple and highly adaptable. It should be useful for applications in biotechnology that require analysis of individual cells cultured at high density or repeated access to cells placed in specified locations.

Primary Human Lung Alveolus-on-a-chip Model of Intravascular Thrombosis for Assessment of Therapeutics.

Pulmonary thrombosis is a significant cause of patient mortality; however, there are no effective in vitro models of thrombi formation in human lung microvessels that could also assess therapeutics and toxicology of antithrombotic drugs. Here, we show that a microfluidic lung alveolus-on-a-chip lined by human primary alveolar epithelium interfaced with endothelium and cultured under flowing whole blood can be used to perform quantitative analysis of organ-level contributions to inflammation-induced thrombosis. This microfluidic chip recapitulates in vivo responses, including platelet-endothelial dynamics and revealed that lipopolysaccharide (LPS) endotoxin indirectly stimulates intravascular thrombosis by activating the alveolar epithelium, rather than acting directly on endothelium. This model is also used to analyze inhibition of endothelial activation and thrombosis due to a protease activated receptor-1 (PAR-1) antagonist, demonstrating its ability to dissect complex responses and identify antithrombotic therapeutics. Thus, this methodology offers a new approach to study human pathophysiology of pulmonary thrombosis and advance drug development.

Integrating with integrins.

Our central claim is that signaling by integrins provides a mechanism by which signals generated in response to adhesion, soluble hormones, and mechanical forces can interact. Such interactions permit cells to integrate these different classes of external stimuli and hence to orchestrate an efficient response. This integrating function of integrins is likely to be essential for much of development and physiology, as well as complex pathologies such as cancer. Understanding in detail how these signals are transduced and processed is likely to be an important area of research in the near future.

Cooperative control of Akt phosphorylation, bcl-2 expression, and apoptosis by cytoskeletal microfilaments and microtubules in capillary endothelial cells.

Capillary endothelial cells can be switched between growth and apoptosis by modulating their shape with the use of micropatterned adhesive islands. The present study was carried out to examine whether cytoskeletal filaments contribute to this response. Disruption of microfilaments or microtubules with the use of cytochalasin D or nocodazole, respectively, led to levels of apoptosis in capillary cells equivalent to that previously demonstrated by inducing cell rounding with the use of micropatterned culture surfaces containing small (<20 microm in diameter) circular adhesive islands coated with fibronectin. Simultaneous disruption of microfilaments and microtubules led to more pronounced cell rounding and to enhanced levels of apoptosis approaching that observed during anoikis in fully detached (suspended) cells, indicating that these two cytoskeletal filament systems can cooperate to promote cell survival. Western blot analysis revealed that the protein kinase Akt, which is known to be critical for control of cell survival became dephosphorylated during cell rounding induced by disruption of the cytoskeleton, and that this was accompanied by a decrease in bcl-2 expression as well as a subsequent increase in caspase activation. This ability of the cytoskeleton to control capillary endothelial cell survival may be important for understanding the relationship among extracellular matrix turnover, cell shape changes, and apoptosis during angiogenesis inhibition.

Control of cyclin D1, p27(Kip1), and cell cycle progression in human capillary endothelial cells by cell shape and cytoskeletal tension.

The extracellular matrix (ECM) plays an essential role in the regulation of cell proliferation during angiogenesis. Cell adhesion to ECM is mediated by binding of cell surface integrin receptors, which both activate intracellular signaling cascades and mediate tension-dependent changes in cell shape and cytoskeletal structure. Although the growth control field has focused on early integrin and growth factor signaling events, recent studies suggest that cell shape may play an equally critical role in control of cell cycle progression. Studies were carried out to determine when cell shape exerts its regulatory effects during the cell cycle and to analyze the molecular basis for shape-dependent growth control. The shape of human capillary endothelial cells was controlled by culturing cells on microfabricated substrates containing ECM-coated adhesive islands with defined shape and size on the micrometer scale or on plastic dishes coated with defined ECM molecular coating densities. Cells that were prevented from spreading in medium containing soluble growth factors exhibited normal activation of the mitogen-activated kinase (erk1/erk2) growth signaling pathway. However, in contrast to spread cells, these cells failed to progress through G1 and enter S phase. This shape-dependent block in cell cycle progression correlated with a failure to increase cyclin D1 protein levels, down-regulate the cell cycle inhibitor p27(Kip1), and phosphorylate the retinoblastoma protein in late G1. A similar block in cell cycle progression was induced before this same shape-sensitive restriction point by disrupting the actin network using cytochalasin or by inhibiting cytoskeletal tension generation using an inhibitor of actomyosin interactions. In contrast, neither modifications of cell shape, cytoskeletal structure, nor mechanical tension had any effect on S phase entry when added at later times. These findings demonstrate that although early growth factor and integrin signaling events are required for growth, they alone are not sufficient. Subsequent cell cycle progression and, hence, cell proliferation are controlled by tension-dependent changes in cell shape and cytoskeletal structure that act by subjugating the molecular machinery that regulates the G1/S transition.

