ENDOTHELIAL CELL CULTURE

Human umbilical vein endothelial cells (HUVEC) were isolated from fresh umbilical cords. Cells prior to passage 3 were cultured in endothelial cell growth medium supplemented with 10% fetal bovine serum (FBS). Bovine aortic endothelial cells (BAEC) prior to passage 10 were maintained in DMEM media supplemented with 10% FBS. All cell cultures were maintained in a humidified 5% CO2-95% air incubator at 37°C.

SHEAR STRESS EXPERIMENTS

ECs cultured on glass slides were subjected to laminar shear in a rectangular flow channel (3) with shear stresses of 12-15 dynes/cm2, which are in the range found in normal arteries. The pH of the medium was maintained constant by gassing with a mixture of 5% CO2 and 95% air, and the temperature was maintained at 37°C. Static control experiments were performed on cells in flow chambers not exposed to shear stress.

RNA ISOLATION AND NORTHERN BLOT ANALYSIS

Total cellular RNA were isolated from sheared or static cells by using the guanidinium isothiocyanate method. The isolated RNA were electrophoresed on a 1.5% formaldehyde agarose gel and transferred to a nylon membrane for hybridization with the appropriate probes.

PLASMIDS AND DNA TRANSFECTION

DNA plasmids linked with a luciferase reporter were transfected into BAEC at 70% confluence, and the pSV-b-galactosidase plasmid was co-transfected to monitor the transfection efficiency (4). After incubation in 5% CO2-95% air, at 37°C for 6 hr, DMEM media containing 13% FBS were added to the transfected cells for incubation of another 24-48 hr to reach confluence. X-gal staining showed that the transfection efficiency was 10-15%. The cells were then seeded for either the shear experiments or kept as static controls. A lysis buffer containing Triton X-100 was used to release the reporter protein luciferase and b-galactosidase for these assays. The expression of luciferase was normalized to that of b-galactosidase.

IMMUNOPRECIPITATION AND IMMUNOBLOTTING ANALYSIS

In order to study the association of two proteins (say, X and Z), the method of immunoprecipitation followed by immunoblotting was used. The isolated proteins were immunoprecipitated with an antibody against protein X, and the immunoprecipitated complex was electrophoresed on an SDS polyacrylamide gel. The proteins on the gel were transferred to a nitrocellulose membrane and probed with an antibody against protein Z to assess X-Z association. For the assessment of tyrosine phosphorylation of a protein (say X), the immunoprecipitated complex after electrophoresis and transfer was probed with an antibody against phosphotyrosine.

SHEAR STRESS-INDUCTION OF MCP-1 AND TISSUE FACTOR GENES

Shear stress causes time-dependent induction of the expression of a number of immediate early (IE) genes. We found that shear stress at arterial level caused the activation of the MCP-1 gene (5) and the TF gene (6). Northern blot analysis indicates that the activation of the MCP-1 gene in HUVEC was detectable at 0.5 hr, reached a peak in 1.5 hr and returned to the baseline in 4 hr (5). In experiments with sequential deletion in the promoter region of the MCP-1 gene, we found that a divergent TPA-responsive element (TRE) was the cis-element responsible for shear stress induction (4). Site-specific mutagenesis showed that the nucleotide sequence of TGACTCC was critical. The transcription factor for this TRE site is AP-1, which is a Jun-Fos heterodimer or a Jun-Jun homodimer.

Laminar shear stress caused an increase of TF gene expression in HUVEC and BAEC within 1 hr, and that the increase reached a peak at 2 hr and disappeared at 6 hr (6). Functional analysis of the promoter region of the TF gene indicates that a GC-rich region containing three copies each of the Egr-1 and Sp-1 sites was required for the shear-induction of the TF. Mutation of the Sp-1 sites, but not the Egr-1 sites, attenuated the response of the TF promoter to shear stress, suggesting that Sp-1 is critical for shear inducibility of the TF gene.

SHEAR STRESS-INDUCTION OF THE RAS-MEKK-JNK PATHWAY IN GENE ACTIVATION MEDIATED BY TRE (E.G., MCP-1)

The upstream signaling molecules leading to the shear induction of AP-1/TRE-mediated transcriptional activation of genes such as MCP-1 were studied in BAEC (7). Shear stress induced a transient and rapid activation of the small GTPase Ras within 1 min. This was followed by the transient activation of mitogen activated protein kinase (MAPK) pathways, as manifested by the induction of c-jun NH2 terminal kinases (JNK) and extracellular signal-regulated kinases (ERK). Co-transfection of a dominant negative mutant of Ras, i.e., RasN17, attenuated the shear stress activation of JNK and AP-1/TRE-mediated promoter. The relative importance of the JNK and ERK pathways were studied by using negative mutants of these and their upstream molecules. The catalytically inactive mutants of JNK1 and MEKK, i.e., JNK(K-R) and MEKK(K-M) respectively, attenuated the shear-induction of the AP-1/TRE-mediated promoter. In contrast, dominant negative mutants of ERK-1, ERK-2, and Raf-1 had little inhibitory effect. These results indicate the importance of the JNK pathway in the shear-induction the AP-1/TRE-mediated promoter, such as that in the MCP-1 gene.

