INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

ENDOTHELIAL CELLS (EC), which form the inner lining of blood vessels, are subjected not only to a continually changing chemical environment but to the forces of the circulation. These forces include 1) shear stress, a frictional force generated by direct contact with blood flow; 2) hydrostatic pressure, systolic and diastolic pressure with each pulse of the cardiac cycle; and 3) cyclic strain, repetitive deformation as the vessel wall rhythmically distends and relaxes with the cardiac cycle (26). These hemodynamic forces recognized by EC have been shown to play an important role in the modulation of vessel function and structure. Shear stress stimulates the secretion of several factors that regulate vessel function, such as nitric oxide (6), prostacyclin (16), endothelin (27), and tissue plasminogen activator (11), and alters cell morphology and orientation (16). Hydrostatic pressure stimulates EC proliferation and morphological change (43).

Cyclic strain can influence EC function and structure as well. Cyclic strain stimulates the production of prostacyclin (40), endothelin (42), nitric oxide (2), and tissue plasminogen activator (19). EC subjected to cyclic strain show significant increase in their proliferative response (30), and elongate and align perpendicular to the vector force (49). Yano et al. (49) demonstrated that focal adhesion kinase (pp125^FAK) and paxillin are tyrosine phosphorylated in EC exposed to cyclic strain, and these events regulate the morphological change and migration induced by cyclic strain. In addition, ₅₁ and ₂₁ integrins play an important role in transducing mechanical stimuli into intracellular signals (50). Protein kinase C (PKC) in EC subjected to cyclic strain is activated and mediates the cyclic strain effect on EC proliferation (33, 34). However, the exact mechanism by which EC sense and react to these hemodynamic forces remains unclear.

The mitogen-activated protein (MAP) kinase family, ubiquitous intracellular serine/threonine kinases including extracellular signal-related kinases (ERK) 1 and 2, plays an important role in the transduction of mitogenic and differentiation signals from the cytoplasm to the cell nucleus (9). Recently, shear stress has been shown to stimulate ERK1/ERK2 in EC (20, 22, 47), and mechanical loading activates ERK2 in rat cardiac myocytes (48). ERK2 has been shown to be involved in the wound-repair process, which requires cell migration and proliferation (12), and an important role of ERK1/ERK2 activation in cell migration has been demonstrated (24).

Because EC exposed to cyclic strain show significantly elevated cell proliferation (30) and have a propensity to elongate and align perpendicular to the vector force (49), the first aim of this study was to determine whether ERK1/ERK2 are phosphorylated or activated by cyclic strain. Second, we examined the functional involvement of ERK1/ERK2 in the signal transduction of strain-induced cell growth or cell orientation by assessing the effect of PD-98059, a specific MAP kinase kinase (MEK) inhibitor (14). Third, we examined the effect of intracellular and extracellular Ca²⁺ mobilization and PKC and tyrosine kinase inhibitors on the activation of ERK1/ERK2 induced by cyclic strain.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Culture of bovine aortic EC. Bovine aortic EC (BAEC) were obtained by gentle scraping of the intimal surface of bovine aorta, obtained from freshly killed calves at a local slaughterhouse (30). Cells were maintained in Dulbecco's modified Eagle's medium high glucose-Ham's F-12 (GIBCO BRL, Gaithersburg, MD) 1:1 mixture, supplemented with 10% heat-inactivated FCS (HyClone Laboratories, Logan, UT), 5 µg/ml deoxycytidine/thymidine (Sigma Chemical, St. Louis, MO), antibiotics (penicillin 100 U/ml, streptomycin 100 µg/ml), and amphotericin B (250 ng/ml) (GIBCO BRL), and grown to confluence at 37°C in a humidified 5% CO₂ incubator. BAEC were identified by their typical cobblestone appearance and indirect immunofluorescence staining for factor VIII antigen. Cells used in this study were between passages 3 and 6. For assessing cell proliferation, MAP kinase immunodetection, and MAP kinase in-gel assay study, EC were synchronized by incubating in media containing 0.1% FCS for 2 days and in serum-free media for the following day.

Experimental protocol. BAEC were seeded (200,000 cells) on flexible-bottomed 25-mm culture plates coated with collagen I (Flex I plate; Flexcell International, McKeesport, PA) and synchronized as above. Cyclic strain was applied utilizing a Flexercell Strain Unit (Flexcell FX-2000, Flexcell International), which consists of a vacuum manifold with recessed ports (3, 17). The flexible bottom membranes were deformed to a known percentage elongation. For the experiments described herein, cyclic strain was applied by deforming the membrane with 150 mmHg of vacuum, which produces an average strain of 10% at a frequency of 60 cycles/min (0.5 s of deformation alternating with 0.5 s of relaxation). For study of strain dependence, the membranes were subjected to 37.5-mmHg vacuum at a frequency of 60 cycles/min, which produces an average strain of 6% (17).

Because cyclic strain elevates intracellular Ca²⁺ levels ([Ca²⁺]_i) in EC (32) and there are calcium-dependent ERK1/ERK2 activation pathways in some cell types (7, 15), we studied the effect of Ca²⁺ on cyclic strain-induced ERK1/ERK2 activation. We employed 10 µM gadolinium chloride (GdCl₃; Aldrich, Milwaukee, WI) and 25 mM 2,5-di-tert-butyl-1,4-benzohydroquinone (BHQ; Calbiochem, San Diego, CA) to inhibit cyclic strain-activated Ca²⁺ influx and mobilization, respectively (29, 31).