Integrin binding and cell spreading on extracellular matrix act at different points in the cell cycle to promote hepatocyte growth.

This study was undertaken to determine the importance of integrin binding and cell shape changes in the control of cell-cycle progression by extracellular matrix (ECM). Primary rat hepatocytes were cultured on ECM-coated dishes in serum-free medium with saturating amounts of growth factors (epidermal growth factor and insulin). Integrin binding and cell spreading were promoted in parallel by plating cells on dishes coated with fibronectin (FN). Integrin binding was separated from cell shape changes by culturing cells on dishes coated with a synthetic arg-gly-asp (RGD)-peptide that acts as an integrin ligand but does not support hepatocyte extension. Expression of early (junB) and late (ras) growth response genes and DNA synthesis were measured to determine whether these substrata induce G0-synchronized hepatocytes to reenter the growth cycle. Cells plated on FN exhibited transient increases in junB and ras gene expression (within 2 and 8 h after plating, respectively) and synchronous entry into S phase. Induction of junB and ras was observed over a similar time course in cells on RGD-coated dishes, however, these round cells did not enter S phase. The possibility that round cells on RGD were blocked in mid to late G1 was confirmed by the finding that when trypsinized and replated onto FN-coated dishes after 30 h of culture, they required a similar time (12-15 h) to reenter S phase as cells that had been spread and allowed to progress through G1 on FN. We have previously shown that hepatocytes remain viable and maintain high levels of liver-specific functions when cultured on these RGD-coated dishes. Thus, these results suggest that ECM acts at two different points in the cell cycle to regulate hepatocyte growth: first, by activating the G0/G1 transition via integrin binding and second, by promoting the G1/S phase transition and switching off the default differentiation program through mechanisms related to cell spreading.

Extracellular matrix controls tubulin monomer levels in hepatocytes by regulating protein turnover.

Cells have evolved an autoregulatory mechanism to dampen variations in the concentration of tubulin monomer that is available to polymerize into microtubules (MTs), a process that is known as tubulin autoregulation. However, thermodynamic analysis of MT polymerization predicts that the concentration of free tubulin monomer must vary if MTs are to remain stable under different mechanical loads that result from changes in cell adhesion to the extracellular matrix (ECM). To determine how these seemingly contradictory regulatory mechanisms coexist in cells, we measured changes in the masses of tubulin monomer and polymer that resulted from altering cell-ECM contacts. Primary rat hepatocytes were cultured in chemically defined medium on bacteriological petri dishes that were precoated with different densities of laminin (LM). Increasing the LM density from low to high (1-1000 ng/cm2), promoted cell spreading (average projected cell area increased from 1200 to 6000 microns2) and resulted in formation of a greatly extended MT network. Nevertheless, the steady-state mass of tubulin polymer was similar at 48 h, regardless of cell shape or ECM density. In contrast, round hepatocytes on low LM contained a threefold higher mass of tubulin monomer when compared with spread cells on high LM. Furthermore, similar results were obtained whether LM, fibronectin, or type I collagen were used for cell attachment. Tubulin autoregulation appeared to function normally in these cells because tubulin mRNA levels and protein synthetic rates were greatly depressed in round cells that contained the highest level of free tubulin monomer. However, the rate of tubulin protein degradation slowed, causing the tubulin half-life to increase from approximately 24 to 55 h as the LM density was lowered from high to low and cell rounding was promoted. These results indicate that the set-point for the tubulin monomer mass in hepatocytes can be regulated by altering the density of ECM contacts and changing cell shape. This finding is consistent with a mechanism of MT regulation in which the ECM stabilizes MTs by both accepting transfer of mechanical loads and altering tubulin degradation in cells that continue to autoregulate tubulin synthesis.

