Fluid Flow Rapidly Activates G Proteins in Human Endothelial Cells
Involvement of G Proteins in Mechanochemical Signal Transduction
Fluid shear stimulates endothelial cells, with the external hemodynamic forces transduced across the plasma membrane to modulate intracellular events. We report the first direct evidence that identifies specific GTP binding proteins (G proteins) activated within 1 second of flow onset, representing one of the earliest mechanochemical signal transduction events reported to date in shear-stimulated endothelium. A nonhydrolyzable GTP photoreactive analogue, azidoanilido [α-32P]GTP (AAGTP), allowed irreversible labeling of flow-stimulated G proteins, with two protein bands (42 kD and 31 kD) identified in human umbilical vein endothelial cells (HUVECs) subjected to laminar flow (10 dyne/cm2) in a parallel-plate flow chamber. Immunoprecipitation of labeled whole-cell lysates identified the specific G-protein subunits Gαq/α11 and Gαi3/αo as being activated by flow. Endothelial cell membrane vesicles were sheared in a cone-and-plate viscometer, with the 42-kD protein band labeled by AAGTP, but the 31-kD protein absent, indicating that the 42-kD G protein is membrane associated and activated independently of intact cytoskeletal or cytosolic components. Our results describe one of the earliest flow-induced signaling events reported in HUVECs, providing insight into the primary mechanosensing and signal transduction mechanisms.
Endothelial cells lining the vasculature are continuously exposed to fluid flow and to the ensuing mechanical forces.1 They are able to sense fluid flow and effect changes that serve to maintain a specific level of flow or associated shear stress in blood vessels,2 although the molecular mechanisms by which flow is sensed and the signal transduced across the endothelial plasma membrane are not well understood. Transfer of fluid-flow forces to the cell occurs at the luminal cell surface, with the plasma membrane and associated transmembrane proteins likely candidates for mechanoreceptors that transduce the extracellular stimulus into intracellular biochemical signals.
Fluid flow stimulates numerous responses in endothelial cells. These include elevated production of the second messengers IP33 4 5 and cGMP6 7 ; increased release of the vasoactive compounds PGI2,8 9 10 NO,6 7 and endothelin-111 ; increased mitogen-activated protein kinase activity12 ; and elevated levels of platelet-derived growth factor13 and c-fos gene expression.14 Many of these flow-induced responses are mediated by G-protein activation, as demonstrated by their inhibition by GDPβS,6 9 12 13 a nonhydrolyzable analogue of GDP. PTX inhibited PGI2,9 while having no effect on NO or cGMP production,6 indicating that both PTX-sensitive and -insensitive G proteins are activated by flow.
Heterotrimeric G proteins, composed of α, β, and γ subunits, transduce signals from activated transmembrane receptors to intracellular effectors. When associated with an activated receptor, the α subunit exchanges bound GDP for GTP and separates from the βγ dimer. Both the GTP α and βγ subunits mediate downstream signaling events.15 The α subunit is inactivated by GTPase activity, restoring the GDP-bound state. Subclasses of Gα subunits involved in signal transduction2 include Gs, which stimulates adenylyl cyclase (and hence cAMP generation) and activates Ca2+ channels; Gi, which inhibits adenylyl cyclase and activation of K+ channels; and Gq, which activates phospholipase C. The activity of G proteins can be discriminated by the use of antagonists such as GDPβS, a general inhibitor of G proteins, and PTX, which preferentially inhibits the Gi class, while having no effect on Gq.
Whereas intracellular events triggered by fluid shear have been elucidated, the primary mechanosensor that transduces this mechanical stimulus into a biochemical signal remains unknown. Many mechanisms have been proposed (see Davies2 for review), including the direct stimulation of transmembrane proteins exposed on the luminal surface or distorted within the strained membrane (including G-protein receptors), activation of ion channels, which alter membrane polarization or intracellular Ca2+, transduction of stress along cytoskeletal elements to other regions of the cell (focal adhesions, nuclear membrane), and changes in the physical properties of the plasma membrane itself under flow. Although indirect evidence or secondary events have been used to infer their involvement, no direct evidence of these mechanisms directly activating signaling pathways due to fluid shear has been demonstrated.
In the present study, we show that membrane-associated G proteins are activated during the first second of flow, representing one of the earliest responses to fluid shear reported to date in endothelial cells. Photoreactive radiolabeling and immunoprecipitation allowed the identification of Gαq/α11 and Gαi3/αo as two of the G proteins activated, indicating that both PTX-insensitive (Gq) and PTX-sensitive (Gi) subunits are involved in this rapid response.
