1 Department of Bioengineering and the Whitaker Institute of Biomedical Engineering, University of California, San Diego, La Jolla 92093-0427
2 Division of Biomedical Sciences, University of California, Riverside, Riverside, California
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
DNA microarray; endothelial cells; gene expression; shear stress
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Hemodynamic forces regulate the structure and function of the blood vessel wall. The endothelial cells (ECs) lining the blood vessels are exposed to the shear stress resulting from the tangential forces exerted by blood flow. Previous in vitro and in vivo findings have shown that a sudden onset of shear stress activates EC signaling pathways and modulates EC gene expression (for review see Refs. 5, 7, 19, 24). In these studies, the flow-modulated genes have been investigated mainly by using Northern blotting to detect the expression of mRNA, and the translated proteins have been detected by the use of immunoblotting, enzyme-linked immunosorbent assay (ELISA), or radioimmunoassay (RIA). There have been efforts to compare the shear stress-induced gene regulation by using the differential display strategy (3, 50). Although differential display is useful for the identification of shear-stress-responsive genes, it can only cover a limit number of genes, with laborious procedures and relatively low reproducibility. The DNA microarray technology provides an efficient tool to rapidly compare the differential expression of a large number of genes in a single test, and this is ideally suited for studying the pattern of gene expression in ECs induced by various experimental conditions (4).
There is increasing evidence that the gene expression induced by laminar shear stress is time dependent. The sudden application of shear stress to ECs previously in static culture causes the transient activation of a large number of genes, but most of these genes become downregulated during the sustained application of shear stress, e.g., for 24 h (5). In the current study, we used the DNA microarray technology to investigate the EC gene expression profiles in response to 24-h shear stress. The aims are to confirm and expand previous knowledge with this relatively new technology for the elucidation of the genomic programming of shear-adapted ECs.
![]() |
MATERIALS AND METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Shear stress experiments.
A flow system was used to impose shear stress on cultured ECs as described by Frangos et al. (15). In brief, a 10 x 8-cm glass slide seeded with a confluent monolayer of HAECs was mounted in the flow channel created by sandwiching a silicone gasket between the glass slide and an acrylic plate, with inlet and outlet for exposing the cultured cells to the shear stress imparted by flow. The channel has dimensions of 0.25 mm in height, 6 cm in width, and 7.5 cm in length. A laminar shear stress of 12 dyn/cm2 was generated by the flow resulting from a hydrostatic pressure difference between two reservoirs. This level of shear stress is encountered under physiological conditions in the straight part of the aorta and frequently used to study the effects of shear stress on ECs. The flow system was kept at 37°C and ventilated with 95% humidified air and 5% CO2. The same batch of HAECs at passage 5 was used for three independent shear experiments.
cDNA labeling, microarray, hybridization, and imaging.
Three independent sets of experiments were performed, each consisted of an experiment with HAECs subjected to 24-h shear stress and a paired control experiment with HAECs kept under static condition for 24 h. In each experiment, the total RNA was isolated by using STAT60 total RNA purification reagent (TEL-TEST "B"). The cells were lysed in phenol-containing STAT60 solution and centrifuged. The RNA-containing aqueous phase was isolated, and 0.6 vol of isopropanol was added to precipitate RNA. Thirty micrograms of total RNA was subjected to reverse transcription (RT) reaction and [33P]dCTP labeling using an Atlas Pure Total RNA Labeling System (Clontech). The 33P-labeled cDNA was hybridized to Atlas Human Cancer 1.2 arrays at 68°C for 16 h, followed by washing once in 2x SSC/1% SDS at 68°C for 30 min and three times in 0.1x SSC/0.5% SDS at 68°C for 30 min. The hybridized array was then exposed to Phospho Storage Screen (Molecular Dynamics), and the images were analyzed and quantified by using an AtlasImage 1.01 software (Clontech) to quantify the intensity of each of the 1,185 spots. The Atlas 1.2 array was chosen because it contains many genes involved in growth, proliferation, cell cycle regulation, cell structure remodeling, and inflammatory responses, as well as signal transduction, which were the functions of interest in our investigation. Furthermore, it does not require too much RNA for the hybridization and it is relatively inexpensive.
Data analysis and statistics.
