Nitric Oxide–Dependent and Nitric Oxide–Independent Transcriptional Responses to High Shear Stress in Endothelial Cells
Shear stress modulates gene expression in endothelial cells (ECs) partly through nitric oxide (NO), acting via enhanced cGMP formation by guanylyl cyclase (GC). We addressed non-cGMP–mediated transcriptional responses to shear stress in human umbilical ECs subjected to high-laminar shear stress (25 dyn/cm2; 150 minutes). RNA was isolated, reverse-transcribed, Cy3/5-labeled, and hybridized to 19 K human microarrays. High shear (n=6), high shear with 100 μmol/L l-NAME (n=3), and high shear with 10 μmol/L ODQ (GC inhibitor) in the perfusate (n=3) was compared with samples not subjected to flow. Among genes responding to high shear were HMOX1 (up) and PPARG (down). A high percentage of gene expression modulation by shear was absent during concomitant l-NAME or ODQ. Several transcriptional modulators were found (up: SOX5, SOX25, ZNF151, HOXD10; down: SOX11); a number of genes were regulated by shear and by shear with ODQ, but not regulated during l-NAME, indicating a nitric oxide synthase (NOS)–dependent, guanylyl cyclase (GC)–independent pathway. Several genes only responded to shear stress during l-NAME, others only responded to shear during ODQ. Upstream binding site analysis indicated shear stress and NO-dependent regulation of transcription via SOX5 and SOX9. Although NO importantly modulated the effect of shear stress on EC transcription, HMOX1 was consistently induced by shear stress, but not dependent on NOS or GC. Using bio-informatics software and databases, a promoter analysis identified SOX5 and SOX9 as potential, novel, shear-sensitive, and NO-dependent transcriptional regulators. The role of HMOX1 as a potential backup for NOS and the downstream role of SOXes should be explored.
Endothelial cells (ECs) form the interface between blood and vascular wall and are highly sensitive to physicochemical alterations in the circulation. EC dysfunction is associated with a large number of undesirable consequences for the entire vascular wall. Although an exact definition of EC dysfunction is difficult to give, ingredients of maintenance of vascular wall integrity include intact vasomotion, control of growth, adhesion and proliferation, and normal barrier function.1 Many studies have investigated the consequences of EC-derived nitric oxide (NO) on vascular smooth muscle cell contraction and other vascular smooth muscle functions. Few studies have addressed the actions of NO on EC function and gene transcription.2 Understanding how the EC copes with shear can help to discriminate whether deterioration is caused by aberrant transcriptional responses of the EC, or by other components of the vascular wall.
Shear stress evokes rapid induction of gene transcription, particularly of early growth response 1(Egr-1) and NF-κB. These transcription factors then initiate a cascade of transcriptional events.2 Multiple studies report that NO can dampen EC gene transcription on activation by factors such as IL-1B,3 tumor necrosis factor-alpha,3 and lipopolysaccharide.4 We demonstrated that NO quickly depresses the expression of many genes in quiescent EC, among them several transcription factors.5 Despite all this knowledge, it is currently largely unresolved which of the transcriptional responses of elevated shear stress are mediated by NO. Some studies investigating individual genes have demonstrated that NO mediates gene expression evoked by shear stress;6,7 other studies have demonstrated a dampening effect of NO. Moreover, the exact pathways by which NO transmits its signals to the genome have only been partly elucidated. The activation of soluble guanylate cyclase by NO causes the formation of cGMP and activation of cGMP-dependent kinases and cGMP-dependent calcium channels.8,9 cGMP-independent modulation of intracellular signaling by NO seems to occur by nitrosylation of proteins by NO.10 It is unclear whether a relevant cGMP-independent transcriptional response occurs in response to shear stress.
We hypothesized that short-term high shear by laminar flow activates the EC and that the transcriptional response is importantly modulated by NO, and that NO signaling largely depends on cGMP formation. We also explored whether backup systems independent from NO could be recognized and whether such factors have different transcriptional activators than NO-dependent shear stress–induced genes.
Materials and Methods
The methods for the flow system were adapted from Dekker et al.11 Briefly, Cellmax (catalog number 400-025; Spectrum Europe, Breda, the Netherlands) were seeded with human umbilical vein ECs and connected to a peristaltic pump. In the system, up to 5 cartridges can be run simultaneously. For details, please see http://hyper.ahajournals.org. Figure 1 shows a schematic of the system (Figure 1A) and the protocol (Figure 1B).