Materials and Methods
AAGTP (specific activity, 6.1 Ci/mmol) was obtained from ICN Radiochemicals. Antibodies to Gα were purchased from Du Pont–New England Nuclear. PTX was from List Biological Laboratories, ATP-free medium 199 and FBS from Hy-Clone, and Protein A–Sepharose CL-4B from Schleicher & Schuell. All other reagents were obtained from Sigma Chemical Co.
Cell Culture and Fluid Flow
Primary HUVECs were harvested from fresh umbilical cords.8 The growth medium (ATP-free medium 199 with 20% FBS, l-glutamine [2 mmol/L], penicillin [100 U/mL], and streptomycin sulfate [100 μg/mL]) was replaced with incubation medium (ATP-free medium 199 supplemented with 2% BSA) and incubated for at least 3 hours prior to flow. Exposure of intact and attached cells to laminar fluid flow was accomplished in a parallel-plate flow chamber16 through which medium was driven by a syringe pump (Harvard Apparatus). Stationary control cells were kept in Petri dishes filled with incubation medium. In other experiments, suspensions of membrane vesicles were sheared in a cone-and-plate viscometer. The cone angle of this device was 0.5 degree.
Endothelial cell membranes were prepared as described,17 with modifications. Membranes were prepared from confluent slides of HUVECs after washing the adherent cells with PBS. Cells were harvested by incubation in PBS (20 mmol/L sodium phosphate, pH 7.4, 150 mmol/L NaCl) and 5 mmol/L EDTA for 15 to 30 minutes at 37°C. Cells were pelleted by low-speed centrifugation and resuspended in HME buffer (20 mmol/L sodium HEPES, pH 7.4, 1 mmol/L EDTA, 2 mmol/L MgCl2). Cells were lysed in HME buffer (containing protease inhibitors PMSF [10 μg/mL], leupeptin [2 μg/mL], pepstatin A [2 μg/mL], and aprotinin [2 μg/mL]) by two freeze-thaw cycles in dry ice/ethanol followed by 30 strokes of a dounce homogenizer on ice. Nuclei were removed from the homogenate by microcentrifugation at 2500 rpm for 5 minutes at 4°C. The membranes in the low-speed supernatant were then pelleted by centrifugation at 100 000g (40 000 rpm, Beckman Ti-70.1) at 4°C in a Beckman L8-M preparative ultracentrifuge. The membranes prepared by this method are composed mostly of vesicles. The membrane pellet was then resuspended in HME buffer to a protein concentration of 5 to 20 mg/mL. Protein content of the membranes was determined using a Bradford assay (Bio-Rad Laboratories), and fresh membrane preparations were used in the experiments.
Photoaffinity Labeling of G Proteins
Photoaffinity labeling of G proteins in membranes with AAGTP18 was carried out as previously reported,19 with modifications. Membranes were diluted with 20 mmol/L HEPES, pH 7.4, to 2.5 mg/mL concentration. The assay mixture containing 20 μL of 3× assay buffer (30 mmol/L HEPES, pH 7.4, 0.1 mmol/L EDTA, 5 mmol/L MgCl2, 100 mmol/L NaCl, and 3 μmol/L GDP), 20 μL of diluted membranes, 10 μL of water, and 10 μL of AAGTP (10 μCi [1.64 nmol]) was added to a cone-and-plate viscometer. The membrane suspensions were sheared at 75 dyne/cm2 for 1 minute at 37°C and immediately irradiated with UV light (Mineralight; 254 nm, 90 W) for 1 minute. UV light serves to covalently incorporate the bound radiolabeled guanine nucleotide. The controls were incubated at 37°C for a similar duration. Samples were placed on ice, and unbound AAGTP was removed by centrifugation. Samples were resuspended in 1× assay buffer containing 2 mmol/L DTT without GDP. Membrane suspensions were then irradiated on ice for 3 minutes at a distance of 4 cm with UV light. Pelleted membranes were dissolved in SDS-PAGE sample buffer.20
Intact confluent HUVEC monolayers on glass slides were incubated in medium containing 20 μmol/L digitonin and AAGTP (10 μCi [1.64 nmol]/106 cells) for 3 minutes at 37°C. Cells were checked for viability after incubation with digitonin, using the trypan blue (0.4%) exclusion method, with 80% of cells viable for up to 15 minutes' incubation. As a detergent, digitonin may act to not only permeabilize the membrane but activate membrane-bound proteins. Data presented in this study are normalized relative to identically treated static controls, which serves to account for these effects. Cells were then sheared in a parallel-plate flow chamber using incubation medium. Cell monolayers were immediately irradiated for 1 minute with UV light, then placed on ice for additional UV exposure (3 minutes). Control stationary cells were placed on a parallel-plate flow chamber and treated identically but were not exposed to flow. Following UV irradiation, cells were rinsed with ice-cold 10 mmol/L HEPES, pH 7.4, 5 mmol/L MgCl2, 100 mmol/L NaCl, and 4 mmol/L DTT and scraped from the slides. The suspensions were centrifuged at 2000g for 10 minutes and the cell pellets dissolved in SDS-PAGE sample buffer20 with 50 mmol/L DTT.