Each Atlas 1.2 array consists of four separate filters. Two filters from each set were used to perform the experiments reported in this study, one for static condition and another for 24-h shearing; the other two filters were used for other experiments. A total of three sets of filters were used, thus providing three sets of experimental data, each with a static and a shear experiment from the same set of filters. The background was determined from the intensity reading around the spotted areas by using the AtlasImage Software and subtracted from the reading of the spot. Comparisons were made between the static and shear filters from the same set; the results obtained from each filter were normalized by using the mean intensity as the reference, i.e., global normalization. The normalized results of the sheared and static filters from the same set were used to calculate the ratio for each gene in each set. The ratios for each gene from three independent experiments (i.e., for the three sets of filters) were used to perform paired Students t-test to assess the significance of the difference of the mean ratio from unity. A P value of 0.05 or lower was taken to indicate a statistically significant difference of the shear/static ratio from unity. No normalization was performed between filters. The shear/static ratios are reported as means ± SE.
RT-PCR.
RT-PCR was carried out by using the following procedures. For RT reaction, 5 µg of total RNA and 0.5 µg of oligo(dT) in 20 µl diethyl pyrocarbonate-H2O were incubated at 70°C for 5 min, and cooled slowly at room temperature for 10 min. To synthesize the first-strand cDNA, 2.5 µl of 10x first-strand buffer, 0.5 µl of RNase inhibitor (40 U/µl), 1 µl of 100 mmol/l dNTPs, and 0.5 µl of Moloney murine leukemia virus reverse transcriptase (50 U/µl) were added. The samples were incubated at 37°C for 1 h, and then at 90°C for 5 min to stop the reaction. Volumes of 0.1 or 0.5 µl of each RT reaction were mixed with 50 µl of a PCR mixture containing 1x PCR buffer (GIBCO BRL), 1.5 mmol/l MgCl2, 200 µmol/l dNTPs, 30 pmol each of 5' and 3' primers, and 2.5 U of Taq DNA polymerase (GIBCO BRL). The PCR reaction was carried out by heating for 3 min at 94°C, followed by 20 to 35 cycles of denaturation (94°C, 45 s), annealing (range from 55 to 60°C, 45 s), and extension (72°C, 1 min), and subjected to final extension for 10 min at 72°C using a Primer 96 thermocycler (MHG Biotech). The PCR products were analyzed by 1% agarose gel electrophoresis. Because of variations in the abundance of mRNA of different genes in the cell, the PCR cycles and titrations of RT reaction were varied to ensure that the PCR reaction was in the linear range.
Northern blot analysis.
The microarray results were selectively confirmed by Northern blot analysis. The cDNA fragments of the selected genes were used as probes labeled with [32P]dCTP by random primer labeling. Northern hybridization was performed according to standard protocol (43). Twenty micrograms total RNA from static or sheared samples was electrophoresed on a formaldehyde-containing 1.5% agarose gel. The RNA was transferred to a nylon membrane for hybridization with the 32P-labeled cDNA. The hybridized RNA was detected by autoradiography.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
Shear stress regulates the expression of genes related to inflammation.
Exposure of ECs to proinflammatory cytokines [such as interleukin-1 (IL-1) and tumor necrosis factor (TNF)] results in the modulation of "prothrombotic-proinflammatory" program (31, 32). In the current microarray study, several proinflammatory and prothrombotic genes were significantly downregulated by 24-h shear stress, including MYD88 (0.37 ± 0.15, an adapter protein essential for signaling via the IL-1 receptor), CD30 ligand (0.44 ± 0.12, a member of the TNF superfamily and an inducer for T-cell proliferation), and platelet basic protein (0.47 ± 0.06, PBP, a chemokine for neutrophils and monocytes). Using RT-PCR analysis (Fig. 1A), we confirmed the decreased expressions of MYD88, CD30 ligand, and PBP genes in response to 24-h shear stress. The downregulation of MYD88 has been confirmed by Northern blotting (Fig. 2). In addition, we found that the expression of IL-1 receptor antagonist protein (0.51 ± 0.10), thymus-expressed chemokine (0.54 ± 0.07), IL-14 (0.64 ± 0.08), and IL-15 (0.67 ± 0.06) were also downregulated in response to 24-h shear stress, although to lesser degrees than the other cytokines mentioned above. The expression of other cytokines-related genes showed no significant change in response to 24-h shear stress. The only inflammation-related gene that showed an upregulation in response to 24-h shear stress is the IL-13 receptor gene (1.85 ± 0.21). It is interesting to note that IL-13 receptor and its ligand IL-13 have been implicated in anti-inflammatory responses (31, 51).
|
|
Shear stress regulates the expression of genes related to EC remodeling.
