Endothelial cells are normally exposed constantly to mechanical forces that significantly influence their phenotype. This symposium presented recent information concerning endothelial cell responses to shear stress associated with blood flow. Endothelial cell shear stress mechanosensors that have been proposed include membrane receptor kinases, integrins, G proteins, ion channels, intercellular junction proteins, membrane lipids (e.g., those associated with caveolae), and the cytoskeleton. These sensors are linked to signaling cascades that interact with or result in generation of reactive oxygen species, nitric oxide, and various transcription factors among other responses. Endothelial cells adapt to sustained shear stress, and either an increase or decrease from normal shear leads to signaling events. In vitro models for the study of endothelial cell responses must consider the pattern of shear stress (e.g., steady vs. oscillatory flow), the scaffold for cell growth (e.g., basement membrane or other cell types such as smooth muscle cells), and the extent of flow adaptation. These cellular responses have major relevance for understanding the pathophysiological effects of increased shear stress associated with hypertension or decreased shear stress associated with thrombotic occlusion.
- cell signaling
it is widely appreciated that cells sense and respond to a broad range of chemical stimuli. Recent evidence (10, 12) has indicated that mechanical factors can also markedly influence cell structure, growth, and function. Cells in situ are subjected to varied mechanical stresses including gravitational force, mechanical stretch or strain, and shear stress. The endothelial cells lining blood vessels are subjected to each of these mechanical forces but, in particular, are exposed to a relatively elevated shear stress associated with blood flow. How do cells, and the endothelium in particular, sense a change in shear stress, and what are the signaling pathways for the cellular response? From a simplistic standpoint, changes in fluid shear stress could be sensed directly by cell membrane components such as membrane proteins, ion channels, or caveolae or by alterations of the cellular cytoskeleton; subsequent cellular signaling cascades through phosphorylation events or generation of reactive oxygen species (ROS) can lead to diverse effects such as the release of cytokines and other mediators, activation of transcription factors, altered gene and protein expression, and cell division or death (7, 8, 14, 18, 22,31). These issues that are summarized in this report were considered at a symposium held during the Experimental Biology Meeting in Orlando, FL, on April 4, 2001. The symposium was sponsored by the Respiration Section of the American Physiological Society. Although the presentations reported results with endothelial cells obtained primarily from nonpulmonary sources, specific differences in the response to shear stress between endothelial cells from pulmonary and other vascular beds have not yet been described. Because endothelial cells associated with the pulmonary vasculature may comprise 30% of the vascular endothelial cells of the body, the content of this symposium has clear relevance for physiologists interested in lung cell function as well as those with a primary interest in endothelial cell biology.
MECHANOTRANSDUCTION IN ENDOTHELIAL CELLS1
Vascular endothelial cells play significant roles in regulating vascular functions in health and disease. Shear stress can modulate endothelial cell functions by sequentially activating the mechanosensors, intracellular signaling pathways, specific transcription factors, and the expression of genes and proteins (7, 8, 15). Our laboratory found that the vascular endothelial growth factor receptor (which is a receptor tyrosine kinase) on the luminal side of endothelial cells (7) and integrins on the abluminal side (15) can serve as mechanosensors. Other potential endothelial cell membrane mechanosensors are G proteins, ion channels, intercellular junction proteins, and membrane lipids. Using 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI) and fluorescence recovery after photobleaching, Butler et al. (5) found that shear stress increases membrane fluidity, preferentially on the upstream side of the endothelial cell. This increase in fluidity may facilitate the lateral mobility of membrane proteins and enhance their interactions, e.g., oligomerization.
The activation of mechanosensors by shear stress leads to the triggering of phosphorylation cascades of signaling molecules. For example, the Ras-mitogen-activated protein kinase kinase (MEKK)-c-Jun amino-terminal kinase (JNK) signaling pathway (17) mediates the shear stress activation of the expression of monocyte chemotactic protein-1 (MCP-1) gene and protein in response to shear stress (25). A sustained application of laminar shear stress results in only transient activation of the Ras-MEKK-JNK signaling pathway and transient MCP-1 expression followed by its downregulation. Furthermore, sustained laminar shear stress activates the genes that inhibit vessel wall growth, e.g., the genes causing cell cycle arrest (19). When endothelial cells are subjected to complex flow patterns simulating those seen at the branch points, the downregulation of MCP-1 in response to sustained flow does not occur. Thus there is a sustained activation of genes such as MCP-1 in areas with complex flow patterns, especially the flow reattachment area, which has a low shear stress and a high shear stress gradient. Furthermore, endothelial cell turnover is accelerated in the areas with complex flow (9). Such accelerated cell turnover leads to an enhanced macromolecular permeability. Therefore, the temporal and spatial variations in shear stress and the flow pattern play important roles in endothelial functions such as leukocyte chemoattraction and macromolecular permeability.