Shear Stress Experiments
On the day of the experiment, flow was redirected to the lumen; systems were always compared with a system in which the flow was pericapillary, thus not rerouted to the lumen. Flow was started at 2.5 mL/min, generating a shear stress of 2.5 dynes/cm2 and increased over a 15-minute period to the desired shear stress. Flow through the cartridges was increased to 12.5 mL/min and 25 mL/min, corresponding to shear stress of ≈12.5 and 25 dyne/cm2. The Reynolds number at 25 mL/min in each capillary is ≈32, which is not associated with turbulent flow. To address whether high shear stress increased the expression of endothelial nitric oxide synthase (eNOS) and Egr-1, the flow rate was maintained at 12.5 and 25 mL/min for 60 and 120 minutes. For the gene array experiments, flow was maintained for 2.5 hours at 25 mL/min. For the microarray analyses, a cartridge subjected to high flow (25 mL/min) was compared with a cartridge not subjected to transluminal flow (thus subjected pericapillary flow) in a total of 6 independent experiments. In 3 of these experiments, 100 μmol/L l-NAME (Sigma Aldrich, St. Louis, Mo) was added to the perfusate of a separate tubing system and cartridge subjected to high flow at the initiation of the high flow. In the other 3 experiments, 10 μmol/L of the guanylate cyclase inhibitor 1H-[1,2,4]oxadiazolo-[4,3–2]quinoxalin-1-one (ODQ) (Sigma Aldrich) was added to the perfusate in a similar fashion.
Visualization of ECs in Capillaries of the Cartridges
For cell fixation, the cartridges were perfused with formaldehyde, cut open, and the fibers were removed, cut, embedded in paraffin, and 10-μm longitudinal and transverse sections were stained with hematoxylin and eosin and with von Willebrand factor antibody.
Harvesting of the Cells and Isolation of RNA
At the end of the experiment, each flow cartridge was repeatedly flushed with Trizol reagent and the collected suspension was immediately stored at −20°C. RNA was isolated according to the manufacturer’s instructions, quantitated using a spectrophotometer (Shimadzu Scientific Instruments, Columbia, Md), and stored at −80°C.
Reverse-Transcriptase Polymerase Chain Reaction to Assess Endothelial NOS and Egr-1 Gene Expression and Confirm Array Experiments
Reverse-transcriptase (RT) was executed using an RT kit after the instructions (Lifetech 11904 to 018; Invitrogen, Carlsbad, Calif). Primer pairs and optimized polymerase chain reaction (PCR) conditions are listed online in Table I.
Statistics and Bioinformatics
Hierachical and K-means clustering was applied by use of the software developed by Jaak Vilo (Expression Profiler; http://ep.ebi.ac.uk/EP/). Gene Ontology Consortium classes of biological processes were coupled to the genes interrogated by the microarray using a proprietary database routine in Microsoft Access (Microsoft Corporation, Redmond, Wash) database. To assess in which groups of genes regulation was prominent, the percentage of genes were counted in each group (with at least 6 genes) that exceeded a log2 ratio >0.7 or <−0.7; because of the differences in the total numbers of genes in a group, a group was marked as regulated if >10% of genes were regulated and the group had >25 genes or if 15% of genes were regulated when the group contained <25 genes.
To assess whether the genes sensitive and insensitive to shear stress had different frequencies of transcription factor binding sites in their promoter regions, and whether NO-dependent transcriptional activation or repression (either via cGMP or independent from cGMP) involved different transcription factors, we analyzed the 1000-bp upstream regions of these genes. First, the upstream sequence was obtained via TRASER (http://genome-www6.stanford.edu/cgi-bin/Traser/traser). This sequence was subjected to transcription factor binding site analysis using MATCH,12 and matrix and core similarity cutoffs were maximized by the program. The frequency of binding sites for transcription factors was compared with a frequency table generated from 500 genes that were not regulated, ie, in the center of the log2 distribution.
Data and procedures of the microarray experiments were submitted in MIAME format13 to the European Bioinformatics Institute. Accession of the experiment numbers are listed online in Table II.
Visualization of ECs in the Capillaries
Cross-sections of one of the capillaries obtained from the flow system show that ECs cover the luminal wall (Figure 2), which was the case for most capillaries. Immunohistochemistry with anti–von Willebrand factor confirmed the presence of EC.