Shear stress plays an important role in modulating EC morphology, cytoskeleton, and extracellular matrix (ECM) remodeling (7). In our microarray study, a few cytoskeleton- and ECM-related genes were downregulated by the 24-h shearing, e.g., cytokeratin 4 (0.34 ± 0.04), cytokeratin 2E (0.37 ± 0.10), microtubule-associated protein (0.53 ± 0.11), laminin-4 (0.54 ± 0.08), and integrin subunits ß3 (0.54 ± 0.06) and
7b (0.55 ± 0.08). The expression of matrix metalloproteinase-1 (MMP1 or interstitial collagenase) was significantly upregulated (2.77 ± 0.35) in response to the 24-h shear stress, and this was confirmed by RT-PCR and Northern blotting (Figs. 1B and 2). The expression of MMP3 has been shown to be regulated in parallel to MMP1 in many different experiment settings (33, 34). However, the 24-h shearing did not affect the expression of MMP3 (0.94 ± 0.23) nor the metalloproteinase inhibitors TIMP1 (1.04 ± 0.14) and TIMP2 (0.81 ± 0.05).
Shear stress regulates the expression of genes related to signal transduction.
Shear stress, like chemical stimuli, induces many signaling pathways in ECs to modulate vascular biology (7). Our microarray expression study has shown that 24-h shear stress causes the selective modulation of a few signaling molecules. Two endothelium-specific receptor tyrosine kinases were significantly upregulated, i.e., Tie2, which is angiopoeitin-1 and -2 receptor (1.90 ± 0.01), and Flk-1/KDR, which is VEGF receptor 2 (1.51 ± 0.02). These increases were confirmed by RT-PCR analysis (Fig. 1C). Tie2 and Flk-1 have been shown to play important roles in vasculogenesis during embryonic development and the subsequent angiogenesis (14, 39). Endothelin-3 was downregulated by 24-h shear stress (0.52 ± 0.11).
The expression of caveolin-1, another signaling molecule, was significantly downregulated (0.42 ± 0.04) after 24 h of shearing, and this was confirmed by the RT-PCR analysis and Northern blotting (Figs. 1C and 2).
Other genes affected by 24 h of shear stress.
As shown in Table 1, the expression of several transcription- and metabolism-related genes was also modulated by 24-h shear stress, with most of them being downregulated. In addition, 24-h shear stress also modulated the expression of genes involved in RNA splicing (suppressor for yeast mutant, arginine/serine-rich splicing factor 8), protein translation and transportation (RNA binding protein PPTB-1, elongation factor 2, importin 3), development (lunatic fringe), and redox response (glutathione-S-transferase homolog, glutathione synthase), as well as some genes with unknown functions.
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Fluid shear stress exerts many effects on vascular ECs and plays an important role in maintaining the homeostasis of blood vessels. It has been shown that shear stress modulates the expression of EC genes encoding growth factors, adhesion molecules, coagulation molecules, chemoattractants, proto-oncogenes, and vasoactive substances (5, 7). For example, the platelet-derived growth factors (PDGFs), intercellular adhesion molecule-1, monocyte chemotactic protein-1 (MCP-1), and c-fos genes are all transiently upregulated by shear stress (23, 35, 42, 46). However, most of these studies reported the effects of short-term shear stress, which are different from those of longer periods of shearing. We have performed some preliminary studies using the same Atlas 1.2 Cancer array, on the short-term (1 h) effect of shear stress on gene expression profile. The results show an induction of genes such as PDGF A, PDGF B, EGF, tissue plasminogen activator (tPA), c-jun, and c-fos at this early time point, as previously reported in the literature. In addition, genes such as IL-1, IL-6, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and actin did not show a modulation by shear stress, as found in previous studies on individual genes.
Prolonged shearing can decrease EC turnover, including cell proliferation and apoptosis (1, 11, 12, 29). Since relatively long periods of shearing is more relevant to the flow condition in the athero-resistant, straight parts of the arterial tree, the current study focused on the effects of 24-h shear stress on the gene expression profiles in HAECs. The aim is to elucidate the genomic programming of ECs in response to sustained shear stress.
Long-term shear stress decreases EC DNA synthesis and arrests EC cell cycle at G1/G0 phase (6, 10, 29, 36). One of the mechanisms for the shear stress-induced EC arrest was through the posttranslational modifications of Rb and p53 (29). Indeed, the transcription levels of Rb and p53 were not affected in our DNA microarray study. Our findings that 24-h shear stress downregulates the expression of many cell proliferation genes, notably the genes related to growth factors (e.g., BMP4, TGF-ß2, HDGF, and FGF6), suggest that autocrine/paracrine mechanisms may be involved in the shear stress-induced EC growth arrest.