SDS-PAGE and Autoradiography
SDS-PAGE was performed on a discontinuous slab gel system with a 4% acrylamide stacking and 10% acrylamide separating gel.20 After electrophoresis, the gels were dried and exposed to Kodak XR-OMAT film with an intensifying screen for 2 to 5 days at −80°C. Autoradiographs were quantified using an image analyzer (Alpha Innotech, Model IS-1000).
Immunoprecipitation of Labeled G Proteins
To identify the AAGTP-labeled G-protein subunits following exposure to shear, cells were lysed for 30 minutes in buffer (50 mmol/L Tris HCl, pH 7.4, 100 mmol/L NaCl, 1 mmol/L EDTA, 1% Triton X-100, 0.5% nonidet P-40, 0.2 mmol/L sodium vanadate, PMSF [10 μg/mL], leupeptin [2 μg/mL], pepstatin A [2 μg/mL], and aprotinin [2 μg/mL]) on ice. Lysates were ultracentrifuged at 100 000g for 30 minutes. The clear supernatants were incubated at 4°C for 3 hours with polyclonal antibodies specific for Gα subunits: Gαq/Gα11 (QL), GαT/Gαi1/Gαi2 (AS/7), and Gαi3/Gαo (EC/2). Mixtures were incubated with Protein A–Sepharose CL-4B for 4 to 5 hours at 4°C and washed four times with NET buffer (150 mmol/L NaCl, 0.5 mmol/L EDTA, 50 mmol/L Tris HCl, pH 8.0). Immunoprecipitates were solubilized in electrophoresis sample buffer20 and analyzed by SDS-PAGE autoradiography.
Inhibition of GTP Binding
Endothelial cells were treated with PTX (1 μg/mL) in the incubation medium at 37°C for 6 hours. Under similar experimental conditions, ribosylation of G proteins with PTX has been demonstrated earlier in our laboratory.6 It was shown that ribosylation by PTX was time dependent and was not complete until 4 hours of incubation with the toxin. After incubation with PTX, cell monolayers were washed twice with phosphate buffer, incubated with AAGTP (10 μCi [1.64 nmol]/106 cells) for 3 minutes, and subjected to fluid flow (10 dyne/cm2) for 10 seconds followed by UV irradiation.
Fluid Flow–Induced GTP Binding in HUVECs
In experiments with intact endothelial cells, confluent HUVEC monolayers were loaded with AAGTP and then exposed to fluid flow in a parallel-plate flow chamber for 10 seconds at 10 dyne/cm2. In adherent cells, autoradiography revealed a 42-kD protein band as well as a prominent 31-kD protein band (Fig 1⇓). Time and flow rate (dose) dependency for flow-induced GTP binding to the 42-kD protein is shown in Fig 2⇓. Cells exposed to flow for 0, 1, 3, or 10 seconds at a shear stress of 10 dyne/cm2 displayed a significant increase in bound GTP within the first second of flow and reached a plateau by 3 seconds. Cells exposed to 0, 3, 10, or 20 dyne/cm2 for 3 seconds displayed no significant dose dependence of G-protein activation with increasing levels of shear.
Activation of Membrane-Bound G Proteins in Vesicles
To determine whether fluid flow–activated G proteins in endothelial cells are membrane bound or cytosolic, membrane vesicles prepared from endothelial cells were incubated with AAGTP and subjected to fluid shear at 75 dyne/cm2 for 1 minute in a cone-and-plate viscometer. A 42-kD protein band was observed (Fig 3⇓), and there was complete absence of the 31-kD band for the intact cells. There was no significant binding in stationary controls. Under the stationary conditions, bradykinin (10 nmol/L) stimulated G proteins, as indicated by AAGTP binding in endothelial cell membranes, providing the positive control.