High Shear Stress Rapidly Induces Egr-1 and eNOS
In pilot experiments, Egr-1 gene expression was induced by 12.5 dyn/cm2 and 25.0 dyn/cm2 shear stress rates for 60 minutes. RT-PCR revealed transient and shear stress level dependent induction of Egr-1 (Figure 3A). eNOS was induced after 2.5 hours of shear at 25 dyn/cm2. Interestingly, induction of eNOS was inhibited by 100 μmol/L l-NAME (Figure 3B).
High Shear Stress Induces and Represses Multiple Genes
High shear stress induced 541 genes and depressed 436 genes (n=6). A selection of the differentially expressed genes with a known function on high shear stress is shown in Tables III and IV. The number of experiments in which a gene ratio did not exceed the cutoff for the background, the number of experiments in which the ratio exceeded 0.7, and the percent of valid ratios that exceeded 0.7 are also indicated. The complete set of regulated genes is at www.nephrogenomics.net/data/appendices/Shear2004. There was considerable variation between the different experimental conditions. Marked induction was observed for heme oxygenase 1 (HMOX1), phospholipase A2 (PLA2G4B), and complement component 1S. The slight induction of IL-8 (log2 ratio 0.62) was included because of its relevance. Induction of HMOX1 and IL-8 was confirmed by RT-PCR (Figure 4). Analysis of biological processes with the GO Consortium (web appendix Figure V) points at regulation in inflammatory and defense processes, indicating EC activation.
Inhibition of NOS Strongly Dampens Shear-Induced Gene Expression
With 100 μmol/L l-NAME (n=3), a large fraction of the differentially expressed genes by shear was not regulated any more (87% when compared with the corresponding 3 experiments without l-NAME). The tendency for more prominent induction than repression of gene expression by flow was inverted (Figure 5A). Figure 5B only includes those genes that exceeded background in both comparisons. Web Tables III and IV show known genes that were maximally differentially expressed by flow and the response to l-NAME. In the presence of l-NAME, shear still consistently induced expression of HMOX1. A relatively small number of genes that were not affected by flow, however, became induced during flow with concomitant NOS inhibition. High shear stress with concomitant NOS inhibition repressed inducible NOS (log2 ratio −0.71 versus −0.15 during flow without l-NAME) and C-reactive protein expression (log2 ratio −-0.92 versus −0.27 during flow without l-NAME). l-NAME (n=2) added to a cartridge without flow barely affected gene expression (101 modulated; data not shown).
Inhibition of GC Does Not Affect Gene Expression in a Similar Fashion as NOS Inhibition
Inhibition of GC during flow also affected gene expression; however, the tendency for more pronounced upregulation than downregulation persisted (Figure 6A). Of the regulated genes by flow, 92% were not regulated during concomitant ODQ, compared with the corresponding 3 experiments without ODQ (Figure 6B). Note that the experiments with ODQ and high flow were compared with their respective controls. Interestingly, a relatively large number of genes became regulated during ODQ and flow. Note that comparison of the experiments with ODQ in the perfusate was performed with the experiments from the same dates without ODQ. Web appendix Tables III and IV show known genes that were maximally differentially expressed by flow and the response to ODQ. Remarkably, HMOX1 expression was still induced by flow during ODQ administration (log2 ratio 0.78 versus 1.70 during flow without ODQ).
Genes Regulated by Flow Through NO and cGMP
Genes acting via NOS and GC were selected as those genes with an absolute log2 ratio exceeding 0.7 during shear and log2 ratios during NOS inhibition and GC inhibition that were >0.5 less. Using the GO classification, NO-dependent and cGMP-dependent upregulation of genes related to cell structure and cell–cell contact was recognized. Genes downregulated by flow through NOS and via GC revealed some prominent responses of cadherins and associated proteins (procadherin beta 8, desmoglein 1, N-cadherin 2; Table 1). Furthermore, several apoptosis genes were depressed by flow through NO (BAX, ICEBERG, DAP, DAPK3, CD38). Finally, PPARG was depressed by flow through NO and cGMP, whereas the PPARG coactivator 1 was increased. Intriguing is the complex response observed in the SOX family of transcriptional regulators: SOX28 and SOX29 were decreased by flow through NO, and SOX25 and SOX26 were increased by flow. The full data set is available online.