The activation of JNK was also accompanied by an increased c-Jun transcriptional activity, which was attenuated by a negative mutant of Son of sevenless (Sos) which is a Ras-activating guanine nucleotide exchange factor. Thus, shear stress activates primarily the Ras-MEKK-JNK pathway in inducing the AP-1/TRE-mediated gene expression in ECs. Upstream to Sos, the adaptor molecules Shc and growth factor receptor binding protein 2 (Grb2) play significant roles in mediating the shear activation since negative mutants of these molecules attenuate the shear inducibility of the Ras-MEKK-JNK pathway.

SHEAR STRESS-INDUCTION OF THE IKK-IKB PATHWAYS IN GENE ACTIVATION MEDIATED BY NFKB (E.G., PDGF-B)

The shear stress induction of several genes, including PDGF-B, has been shown to be mediated through the transcription factor NFkB acting on the SSRE. We determined the role of the recently identified IkB kinases (IKKs) (8) in the shear-activation of NFkB in ECs. Shear stress caused a rapid and transient activation of IKKs, which was followed by IkB degradation and NFkB translocation into the nucleus. Transfection of plasmids encoding catalytic inactive mutants of IKKs inhibited the shear-induced NFkB translocation. In addition, constructs encoding antisense IKKs attenuated shear stress induction of a promoter driven by the kB enhancer element. These results indicate that shear stress, by activating IKKs, caused the degradation of IkB, thus allowing the translocation of NFkB into the nucleus to cause the induction of genes such as PDGF-B.

SHEAR STRESS-INDUCTION OF KINASES IN THE EC-MATRIX FOCAL ADHESIONS

We found that the tyrosine kinases such as focal adhesion kinase (FAK) and c-Src in the focal adhesions are involved in the mechanotransduction in ECs in response to shear stress. Shear stress caused FAK to increase its tyrosine phosphorylation, kinase activity, and association with Grb2 in a rapid (< 1 min) and transient manner (9), suggesting that FAK may be linked to the Ras signaling pathway through the Grb2/Sos complex. A dominant negative mutant of FAK, i.e., FAK(F397Y), attenuated the shear stress-induced kinase activity of hemagglutinin (HA) epitope-tagged JNK1. A dominant negative mutant of Sos (DmSos1), in which the guanine nucleotide exchange domain has been deleted, also attenuated the shear-activation of HA-JNK1.

Shear stress also activated c-Src in a similarly rapid and transient manner (10). The shear stress induction of HA-JNK1 was significantly attenuated by the kinase-defective mutants of v-src and c-src, i.e., v-src(K295R) and c-src(K295R), respectively. The combined uses of positive and negative mutants of c-src and Ras indicate that c-src is upstream to Ras in the signal transduction pathway.

ROLE OF INTEGRINS IN THE MECHANOTRANSDUCTION IN ENDOTHELIAL CELLS

To investigate the role of avb3 integrin in the mechanotransduction, we treated EC monolayers with an anti- avb3 mAb (LM609) prior to the shear stress experiment (unpublished studies by Shila Jalali et al.). We found that LM-609 inhibited the shear-induction of JNK and IKK, indicating that the integrin-mediated signaling pathway regulates JNK and IKKs in ECs in response to shear stress.

ROLES OF RETINOBLASTOMA PROTEIN AND PROTEIN DEPHOSPHORYLATION IN EC RESPONSE TO SUSTAINED LAMINAR SHEAR STRESS

The retinoblastoma (Rb) protein can regulate cell cycle through its phosphorylation and dephosphorylation. When unphosphorylated, Rb binds to the transcription factor E2F, which can then bind to the promoter region of the target gene to suppress DNA synthesis. The phosphorylation of Rb (Rb-P) resulting from the action of cyclin-dependent kinases causes dissociation of Rb-P from E2F and the consequent increase in DNA synthesis. The dephosphorylation of Rb-P back to Rb is mediated by protein phosphatases 1 and 2A. We found that sustained laminar shear stress (> 1-2 hr) caused the progressive dephosphorylation of Rb, thus reducing DNA synthesis and cell turnover (unpublished studies by M.C. Lin et al.). Sustained shear stress also caused increases in the activities of serine-threonine phosphatases and protein phosphatases 1 and 2A.