Two EC survival genes, Tie2 and Flk-1, are upregulated by the 24-h shear stress. Tie2 has been shown to be expressed not only in embryonic vasculogenesis and subsequent angiogenesis, but also in the entire adult vasculature (arteries, veins, and capillaries). It has been suggested that Tie2 plays an important role in the maintenance of the normal function of the adult quiescent vasculature (53). Both Tie2 and Flk-1 are connected with the downstream phosphatidylinositol-3-kinase (PI3K)/Akt survival pathway (18, 27). Thus the increased expressions of Tie2 and Flk-1 in response to 24-h shear stress may enhance the activity of their downstream PI3K/Akt pathway to promote EC survival.
Our findings of the downregulation of a number of inflammatory genes, including MYD88, CD30 ligand, PBP and several cytokines, indicate that shear stress may antagonize inflammatory responses through such gene modulation. Activation of the IL-13 pathway has been shown to decrease the expression of pro-inflammatory cytokines and chemokines (51). Therefore, the upregulation of the IL-13 receptor in response to 24-h shear stress also would lead to an anti-inflammatory action. Cytokines-mediated activation of ECs in many physiological or pathophysiological processes may enhance the interaction between EC and immunocompetent cells (e.g., monocytes) and lead to the disorganization of EC cell junctions to increase the passage of macromolecules such as low-density lipoprotein (LDL) (8, 9, 52). By decreasing inflammatory responses, shear stress could prevent the induction of the atherogenic "prothrombotic-proinflammatory" program and promote the normal function of ECs as a permeability barrier and nonthrombogenic surface lining the vascular wall (32).
In vivo and in vitro studies have suggested that shear stress promotes angiogenesis (2). In the current study, several lines of evidence are consistent with this concept. First, as mentioned above, 24-h shear stress upregulates the endothelium-specific Tie2 and Flk-1 receptor tyrosine kinases, which play crucial roles in both vasculogenesis and angiogenesis (14, 39). Second, the expression of MMP1 is upregulated in response to 24-h shear stress. Collagen is a major secreted ECM protein in the blood vessel. The degradation of collagen by MMP1 may initiate and promote angiogenesis (21, 25). Third, the expression of caveolin-1 is downregulated by 24-h shear stress. It has been shown that angiogenesis inducers [VEGF, basic FGF, hepatocyte growth factor (HGF)] cause dramatic reduction of the caveolin-1 mRNA level in human umbilical vein ECs (HUVECs) and that angiogenesis inhibitors block the VEGF-mediated downregulation of caveolin-1 (30).
Consistent with the current microarray result of the downregulation of caveolin-1 by 24-h shear stress, we have also found that immunostaining of caveolin-1 in cultured bovine aortic ECs decreases upon 24-h shear stress (S. Li and S. Chien, unpublished results). Caveolin-1 is the essential component for assembling the plasma membrane caveolae structure, which has a lipid-ordered membrane domain mainly composed of cholesterol and sphingolipids and has a rich content of signaling molecules (47). It has been proposed that the caveolae function as signaling centers or scaffolds for the preassembly of signaling complexes and the cross talk between various signaling pathways (47). Caveolin-1 also plays an important role in the interaction and recruitment of signaling molecules (37). The decreased expression of caveolin-1 in response to 24-h shear stress may lead to the decreases of caveolae structure and preassembled signaling complexes on the EC surface. Consequently, this could keep ECs in a less active state in their response to various forms of stimuli.
Long-term exposure of ECs to shear stress has been shown to cause the elongation and alignment of ECs along the flow direction, with rearranged microfilaments and focal adhesion sites, attenuated cortical F-actin, and enhanced central stress fibers (20, 28). Our microarray results showed that the expression of the majority of cytoskeleton- and ECM-related genes was not affected by 24-h shear stress. It is likely that shear stress induces cytoskeleton and ECM reorganization at the protein level, rather than the transcription level. Another possibility is that after 24-h exposure to shear stress, ECs have already adapted to the shearing condition, and the remodeling events have subsided. Our results did show the downregulation of a few genes encoding cytoskeleton/ECM proteins and integrins by 24-h shear stress. It remains to be established whether the downregulation of these genes plays a role in the shear stress-regulated EC remodeling.