Genes Regulated Independent of NO and cGMP
For a few important genes, induction (HMOX1, USF1, Heat shock 70kD protein 1B, ADAMTS1) or suppression (Nuclear RNA export factor 3, Signal sequence receptor, gamma, SLC28A1, Absent in melanoma 2, Hypothetical protein FLJ2184) was completely NOS-independent and cGMP-independent. The most consistently regulated gene by shear stress was HMOX1. In 6 of 6 experiments, HMOX1 was induced by flow; in 3 of 3 experiments, this was unaffected by l-NAME; and in 2 of 3 experiments, this was unaffected by ODQ.
K-Means Clustering Reveals Alternative Transmission Pathways
To analyze other potential pathways mediating the transcription induced by shear, K-means analysis with 20 clusters was performed (Figure 7). Genes were selected that displayed at least 1 absolute ratio >0.7. Inspection of the clusters reveals that besides patterns that are compatible with the hypotheses that flow (and shear stress) can act via NO either via cGMP or not via cGMP to induce or repress gene expression, other patterns are also observed. Flow-mediated depression (cluster 15) and activation (cluster 12 and 16) of gene expression independent from NOS but dependent on GC is recognized. Clusters 6 and 18, and 7 and 14, are sets of genes not modulated by shear stress with and without NOS inhibition, but activated and depressed during GC inhibition. The database was also interrogated to identify genes regulated by shear and by shear with ODQ, but not regulated by l-NAME (hidden in several clusters), indicating a NOS-dependent GC-independent pathway.
Analysis of Transcription Factor Binding Sites
Table 2 displays the most pronounced differences in binding sites between shear stress–regulated genes and shear-regulated genes that were modulated via NO. Besides more frequent binding sites for NF-κB in the NO-dependent genes, the presence of 2 SOX proteins, which have a more pronounced representation in genes regulated via NO, is remarkable. Figure 8 shows the comparison of the representation of binding sites for transcription factors in genes modulated by shear through NO acting via GC or not acting via GC. Of interest is that SOX5, SOX9, and PAX6 binding sites were more frequently observed in genes responsive to shear, but not modulated via GC, than in genes modulated through GC. Such a difference could also be observed for several other transcription factors. Note the rather high difference in the presence of binding sites for SOX5 and SOX9 between shear-insensitive genes and shear-sensitive genes modulated by NO but not via GC (32% versus 47%, 39% versus 54%).
It was addressed which part of the transcriptional response of the EC to an acute increase in shear stress is mediated by NO, whether NO acts via GC, and which transcription factors could be involved in the response to shear stress. High laminar shear stress resulted in a more pronounced induction than depression of gene expression. Concomitant exposure to high shear and NO synthesis inhibition prevented the modulation of a large number of genes. Moreover, genes became regulated by shear when NOS was inhibited. Inhibition of GC also strongly affected the response to laminar shear stress, indicating that part of the transmission of shear is via GC. Sets of genes could be identified that acted via NOS and via GC, and a set that was NO-dependent but not GC-dependent. Interestingly, heme oxygenase 1 was very consistently induced by high shear and not dependent on either NO or cGMP. Using K-means cluster analysis, we identified groups of genes that were modulated by shear, during NOS inhibition, which, however, were not regulated during GC inhibition. Very few shear stress–modulated genes were not affected by l-NAME as well as ODQ. Using bioinformatics tools, we compared the potential transcription factor binding sites in the promoter regions of shear stress–sensitive and shear stress–insensitive genes, and in the promoter regions of shear stress–sensitive and NO-dependent and NO-independent genes. Interestingly, 2 FOX transcription factor-binding sites (FOXL1, FOXJ2) were more frequently found in shear-sensitive genes. Two SOX binding sites (SOX5, SOX9) were more frequently found in the shear-sensitive, NO-dependent genes. The present transcriptional analysis of EC points to a diversity of pathways mediating the acute response to laminar shear, which are likely to be associated with different transcriptional regulators.