The results of our experiments, together with those reported by others, indicate that shear stress can activate different cis-elements in different genes. Examples are the SSRE in PDGF-B (11), TRE in MCP-1, and Sp-1 in TF. The shear stress-inducible element for one gene may be present in the promoter in another gene but not responsible for shear inducibility. For example, the SSRE is present in the promoter region of the MCP-1 gene but not critical for its activation, and the TRE is present in the promoter region of the TF gene but not critical for its activation. Therefore, different genes may use different sets of cis-elements in their responses to shear stress, and there is not a single shear-inducible cis-element. For a given gene that possess multiple cis-elements, not all of them are responsible for shear inducibility. However, there is the possibility that the apparently non-essential cis-elements may play a modulating role through their interaction with the critical ones.

Our results show that the shear-induced gene activation is modulated by a complex network of signaling pathways. Thus, although the AP1-TRE mediated shear-activation of the MCP-1 gene involves both the JNK and ERK pathways, the former plays a more important role as revealed by the use of negative mutants.

The various molecules in the signaling pathway exhibit different time courses in their responses to the applied mechanical stimuli, probably reflecting the temporal sequence of their activation. Those associated with the cell membrane, e.g., focal adhesion kinase, c-Src and Ras, respond in a time frame of a minute or less to reach their peaks in less than 5 min. The downstream cytoplasmic kinases are activated with a slower time course to reach their peak activities in 10-30 min. The transcription factors (e.g., AP-1 of NFkB), which are activated through the protein phosphorylation cascades, can then translocate into the nucleus to act on different target cis-elements in different genes. Transcriptional activation of genes such as MCP-1 and TF reaches their peaks in 1-2 hr and this is followed by a decline. The inductions of such mRNA and gene products are not only transient but are actually followed by a down-regulation, with the levels of mRNA and gene products decreased to below the basal level in response to the sustained laminar shear stress. One possible explanation of the transient nature of the response is the delayed activation of protein phosphatase activities which terminate the phosphorylation cascade.

It seems that the time course of the activation of genes is tuned to their functional roles. IE genes such as c-fos and MCP-1 induce proliferative responses and monocyte attraction, respectively, which are needed for short-term response to vascular injury. These genes are down-regulated by sustained laminar shear. In the physiological situations in vivo, the sustained high shear stress in the straight part of the aorta would cause the down-regulation of these genes.

Our studies on Rb indicate that the cell cycle would also be suppressed by sustained high shear stress. At the branch points and curvatures, where the shear stress undergoes temporal and spatial fluctuations, the IE genes may not be down-regulated and become more susceptible to activation by mechanical and chemical stimuli. Such regional predilection for the activation of atherogenesis-related genes is in agreement with the preferential distribution of atherosclerotic lesions in the arterial tree and provides a molecular basis for the focal nature of the disease. In agreement with the findings by Dimmeler et al. (12), our results indicate that the sustained high levels of laminar shear stress in the straight part of the aorta have a protective effect against atherogenic processes; such protection is least in the branch and curved regions of the aorta where flow is unsteady and undergoes directional changes. Our results on the effects of disturbed flow on increased DNA synthesis (13) and enhanced promoter activity of the cyclin-dependent kinase cdc2 (unpublished results by Pin-Pin Hsu et al.) are in further support of this thesis. The beneficial effects of exercise in protecting atherogenesis may be partially related to the enhanced blood flow and the attendant increase in shear stress extending into the branch and curved regions, thus providing the ECs in these regions a more favorable hemodynamic environment in gene regulation.

Our experimental results show that integrins in ECs can serve as mechanosensors in response to shear stress. The mechano-sensitive integrins (e.g., avb3), by acting through the kinases in the focal adhesions (e.g., FAK and c-Src), can activate JNK and IKK and hence the genes mediated by AP-1/TRE or NFkB/kB, respectively. We have recent evidence that receptor tyrosine kinases such as vascular endothelial growth factor receptor can also be activated by shear stress in terms of its phosphorylation and association with the adaptor molecules such as Shc (unpublished studies by Dennis Chen et al.). The sharing of the sensing and signaling pathways by mechanical and chemical activation indicates that hemodynamic forces can superimpose its effects on the actions of chemical stimuli on gene and protein expression in ECs.

ACKNOWLEDGMENTS

This work was supported in part by grants HL19454, HL43026, HL44147 and HL56707 from the National Heart, Lung, and Blood Institute and the Development Award from the Whitaker Foundation. The author would like to acknowledge the valuable collaboration of Drs. Fanny Almus, Indermeet Bhullar, Benjamin Chen, Dennis Chen, H.J. Hsieh, Tony Hunter, Shila Jalali, Michael Karin, Song Li, Y.S. Julie Li, Ming-Chao Kurt Lin, Nigel Mackman, G.C. Perry, Martin Schwartz, John Y.J. Shyy, Mohammad Sotoudeh, and Shunichi Usami and the excellent work of Gerald Norwich, Pin-Pin Hsu, Ying-Li Hu, and Suli Yuan.