Shear stress enhances endothelial cell-extracellular matrix adhesion strength, formation of stress fibers and focal adhesions, and activation of Rho family GTPases (e.g., Rac), all of which may contribute to the increase in endothelial cell migration velocity under shear. Our results indicate that endothelial cell migration is governed by a balance of forces, including the externally applied shear force, the endothelial cell-extracellular matrix adhesion force, and the intracellular locomotive forces resulting from mechanochemical transduction. The results of such interdisciplinary studies help to enhance our understanding of the fundamental process of mechanochemical transduction and the pathophysiological mechanisms underlying cardiovascular diseases.
CELL SIGNALING WITH SIMULATED ISCHEMIA IN FLOW-ADAPTED PULMONARY ENDOTHELIAL CELLS2
Most studies relating shear stress to endothelial cell function have used the experimental model of increased shear stress in cells that have been cultured under static conditions. However, endothelial cells in vivo are normally exposed and presumably adapted to a normal level of shear stress in the range of 5–20 dyn/cm2. Cells adapted to flow might be expected to respond to either an increase or decrease in shear stress from the normal level. Our studies have investigated the response of endothelial cells that are flow adapted either in situ as a normal adaptation or in vitro by a period of preconditioning to the abrupt loss of shear. This model simulates the effects of lung ischemia on the endothelium as might occur with a pulmonary thromboembolism or in association with pulmonary surgery.
Our initial studies were carried out with the isolated rat lung perfused with synthetic medium. Ventilation of the lung was continued during the ischemic period, thereby providing sufficient oxygen to maintain cellular oxygenation as confirmed by the stable tissue ATP content during 1 h of ischemia (3). We used fluorescent dyes and a microscopy system with Metamorph software (Universal Imaging, West Chester, PA) to image cell membrane potential, the generation of ROS, intracellular Ca2+ concentration, and nitric oxide (NO) generation in the vascular endothelium of subpleural capillaries during flow and subsequent ischemia. Endothelial localization of the reporter dyes was confirmed by cell association of the fluorophore DiI-acetylated low-density lipoprotein and by clear differentiation from epithelial cells labeled in lamellar bodies with Nile red (2, 3). Through the use of high-speed videomicroscopy, our laboratory (1, 26, 29) has demonstrated that membrane depolarization occurs as an immediate response to cessation of flow, followed within the first minute by ROS generation, increased intracellular Ca2+, and NO generation. Thus these events represent a rapid cellular response to the loss of shear. The use of various inhibitors suggested that membrane depolarization results from inactivation of a shear stress-sensitive K+ channel (3, 26), that generation of ROS occurs through activation of endothelial membrane NADPH oxidase (lung ROS generation with ischemia was lost in mice with “knockout” of gp91phox, the flavoprotein component of the oxidase) (2), that increased intracellular Ca2+ is due to intracellular release followed by influx from extracellular sources (1, 26, 29), and that the generation of NO is due to activation of endothelial NO synthase (1).
To extend the results obtained in the intact lung to an in vitro system, Manevich et al. (20) developed a cell culture model that permitted the direct spectroscopic evaluation of cells during the initial phase of simulated ischemia. Bovine pulmonary artery endothelial cells were grown in a parallel plate chamber constructed to allow its insertion into a standard spectrophotometer or fluorometer chamber. Cells that were adapted to laminar flow (5 dyn/cm2 for 24 h) exhibited the characteristic structural reorganization, with their long axis in the direction of flow. We confirmed that oxygenation of the medium remained adequate during the initial 3 min of simulated ischemia because Po 2 was >40 mmHg and the observed initial effects of ischemia were not altered by preequilibration of the medium with 100% oxygen. Flow-adapted (but not control) cells subjected to simulated ischemia demonstrated cell membrane depolarization, ROS generation, increased intracellular Ca2+, and NO generation as observed for the endothelium in the intact lung. As expected for the membrane-bound NADPH oxidase, generation of O · was demonstrated to be extracellular by reduction of exogenous cytochrome c and its inhibition with superoxide dismutase; H2O2formed from O · would diffuse freely into cells during continued ischemia, serving as a mechanism for signal transduction. An additional in vitro study (31) utilized bovine pulmonary artery endothelial cells grown in an artificial capillary system where the oxygenation of cells during simulated ischemia could be maintained for prolonged periods. These studies demonstrated that ROS production led to activation of the transcription factors nuclear factor-κB and c-Jun after 1 h of ischemia and increased cell division at 24 h of ischemia.