The EC is the interface between blood and the vascular wall and serves as pivotal regulator of the maintenance of vascular wall. ECs have been found to be highly sensitive to a variety of stimuli, such as cytokines,3 lipopolysaccharide,4 and shear stress.2 A variety of shear stress targets has been discovered and some of the signaling pathways have been investigated.2 In recent years, several studies applying expression profiling on ECs11,14–17 have underscored that the EC gene expression is highly responsive to shear stress, differentially affected by turbulent and laminar flow-induced shear,11,16 and is dependent on duration of exposure to shear.11 These studies not only confirmed the regulation of genes that had been previously identified as shear-responsive but also identified new candidates mediating EC responses to shear. Recently, 2 studies have also successfully correlated the expression of shear-responsive genes to vascular endothelial growth factor, thrombin,11 tumor necrosis factor-alpha, IL-1β, and transforming growth factor-β1 induced expression in ECs.11,14 We have extended these observations by analysis of NO-sensitive and NO-insensitive expression regulation and analysis of 1 of the targets of NO, GC. Acute, laminar, short-term shear stress induced a large number of genes. Three genes have been described by Dekker et al,11 using a similar setup, as being induced after 6 hours of similar shear stress: diaphorase 4 (log2 ratio 0.40), cytochrome P450, subfamily I, polypeptide 1 (CYP1B1, log2 ratio 0.81), and the transcription factor Kruppel-like factor 2 (log2 ratios 0.68 and 1.27). Note that we evaluated the expression of these genes at 2.5 hours. Certainly there are many differences among the various published data sets, and between the present and the published data. It should be noted that we studied acute modulation of gene expression to high shear stress, the current capillary tube-based model differs from the cone and plate models in their characteristics and the microarrays are different.
One of the aims of this study was to assess whether NO dampened shear stress–induced EC transcription. An interesting observation in this respect is shear stress–induced eNOS expression via NO and the decrease in expression of inducible NOS on shear on the microarray. Many of the shear-induced changes were mediated via NO. Still, a set of genes could be identified that only were induced or suppressed by shear stress in the presence of NOS inhibition (Figure 5). Under static conditions, NO has been shown to inhibit endothelin-1,18 E-selectin, and intercellular adhesion molecule-119 gene expression in EC. Furthermore, exogenous NO has been demonstrated to inhibit lipopolysaccharide-induced expression of tissue factor.4 We have recently shown that the NO donor DETA-NONOate depressed both RELB and the p100 precursor of NF-κB2 (p49).5 Nevertheless, there was no predominant dampening effect of NO on shear-induced gene expression in the current experiment. NO can function independently from GC and the resulting cGMP formation. Evidence is rapidly accumulating for the hypothesis that NO can nitrosylate proteins and that protein nitrosylation is a regulated process.10 The literature contains scattered data about cGMP-independent endothelial gene expression. Sata et al have demonstrated that NO inhibited apoptosis induced by serum deprivation of human umbilical vein endothelial cells, and that this effect was not affected by soluble guanylate cyclase inhibition using ODQ.20 Support is available that this antiapoptotic effect of NO is mediated by nitrosylation of caspase-3 and is cGMP-independent.21 Besides protein modification, NO can cause acute cGMP-independent increase in EC calcium levels.22 The activity of important EC transcription factors such as NF-κB is influenced by NO, perhaps to some extent via the formation of peroxynitrite.23 We identified multiple genes that were regulated via NO but not via GC, indicating the relevance of this pathway (Figure 7).
HMOX1 was strongly and consistently induced and not affected by NOS or GC inhibition. HMOX1 reduces heme to biliverdin and liberates carbon monoxide that acts via GC to form cGMP.24 Besides its function to act as an antioxidant system, HMOX1 induction has been suggested as a backup for the NO system.25 HMOX1 has been shown to be induced by low-laminar and turbulent shear stress in human EC,26 and by high-laminar flow.24 Interestingly, NO has been shown to induce HMOX1 in vascular smooth muscle cells and ECs27,28 and act via cGMP.28 The induction by shear stress could be prevented by concomitant administration of the scavenger substance N-acetylcysteine.26 As mentioned, HMOX1 induction was independent from NO and GC. One could argue that to serve as a backup for NO, induction via NO does not seem logical. With respect to HMOX1 and the possibility of CO/cGMP signaling, it is intriguing that in the K-means cluster analysis, subsets of genes could be recognized that were not induced or repressed by shear stress, which could not be inhibited by NOS inhibition; however, they were normalized during GC inhibition. Between the genes induced by flow and silenced by GC inhibition were the tight junction proteins claudin 4 and 14 and profilin 2, a protein involved in actin multimerization, the small inducible cytokine subfamily C, member 1 (lymphotactin), monocytes chemotactic protein 1, and phospholipase A2, group IVB (cytosolic). Together, these observations may indicate that a substance other than NO, possibly CO, activates GC on shear stress and induces genes that are involved in maintenance of EC structure and in inflammation.