Immunoprecipitation of G Proteins
To identify specific heterotrimeric G proteins involved, polyclonal antibodies to Gα subunits Gαq/Gα11 (QL), GαT/Gαi1/Gαi2 (AS/7), and Gαi3/Gαo (EC/2) were used for immunoprecipitation. In sheared endothelial cell lysates, antibodies specific to Gαq/Gα11 and Gαi3/Gαo were bound to corresponding AAGTP-labeled proteins (Fig 4⇓), indicating these two classes of Gα subunits were activated by fluid shear. In the case of Gαq/α11, basal activation can be seen under stationary conditions. This may be attributed to the sensitivity of G proteins belonging to the Gαq subclass, which have a high affinity for GTP binding. Antibodies specific to Gαi2/Gαi1/GαT did not bind labeled protein (data not shown).
PTX Inhibition of GTP Binding in HUVECs During Flow
The flow-activated Gα subunit Gαi3/Gαo identified by immunoprecipitation is known to be PTX sensitive, whereas Gαq/Gα11 is PTX insensitive.15 To further characterize this flow-induced G-protein activation, cells were pretreated with PTX (1 μg/mL) for 6 hours and subjected to flow at 10 dyne/cm2 for 10 seconds (Fig 5⇓). There was an 85% inhibition of flow-stimulated GTP binding to the 42-kD protein.
This study has demonstrated that G proteins are activated within the first second of exposure to fluid flow, which represents one of the earliest responses reported to date in shear-stimulated endothelium. Immunoprecipitation of AAGTP-labeled proteins with polyclonal anti-Gα antibodies identified the heterotrimeric G-protein subunits Gαq/α11 and Gαi3/αo as being activated in this rapid response. Exposure of membrane vesicles to shear resulted in the activation of heterotrimeric G proteins, indicating that the intact cytoskeleton was not required in this flow-stimulated mechanotransduction pathway.
Activation of PTX-sensitive and -insensitive G proteins by fluid flow in the present study is consistent with known flow-induced responses in endothelial cells. Fluid shear was shown to activate phospholipase C,3 10 which rapidly stimulates the hydrolysis of phosphatidylinositol 4,5-diphosphate into the second messengers IP33 4 5 and diacylglycerol.5 In the ensuing second-messenger cascade, IP3 initiates Ca2+ release from internal stores,21 22 23 while diacylglycerol activates protein kinase C, which in turn mediates platelet-derived growth factor gene expression13 and endothelin-1 release.11 The phospholipase C activation and subsequent phosphoinositide turnover is linked to heterotrimeric G-protein activation,13 specifically members of the PTX-insensitive Gq class.15 Together, these studies demonstrate that the PTX-insensitive Gq subunits are activated by flow, resulting in a cascade of intracellular events, with the present study identifying the involvement of the specific subunit Gαq/α11.
A second fluid shear–stimulated pathway involves the activation of phospholipase A2 to liberate arachidonic acid from membrane phospholipids, which is converted intracellularly to produce PGI2.8 9 10 This flow-induced PGI2 production was inhibited by PTX,9 indicating that a separate class of PTX-sensitive G proteins mediates this pathway. In the class of G proteins inactivated by PTX are the Gα subunits Gi1, Gi2, Gi3, and G0.15 Immunoprecipitation with antibodies in the present study showed that this Gi activation step occurs within the first second of flow, identifying the Gαi3/αo subunit. Gαi1/αi2 antibody was negative for immunoprecipitation.
We have shown that G proteins are rapidly activated in endothelium exposed to fluid shear, with labeled GTP binding to the α subunit within the first second of flow. Ion channel activation and alterations in cytosolic free calcium ([Ca2+]i) are two flow-induced responses that occur within seconds (see Davies2 for review) and that have been proposed as primary steps in flow transduction. Calcium is an important signaling molecule that mediates critical intracellular pathways. Elevation of [Ca2+]i has been shown to be due to release from internal Ca2+ stores21 22 by an IP3-dependent mechanism.23 Peak [Ca2+]i levels were seen 15 to 40 seconds after the onset of flow and were independent of extracellular Ca2+, cell preconditioning by flow, and membrane depolarization with high K+.21 The time course was further refined recently for endothelial cells exposed to fluid shear when [Ca2+]i and NO release were measured by using fura 2 fluorescence and a porphyrinic microsensor.24 Shear stress (0.2 to 10 dyne/cm2) caused transient NO release 3 to 5 seconds after the initiation of flow and 1 to 3 seconds after the initial rise in [Ca2+]i. The IP3 dependence, combined with the temporal studies, demonstrates that changes in [Ca2+]i occur downstream of primary mechanosensing events at times later than G-protein activation.