Our interpretation of these results is that the loss of shear stress associated with flow cessation in flow-adapted endothelial cells leads to a complex signaling response, possibly initiated by flow-sensitive K+ channels, leading to membrane depolarization and propagated by ROS generation and Ca2+ release. The physiological significance of these signaling events may be the net generation of vasoactive mediators and signals for capillary angiogenesis in an attempt to restore the compromised circulation.
ROLE OF ION CHANNELS IN SHEAR STRESS SENSING IN VASCULAR ENDOTHELIUM3
The ability of arterial endothelial cells to sense and respond to changes in fluid mechanical shear stress is essential for vasoregulation and for vascular wall remodeling and may play a role in the development and localization of early atherosclerotic lesions. Exposure to shear stress elicits humoral, metabolic, and structural responses in endothelial cells. A virtually immediate endothelial response to shear stress is the activation of flow-sensitive ion channels. Because of their very rapid response to flow, these ion channels have been hypothesized to play a role in shear stress sensing and transduction.
The activation of flow-sensitive ion currents in endothelial cells was first reported by Olesen et al. (23), who used whole cell patch-clamp recordings to demonstrate that steady shear stress stimulates inward rectifying K+ channels, the activation of which leads to cell membrane hyperpolarization. More recently, Barakat et al. (4) used whole cell patch-clamp recordings and measurements from membrane potential-sensitive fluorescent dyes to demonstrate that steady shear stress also induces an outward rectifying Cl− current in endothelial cells and that activation of this current reverses the K+ channel-mediated hyperpolarization to depolarization within ∼100 s of the onset of flow. On cessation of flow, the membrane potential returns to preflow baseline levels, but this process is relatively slow. The K+ and Cl− currents are independently activated; pharmacologically blocking either current does not interfere with the activation of the other by flow. Although both of these currents are activated very rapidly on the initiation of flow, the fact that the cell membrane initially hyperpolarizes and then depolarizes suggests that flow-sensitive K+ channels attain full activation and/or desensitize more rapidly than the Cl−channels. Differences in the dynamics of activation and desensitization between the two channel types may have important implications for the vascular endothelial responsiveness to a sustained flow stimulus.
Endothelial cells in arterial regions prone to the development of early atherosclerosis are cuboidal (or round), whereas in athero-resistant regions, they are elongated. Therefore, there is a need for understanding how the shape of endothelial cells may regulate their function. Toward this goal, endothelial cells were exposed for 24 h to a steady shear stress of 19 dyn/cm2 that led to extensive cytoskeletal remodeling and cellular elongation and alignment in the direction of flow. Subsequent whole cell patch-clamp recordings on flow-elongated endothelial cells revealed that, as in previously unsheared cells, flow results in cell membrane hyperpolarization followed by depolarization and hence activates both the flow-sensitive K+ and Cl− channels. These results may be interpreted in one of two ways: either cell shape does not play a role in regulating ion channel responsiveness to flow or the extensive signaling cascades activated during the 24-h flow-conditioning period serve an adaptive function that acts to restore ion channel sensitivity to flow. These two interpretations can be separated with methods that allow more noninvasive control of cell shape. Indeed, a recent study (13) has demonstrated that the extent of endothelial cell elongation can be controlled by culturing the cells in microchannels of different width. Studies of ion channel responses to flow in endothelial cells cultured in these microchannels should shed light on the possible relationship between flow-sensitive ion channels and endothelial cell shape.
The precise role of flow-induced K+ and Cl−currents in overall shear stress sensing and transduction in endothelial cells remains to be elucidated. Pharmacological blockers of these channels considerably attenuate, or in some cases entirely abolish, several shear stress-induced endothelial responses including increased transforming growth factor-β1 mRNA and Na-K-Cl cotransport protein levels (22, 27). This suggests that flow-activated ion channels play an essential role in regulating certain endothelial cell responses to flow. However, recent data (18) demonstrating that flow-sensitive ion channels are regulated by pertussis toxin-sensitive GTP-binding proteins (or G proteins) suggest that stimulation of these ion channels by shear stress occurs downstream of G protein activation. An understanding of the complex interactions among flow-sensitive ion channels, G proteins, and other cellular structures such as the cell membrane and cytoskeleton is essential for elucidating the mechanisms by which endothelial cells sense and respond to fluid mechanical stimulation.