In the current study, MMP1 is the only cytoskeleton/ECM-related gene significantly upregulated by 24-h shear stress. There is evidence that MMP1 may be involved in EC migration and wound healing. Shear stress can enhance EC migration and wound healing in both animal studies and cell cultures (40, 49, 54). MMP1 has been shown to promote wound healing by facilitating keratinocyte migration over the collagen-rich dermis during re-epithelialization (16). Hence, the increase in MMP1 expression by shear stress would be beneficial for EC migration and wound healing. In addition, Tie2 and Flk-1 have been connected to the PI3K signaling pathway, which not only plays a role in cell survival, but also is involved in cell migration. PI3K has been shown to be upstream of the Rho family small GTPases (13, 16). Through its regulation of Rho family GTPases, PI3K affects actin polymerization and focal adhesion formation and consequently influences cell migration (41). Taken together, the upregulation of MMP1, Tie2, and Flk-1 could be synergistic in promoting EC migration and wound healing under long-term shear stress.
There had been relatively little information on the effects of prolonged shear stress on gene expression. During the revision of this manuscript, the report by Garcia-Cardena et al. (17) was published to provide for the first time extensive information on the modulation of EC gene expression by 24-h shearing. They also found that the majority of the gene expression remained unchanged after 24-h laminar shear stress. They reported a set of experiments that consisted of control, laminar shear, and turbulent flow; no statistical analysis was given in the report. A number of the genes that showed a twofold change in Ref. 17 in response to laminar shear were not present in the microarray used in the present study. For example, the genes related to nitric oxide synthase-associated regulatory factors, which are important to understanding vascular tone regulation and other functions, were not present in the Atlas filter. Some of the genes, however, were common to both studies. Laminar shear stress (24 h) caused an increase of MMP1 in both studies. Rb binding protein 2, FGF receptor 3 precursor, CRM1 protein, and BRCA2 decreased in both studies. Thus some of the changes in gene expression are in the same direction, but others are not. By using a twofold change as a criterion, Garcia-Cardena et al. (17) found a shear-induced upregulation of fibronectin precursor, thioredoxin reductase, osteonectin, and protein tyrosine phosphatase. In the present study, however, these genes either did not show a statistically significant change.
In comparing the results of these two studies, we must note several differences in materials and methods. Garcia-Cardena et al. (17) used HUVECs, whereas we used HAECs derived from an adult female donor. The difference in the origin of cells and also the use of growth factors at various concentrations may contribute to differences in expression of genes. As mentioned above, the gene list used as probes were different between the two studies, and even genes with the same name may have different portions of the sequences. Garcia-Cardena et al. (17) used a microarray with a large number of genes to conduct a set of experiments on control, laminar flow, and disturbed flow, whereas we used a much smaller microarray but performed statistical assessment of significance of differences on three sets of experiments on laminar flow and control. These differences in materials and methods, may contribute to the discrepancies in the results obtained in the two studies. With the DNA microarray technology being in its early development stage, there is a need to conduct and report results under different conditions, but each one must be clearly documented for future data mining and bioinformatic analysis.
In the living cardiovascular system, the shear condition may be the normal situation and the absence of shear may be the pathological situation. Thus the straight part of the aorta, which is continuously exposed to laminar shear stress, is generally spared from atherogenesis; in contrast, areas with disturbed flow or low flow are the areas with predilection. It is interesting that Garcia-Cardena et al. (17) found similar gene expression patterns for disturbed flow and static condition, both being different from 24-h laminar flow. Therefore, although this report uses the static value as a control, perhaps one should take the view that the results obtained under shear are the reference and the absence of shear represents a deviation from the norm.
Conclusions.
We investigated the effects of 24-h shear stress on gene expression profiles of cultured HAECs by using the microarray technique. We report that 24-h shear stress significantly modulates EC gene expression related to inflammatory cytokines, cell proliferation, ECM/cytoskeleton remodeling, and signal transduction. The results suggest that relatively long-term shear stress inhibits EC turnover and inflammatory responses, thus keeping ECs in a relatively noninflammatory and nonproliferative state to maintain their normal function as a permeability barrier and nonthrombogenic surface. Our results also provide evidence that 24-h shear stress promotes angiogenesis, EC remodeling, and migration, which would facilitate vascular formation and remodeling. (For detailed information and databases of the current study, please refer to the Supplementary Material1 for this article, published online at the Physiological Genomics web site.)
![]() |
ACKNOWLEDGMENTS |
---|
![]() |
FOOTNOTES |
---|
Address for reprint requests and other correspondence: S. Chien, Dept. of Bioengineering and the Whitaker Institute of Biomedical Engineering, Univ. of California, San Diego, Mail code 0427, 9500 Gilman Dr., La Jolla, CA 92093-0427 (E-mail: schien{at}bioeng. ucsd.edu).
1Supplementary Material to this article is available online at http://physiolgenomics.physiology.org/cgi/content/full/7/1/55/DC1.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|