Activation of the inward rectifying K+ channel by fluid shear is also rapid, occurring within seconds of the initiation of flow. The influence of the K+ channel in the transduction of flow, however, is not clear. We recently demonstrated25 that membrane depolarization in endothelial cells does not influence the cGMP levels, with cGMP elevation a known flow-stimulated response.6 7 Levels of cGMP in flow-treated and stationary cultures were unaffected when cells were depolarized with KCl (90 mmol/L) or treated with tetraethylammonium chloride, a Ca2+-dependent K+ channel blocker. Flow-induced production of NO levels was also not affected by membrane depolarization. These results are consistent with the findings discussed earlier21 that depolarization had no effect on flow-induced elevation of [Ca2+]i. The K+ channel has also demonstrated a delayed response to shear,26 indicating activation may occur downstream of primary signal transduction events. These results indicate that although activated by fluid flow on the same time scale (seconds) as G proteins, the inward rectifying K+ channel does not appear to play a significant role in the mechanochemical transduction of fluid shear.
One question that remains unanswered is how the extracellular flow stimulus is sensed by the cell and transduced across the lipid bilayer to trigger these intracellular events. While transmembrane protein complexes have been identified in ligand binding–induced G-protein activation, no such receptor has been isolated for the mechanotransduction of flow. To investigate the role of the membrane in mechanotransduction, vesicles were subjected to a bulk fluid shear of 75 dyne/cm2 in a cone-and-plate viscometer, as similar shear stresses were shown to activate larger subcellular particles such as platelets.27 The 42-kD G protein(s) stimulated in the suspended membrane vesicles in our study demonstrated that an intact cytoskeleton is not required in this signal pathway, which suggests a primary receptor associated with the lipid bilayer. This contrasts with other proposed mechanisms, which require the primary mechanoreceptor to connect with the cytoskeleton to transduce external forces into intracellular biochemical signals (see Schwartz and Ingber28 for review).
Increasing evidence points to the plasma membrane itself as a primary mechanotransducer. Fluid shear has been shown to alter physical properties of the membrane, increasing permeability to merocyanine 540,29 an amphipathic fluorescent dye that is incorporated into the phospholipid bilayer in a dose-dependent manner with increasing levels of shear. The increased dye incorporation may be due to changes in the packing of the bilayer, with the reduced phospholipid packing resulting in increased membrane fluidity. We hypothesized that perturbations in the membrane under fluid shear due to this increased fluidity may be transduced directly to the heterotrimeric G proteins on the cytosolic face of the plasmalemma. Recent work from our lab also supports the role of membrane fluidity in mechanotransduction, since stiffening the membranes of HUVEC monolayers after incubation with cholesterol hemisuccinate reduced the response to fluid shear, as indicated by reduced NO production. Focusing on the role of the cytoskeleton, we found that disruption of actin filaments with cytochalasin D had little effect on NO production, while disruption of microtubules by colchicine increased production relative to controls (H.K. Knudsen and J.A. Frangos, unpublished observation, 1996). These observations suggest that the cytoskeleton may have a limited or possible inhibitory role in the mechanotransduction of fluid shear, serving in some cases to restrain the plasma membrane, reducing local membrane deformation and hence the flow signal to the cell. Collectively, these observations indicate that the activation of mechanochemical transduction pathways in HUVECs by fluid shear does not require an intact cytoskeleton and that the plasma membrane itself may serve as the primary receptor in activating the intracellular signal cascade, with alterations in membrane fluidity modulating the receptor sensitivity.
At least one low-molecular-weight (≈31 kD) G protein, however, was labeled in adherent endothelial cells by fluid flow but absent in the sheared vesicles (Fig 3⇑). These unidentified proteins may be cytosolic in origin or complexed with a cytoskeletal protein; their signaling pathway requires further investigation.