Convergence of integrin and growth factor receptor signaling pathways within the focal adhesion complex.

Extracellular matrix controls capillary endothelial cell sensitivity to soluble mitogens by binding integrin receptors and thereby activating a chemical signaling response that rapidly integrates with growth factor-induced signaling mechanisms. Here we report that in addition to integrins, growth factor receptors and multiple molecules that transduce signals conveyed by both types of receptors are immobilized on the cytoskeleton (CSK) and spatially integrated within the focal adhesion complex (FAC) at the site of integrin binding. FACs were rapidly induced in round cells and physically isolated from the remainder of the CSK after detergent-extraction using magnetic microbeads coated with fibronectin or a synthetic RGD-containing peptide. Immunofluorescence microscopy revealed that multiple signaling molecules (e.g., pp60c-src, pp125FAK, phosphatidylinositol-3-kinase, phospholipase C-gamma, and Na+/H+ antiporter) involved in both integrin and growth factor receptor signaling pathways became associated with the CSK framework of the FAC within 15 min after binding to beads coated with integrin ligands. Recruitment of tyrosine kinases to the FAC was also accompanied by a local increase in tyrosine phosphorylation, as indicated by enhanced phosphotyrosine staining at the site of integrin binding. In contrast, neither recruitment of signaling molecules nor increased phosphotyrosine staining was observed when cells bound to beads coated with a control ligand (acetylated low density lipoprotein) that ligates transmembrane scavenger receptors, but does not induce FAC formation. Western blot analysis confirmed that FACs isolated using RGD-beads were enriched for pp60c-src, pp125FAK, phospholipase C-gamma, and the Na+/H+ antiporter when compared with intact CSK or basal cell surface preparations that retained lipid bilayer. Isolated FACs were also greatly enriched for the high affinity fibroblast growth factor receptor flg. Most importantly, isolated FACs continued to exhibit multiple chemical signaling activities in vitro, including protein tyrosine kinase activities (pp60c-src and pp125FAK) as well as the ability to undergo multiple sequential steps in the inositol lipid synthesis cascade. These data suggest that many of the chemical signaling events that are induced by integrins and growth factor receptors in capillary cells may effectively function in a "solid-state" on insoluble CSK scaffolds within the FAC and that the FAC may represent a major site for signal integration between these two regulatory pathways. Future investigations into the biochemical and biophysical basis of signal transduction may be facilitated by this method, which results in isolation of FACs that retain the CSK framework as well as multiple associated chemical signaling activities.

Rapid Isolation of Staphylococcus aureus Pathogens from Infected Clinical Samples Using Magnetic Beads Coated with Fc-Mannose Binding Lectin.

Here we describe how Staphylococcus aureus bacteria can be rapidly isolated from clinical samples of articular fluid and synovial tissue using magnetic beads coated with the engineered chimeric human opsonin protein, Fc-mannose-binding lectin (FcMBL). The FcMBL-beads were used to capture and magnetically remove bacteria from purified cultures of 12 S. aureus strains, and from 8 articular fluid samples and 4 synovial tissue samples collected from patients with osteoarthritis or periprosthetic infections previously documented by positive S. aureus cultures. While the capture efficiency was high (85%) with purified S. aureus strains grown in vitro, direct FcMBL-bead capture from the clinical samples was initially disappointing (< 5% efficiency). Further analysis revealed that inhibition of FcMBL binding was due to coating of the bacteria by immunoglobulins and immune cells that masked FcMBL binding sites, and to the high viscosity of these complex biological samples. Importantly, capture of pathogens using the FcMBL-beads was increased to 76% efficiency by pretreating clinical specimens with hypotonic washes, hyaluronidase and a protease cocktail. Using this approach, S. aureus bacteria could be isolated from infected osteoarthritic tissues within 2 hours after sample collection. This FcMBL-enabled magnetic method for rapid capture and concentration of pathogens from clinical samples could be integrated upstream of current processes used in clinical microbiology laboratories to identify pathogens and perform antibiotic sensitivity testing when bacterial culture is not possible or before colonies can be detected.