ENGINEERING THE IN VIVO ENVIRONMENT FOR THE STUDY OF THE ENDOTHELIAL RESPONSE TO SHEAR STRESS4
Even though in vitro studies over the past two decades on the influence of shear stress on vascular endothelial cell function have provided considerable insights (reviewed in Ref. 21), we must continually strive to engineer the cell culture environment so as to make it more physiological. In vivo, the vascular endothelial cell provides the inner lining of the blood vessel wall, which itself has a three-dimensional architecture and where the vascular smooth muscle cell is the natural neighbor of the vascular endothelial cell. Although much of the focus of vascular tissue engineering is on creating small-diameter blood-vessel substitutes, the same technologies can be used to engineer a model of the blood vessel wall that is more representative of native tissue.
One such approach is to incorporate vascular smooth muscle cells within a collagen scaffold and use this as the underlying substrate onto which one establishes an endothelial monolayer (32). This cell-seeded collagen substrate can be altered by mechanical conditioning (24). Initial experiments using this model system have focused on studying the changes in endothelial cell morphology and proliferation due to the influence of a laminar steady-state flow and the associated shear stress. Without flow, it is the collagen organization that determines the orientation of the endothelial cells and localization of the actin microfilaments; however, when flow is imposed, the endothelial cells do align themselves with the direction of flow. Also observed is a reduction in cell proliferation due to the presence of flow. Although this finding is not new, what is important to note is that the observed reduction in cell proliferation associated with flow is dependent on the organization of the collagen.
In engineering the in vitro environment so as to make it more physiological, the exact characteristics of the flow environment are critical (6, 11). Not only is blood flow in large arteries pulsatile as opposed to steady but exposing endothelial cells to a sudden onset of flow (as is normally done) must be considered somewhat artifactual. A better model might be to first precondition the endothelial monolayer with one level of shear stress and then change the flow and thus the associated shear stress. In recent experiments in which the influence of shear stress on MCP-1 was investigated, endothelial cells were first exposed to a shear stress of 2.5 dyn/cm2 for 44 h, and then the flow was changed to a shear stress of 15 dyn/cm2. In contrast to the response of endothelial cells to a sudden onset of flow at a shear stress of 15 dyn/cm2, the effect of preconditioning is to dramatically change the initial MCP-1 transient. If responses are different, then we must wonder if the pathways involved in the transduction of this mechanical signal may also be different.
Clearly, there is much more that can be done to better engineer the in vitro environment so as to make it more physiological. Only if we do this will we be able to achieve in vitro results that truly predict in vivo behavior.
As the above presentations have shown, endothelial cells clearly exist in a dynamic state and respond to alterations in the level of shear stress and presumably also to forces associated with mechanical stretch and strain by signaling events that can markedly influence their behavior. Many of these responses and the stimuli that elicit them are just beginning to be understood. As we seek to better understand these endothelial responses, we will need to have in vitro model systems that are more representative of a blood vessel and the environment in which an endothelial cell resides. Furthermore, recent work (16) indicates that vascular endothelial cells represent a phenotypically heterogeneous population. Thus it is possible that the responses of endothelial cells to mechanical stresses will differ depending on their vessel of origin. Much future effort will be required to elucidate these differences to gain a full understanding of the role of the cellular milieu in determining endothelial function.
We thank the American Physiological Society for support of this symposium and Dr. Abu Al-Mehdi for reviewing and Elaine Primerano for typing the manuscript.
The research reported in the symposium was supported by National Heart, Lung, and Blood Institute Grants HL-60290 (to A. B. Fisher), HL-62747, and HL-64382 (both to S. Chien); a Biomedical Engineering Research Grant from the Whitaker Foundation (to A. I. Barakat), and National Science Foundation Grant EEC-9731643 (to R. M. Nerem).
Address for reprint requests and other correspondence: A. B. Fisher, Institute for Environmental Medicine, Univ. of Pennsylvania School of Medicine, 1 John Morgan Bldg., 3620 Hamilton Walk, Philadelphia, PA 19104-6068 (E-mail:).
↵1 Presented by Shu Chien.
↵2 Presented by Aron B. Fisher. Drs. Abu Al-Mehdi and Yefim Manevich were major collaborators.
↵3 Presented by Abdul I. Barakat.
↵4 Presented by Robert M. Nerem.
- Copyright © 2001 the American Physiological Society