Interestingly, the frequency of binding sites in the shear-dependent and NO-sensitive genes for 2 members of the SOX family of transcription factors and for 1 of the factors that associates with SOX, Pax6, was increased. SOX5 and SOX9 were indeed expressed in the present EC study, and RT-PCR confirmed that shear stress increased SOX5 expression. The SOX family features a 3-hydroxy-3-methylglutaryl box motif and are unique to animals.29 The proteins are widely distributed in developing tissues30 and act in concert with partners as DNA-binding proteins, adaptor proteins, and nuclear import proteins.29 The present data suggest that shear-induced NO actions and SOX proteins are related, because the frequency of SOX5 and SOX9 binding was not more frequent in the NO-insensitive genes as compared with shear-insensitive genes. There is little information about SOX in EC; however, a recent study indicates the presence of SOX9 in human calcified arteries.31 Interestingly, SOX9 expression has been shown to be extremely sensitive to IL-1 and tumor necrosis factor-alpha,32 2 factors that are known to strongly affect gene expression in EC. SOX18 has been shown to be involved in the embryonic development of the cardiovascular system, is localized in ECs,33 and has been shown to interact with the transcriptional regulator MEF2C in mouse ECs.34 Thus, we have indications for a shear stress–dependent and NO-dependent regulation of genes via SOX5 and SOX9, a matter that deserves further investigation, because these transcription factors are involved in calcification.
In this study, we have used gene expression profiling to investigate NO and GC dependency of transcriptional responses to acute elevations in shear stress. The modulation of expression of genes by acute high shear stress was, to a large extent, dependent on NO. A dampening effect of NO on EC function could be demonstrated; however, NO also mediated the expression of many genes. Interestingly, heme oxygenase 1 was very consistently induced by flow and could form a backup mechanism for NO. By using bio-informatics software and databases, a promoter analysis of the shear-sensitive and NO-dependent and NO-independent genes revealed prominent presence of SOX5 and SOX9 binding sites in NO-sensitive genes. We speculate that such pathway analysis could accelerate the identification of potentially important transcription factors in the development of endothelial dysfunction and atherosclerosis.
The work of B.B. is supported by a fellowship from the Royal Dutch Academy of Arts and Sciences. The technical assistance of Adèle Dijk and Despina Xanthakis is greatly appreciated.
- Received October 9, 2004.
- Revision received November 3, 2004.
- Accepted December 17, 2004.
Braddock M, Schwachtgen JL, Houston P, Dickson MC, Lee MJ, Campbell CJ. Fluid shear stress modulation of gene expression in endothelial cells. News Physiol Sci. 1998; 13: 241–246.
De Caterina R, Libby P, Peng HB, Thannickal VJ, Rajavashisth TB, Gimbrone MA, Jr, Shin WS, Liao JK. Nitric oxide decreases cytokine-induced endothelial activation. Nitric oxide selectively reduces endothelial expression of adhesion molecules and proinflammatory cytokines. J Clin Invest. 1995; 96: 60–68.
Braam B, De Roos R, Dijk A, Boer P, Post JA, Kemmeren PP, Holstege FC, Bluyssen HA, Koomans HA. Nitric oxide donor induces temporal and dose-dependent reduction of gene expression in human endothelial cells. Am J Physiol Heart Circ Physiol. 2004; 287: H1977–H1986.
Bao X, Lu C, Frangos JA. Temporal gradient in shear but not steady shear stress induces PDGF-A and MCP-1 expression in endothelial cells: role of NO, NFκB, and egr-1. Arterioscler Thromb Vasc Biol. 1999; 19: 996–1003.
Chiu JJ, Wung BS, Hsieh HJ, Lo LW, Wang DL. Nitric oxide regulates shear stress–induced early growth response-1: expression via the extracellular signal–regulated kinase pathway in endothelial cells. Circ Res. 1999; 85: 238–246.
Dekker RJ, van Soest S, Fontijn RD, Salamanca S, de Groot PG, VanBavel E, Pannekoek H, Horrevoets AJ. Prolonged fluid shear stress induces a distinct set of endothelial cell genes, most specifically lung Kruppel-like factor (KLF2). Blood. 2002; 100: 1689–1698.