The results of this study indicate G proteins are closely associated with the flow-stimulated mechanochemical transducer within endothelial cells. Within the circulation, endothelial alignment, atherosclerotic lesion formation, vascular remodeling, and vasoactive molecule release are all mediated in part by hemodynamic forces, with G-protein activation a key transduction step in the modulation of many of these responses.2 Detection of rapid G-protein activation, along with identification of the specific subunits involved, provides a powerful tool with which to view the initial events that transduce the mechanical stimulus into an intracellular biochemical signal. The ability to identify and manipulate these pathways at their origin would benefit the understanding of the vasculature as well as the diagnosis and treatment of cardiovascular disorders. Although identifying the downstream biochemical events following the application of fluid shear has met with great success, the receptor(s) responsible for the primary transduction event remains elusive. The lipid bilayer itself has been proposed as a primary mechanosensor/transducer of the flow stimulus, with perturbations due to increased membrane fluidity directly activating G proteins on the cytosolic face.9 29
In summary, we have demonstrated that fluid flow activates heterotrimeric G proteins either directly or indirectly through a G protein–coupled mechanoreceptor in the cell membrane. This G-protein activation within the first second of flow represents one of the earliest signal transduction events reported in flow-stimulated HUVECs and provides valuable insight into how hemodynamic forces are transduced across the plasma membrane.
Selected Abbreviations and Acronyms
|HUVEC||=||human umbilical vein endothelial cell|
This work was supported by National Heart, Lung, and Blood Institute grant HL-40696. J.A. Frangos is a recipient of a National Science Foundation Presidential Young Investigator Award. The authors express their sincere appreciation to Polyclinic of Harrisburg and Harrisburg Hospital of Pennsylvania and to Alvarado Hospital and the UCSD Medical Center of San Diego for supplying umbilical cords.
- Received March 21, 1996.
- Accepted June 17, 1996.
Davies PF. Flow-mediated endothelial mechanotransduction. Physiol Rev. 1995;75:519-560.
Prasad ARS, Logan SA, Nerem RM, Schwartz CJ, Sprague EA. Flow-related responses of intracellular inositol phosphate levels in cultured aortic endothelial cells. Circ Res. 1993;72:827-836.
Kuchan MJ, Jo H, Frangos JA. Role of G proteins in shear stress–mediated nitric oxide production by endothelial cells. Am J Physiol. 1994;267:C753-C758.
Kuchan MJ, Frangos JA. Role of calcium and calmodulin in flow-induced nitric oxide production in endothelial cells. Am J Physiol. 1994;266:C628-C636.
Frangos JA, Eskin SG, McIntire LV, Ives CL. Flow effects on prostacyclin production by cultured human endothelial cells. Science. 1985;27:1477-1479.
Kuchan MJ, Frangos JA. Shear stress regulates endothelin-1 release via protein kinase C and cGMP in cultured endothelial cells. Am J Physiol. 1993;264:H150-H156.
Tseng H, Peterson TE, Berk BC. Fluid shear stress stimulates mitogen-activated protein kinase in endothelial cells. Circ Res. 1995;77:869-878.
Simon MI, Strathmann MP, Guatam N. Diversity of G proteins in signal transduction. Science. 1991;252:802-808.
Dell'Acqua ML, Carroll RC, Peralta EG. Transfected m2 muscarinic acetylcholine receptors couple to Gαi2 and Gαi3 in Chinese hamster ovary cells. J Biol Chem. 1993;268:5676-5685.
Pfeuffer T. GTP-binding proteins in membranes and the control of adenylate cyclase activity. J Biol Chem. 1977;252:7224-7234.
Im MJ, Graham RM. A novel guanine nucleotide-binding protein coupled to the α1-adrenergic receptor. J Biol Chem. 1990;265:18944-18951.
Shen J, Luscinskas FW, Connolly A, Dewey CF Jr, Gimbrone MA Jr. Fluid shear stress modulates cytosolic free calcium in vascular endothelial cells. Am J Physiol. 1992;262:C384-C390.
Geiger RV, Berk BC, Alexer RW, Nerem RM. Flow-induced calcium transients in single endothelial cells: spatial and temporal analysis. Am J Physiol. 1992;262:C1411-C1417.
Kanai AJ, Strauss HC, Truskey GA, Crews AL, Grunfeld S, Malinski T. Shear stress induces ATP-independent transient nitric oxide release from vascular endothelial cells, measured directly with a porphyrinic microsensor. Circ Res. 1995;77:284-293.
Gooch KJ, Frangos JA. Flow- and bradykinin-induced nitric oxide production by endothelial cells is independent of membrane potential. Am J Physiol. 1996;262:C546-C551.
Jacobs ER, Cheliakine C, Gebremedhin D, Davies PF, Harder DR. Shear activated channels in cell-attached patches of aortic endothelial cells. FASEB J. 1993;7:71. Abstract.
Schwartz MA, Ingber DE. Integrating with integrins. Mol Biol Cell. 1994;5:389-393.