Kel AE, Gossling E, Reuter I, Cheremushkin E, Kel-Margoulis OV, Wingender E. MATCHTM: a tool for searching transcription factor binding sites in DNA sequences. Nucl Acids Res. 2003; 31: 3576–3579.
Brazma A, Hingamp P, Quackenbush J, Sherlock G, Spellman P, Stoeckert C, Aach J, Ansorge W, Ball CA, Causton HC, Gaasterland T, Glenisson P, Holstege FC, Kim IF, Markowitz V, Matese JC, Parkinson H, Robinson A, Sarkans U, Schulze-Kremer S, Stewart J, Taylor R, Vilo J, Vingron M. Minimum information about a microarray experiment (MIAME)-toward standards for microarray data. Nat Genet. 2001; 29: 365–371.
Wasserman SM, Mehraban F, Komuves LG, Yang RB, Tomlinson JE, Zhang Y, Spriggs F, Topper JN. Gene expression profile of human endothelial cells exposed to sustained fluid shear stress. Physiol Genomics. 2002; 12: 13–23.
Chen BP, Li YS, Zhao Y, Chen KD, Li S, Lao J, Yuan S, Shyy JY, Chien S. DNA microarray analysis of gene expression in endothelial cells in response to 24-h shear stress. Physiol Genomics. 2001; 7: 55–63.
Garcia-Cardena G, Comander J, Anderson KR, Blackman BR, Gimbrone MA, Jr. Biomechanical activation of vascular endothelium as a determinant of its functional phenotype. Proc Natl Acad Sci U S A. 2001; 98: 4478–4485.
Brooks AR, Lelkes PI, Rubanyi GM. Gene expression profiling of human aortic endothelial cells exposed to disturbed flow and steady laminar flow. Physiol Genomics. 2002; 9: 27–41.
Zampolli A, Basta G, Lazzerini G, Feelisch M, De Caterina R. Inhibition of endothelial cell activation by nitric oxide donors. J Pharmacol Exp Ther. 2000; 295: 818–823.
Sata M, Kakoki M, Nagata D, Nishimatsu H, Suzuki E, Aoyagi T, Sugiura S, Kojima H, Nagano T, Kangawa K, Matsuo H, Omata M, Nagai R, Hirata Y. Adrenomedullin and nitric oxide inhibit human endothelial cell apoptosis via a cyclic GMP-independent mechanism. Hypertension. 2000; 36: 83–88.
Rossig L, Fichtlscherer B, Breitschopf K, Haendeler J, Zeiher AM, Mulsch A, Dimmeler S. Nitric oxide inhibits caspase-3 by S-nitrosation in vivo. J Biol Chem. 1999; 274: 6823–6826.
Chen X-L, Varner SE, Rao AS, Grey JY, Thomas S, Cook CK, Wasserman MA, Medford RM, Jaiswal AK, Kunsch C. Laminar flow induction of antioxidant response element-mediated genes in endothelial cells. A novel anti-inflammatory mechanism. J Biol Chem. 2003; 278: 703–711.
Rabelink TJ, Stroes E. Atherosclerosis: defeat of the defense? Circ Res. 2001; 88: 456–457.
De Keulenaer GW, Chappell DC, Ishizaka N, Nerem RM, Alexander RW, Griendling KK. Oscillatory and steady laminar shear stress differentially affect human endothelial redox state: role of a superoxide-producing NADH oxidase. Circ Res. 1998; 82: 1094–1101.
Durante W, Kroll MH, Christodoulides N, Peyton KJ, Schafer AI. Nitric oxide induces heme oxygenase-1 gene expression and carbon monoxide production in vascular smooth muscle cells. Circ Res. 1997; 80: 557–564.
Polte T, Abate A, Dennery PA, Schroder H. Heme oxygenase-1 is a cGMP-inducible endothelial protein and mediates the cytoprotective action of nitric oxide. Arterioscler Thromb Vasc Biol. 2000; 20: 1209–1215.
Tyson KL, Reynolds JL, McNair R, Zhang Q, Weissberg PL, Shanahan CM. Osteo/chondrocytic transcription factors and their target genes exhibit distinct patterns of expression in human arterial calcification. Arterioscler Thromb Vasc Biol. 2003; 23: 489–494.
Murakami S, Lefebvre V, de Crombrugghe B. Potent inhibition of the master chondrogenic factor Sox9 gene by interleukin-1 and tumor necrosis factor-alpha. J Biol Chem. 2000; 275: 3687–3692.