Because previous studies in our laboratory have documented the activation of PKC and tyrosine kinase(s) by cyclic strain (33, 49), we investigated the involvement of PKC and tyrosine kinase in cyclic strain-induced ERK1/ERK2 activation with calphostin C (Calbiochem), a specific PKC inhibitor (25), and genistein (GIBCO BRL), a tyrosine kinase inhibitor (1).

EC growth curve and cell orientation. Proliferation of EC was assessed by determination of cell counts. EC were seeded onto the six-well Flex I plates at an initial density of 50,000 cells/well. After an overnight attachment period, EC were synchronized and then subjected to cyclic strain or maintained in a static environment in the same incubator in media containing 10% FCS or without FCS and incubated with 0.04% DMSO or 20 or 40 µM PD-98059 (Calbiochem). The media were changed every other day with fresh PD-98059 or DMSO. At each time point, EC were trypsinized and the cell number of an aliquot was counted by Coulter Counter (model ZM; Coulter Instruments, Hialeah, FL). For assessment of cell orientation, EC were fixed in 3.7% formaldehyde for 10 min and stained with 1% crystal violet (Sigma Chemical) for 5 min and observed under phase-contrast microscopy (Olympus IMT-2; Olympus Optical, Tokyo, Japan).

ERK1/ERK2 phosphorylation. After cyclic strain exposure, cells were washed in ice-cold Hanks' balanced salt solution (Sigma Chemical) twice and scraped in lysis buffer containing 50 mM HEPES, pH 7.4, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl₂, 1 mM EGTA, 100 mM sodium fluoride, 10 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 10 µg/ml leupeptin, 10 µg/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride (PMSF). Cell extracts were sonicated and centrifuged for 15,000 g for 10 min, and supernatants were collected. Protein content was determined by the Bradford technique (5), using the Bio-Rad protein assay system (Bio-Rad, Hercules, CA). Laemmli sample buffer was added to equal protein amounts of each sample and boiled for 5 min. Samples were resolved on 10% SDS-PAGE by the method of Laemmli (28) and transferred to nitrocellulose membrane (Amersham, Arlington Heights, IL) (46). To ensure quantitative transfer of proteins, the gels were routinely stained with Coomassie blue. The membrane was probed with primary antibodies (anti-ERK1/ERK2 antibody, Zymed, South San Francisco, CA; anti-active MAP kinase antibody, specific to phosphorylated forms of ERK1/ERK2, Promega, Madison, WI) and horseradish peroxidase-conjugated secondary antibody (anti-mouse and anti-rabbit, Amersham), with washing as suggested by the manufacturer. Immunodetection was carried out by chemiluminescence (Amersham) and quantitated with a phosphoimager densitometer (Molecular Dynamics, Sunnyvale, CA).

ERK1/ERK2 activation. ERK1/ERK2 activation was detected by in-gel kinase assay according to the method of Kameshita and Fujisawa (23) with some modifications. Briefly, after lysate protein was resolved on 10% SDS-PAGE containing 0.5 mg/ml myelin basic protein (MBP), SDS was removed by incubating the gel with 20% 2-propanol in 50 mM Tris · HCl, pH 8.0, and then with buffer A (50 mM Tris · HCl pH 8.0, 5 mM -mercaptoethanol). The protein was denatured by incubating with 6 M guanidine-HCl, and then renatured with buffer A containing 0.04% Tween 40. After renaturation, the gels were preincubated in buffer B (40 mM HEPES pH 7.5, 0.1 mM EGTA, 20 mM MgCl₂, 2.0 mM dithiothreitol). The kinase assay was accomplished by incubating the gel for 1 h at 22°C in buffer B containing 25 µM ATP and 25 µCi [-³²P]ATP. After incubation, the gel was washed, dried, and subjected to an autoradiography. Simultaneously, in-gel kinase assay with 10% SDS-PAGE without MBP was performed in the same manner to detect autophosphorylation.

Immunoprecipitation of ERK1/ERK2. Immunoprecipitation of ERK1/ERK2 were carried out using protein A Sepharose-conjugated ERK1/ERK2 antibody (Zymed). Briefly, cell lysates (500 µg protein) were incubated in 1 ml immunoprecipitation buffer (1% Triton X-100, 150 mM NaCl, 10 mM Tris pH 7.4, 1 mM EDTA, 1 mM EGTA, 0.2 mM sodium vanadate, 0.2 mM PMSF, 0.5% Nonidet P-40) containing protein A Sepharose-conjugated ERK1/ERK2 antibody at 4°C for 2 h. The samples were centrifuged, and the pellets were washed three times with immunoprecipitation buffer. After the last wash, 50 µl of SDS sample buffer was added to the pellet beads and boiled for 5 min. The immunoprecipitates were then resolved on a MBP-containing SDS gel followed by in-gel kinase assay as described above.

Statistical analysis. Data shown represent means ± SE. Statistical significance for densitometric and proliferation studies was determined by one-way ANOVA with a multiple-comparison method. P < 0.05 was considered significant.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Cyclic strain stimulates ERK1/ERK2 phosphorylation and activation. As shown in Fig. 1, A and B, cyclic strain stimulated the phosphorylation of 44-kDa ERK1 and a 42-kDa ERK2 in a time-dependent manner. Phosphorylation of ERK1/ERK2 as detected by an anti-active MAP kinase antibody was significantly increased by 5 min and was maximal at 10 min (ERK1: 150 ± 12%, ERK2: 151 ± 25%, n = 3) and decreased thereafter. The immunoblotting study with ERK1/ERK2 antibody, where phosphorylated ERK1/ERK2 exhibit retarded mobility, only showed clear phosphorylation of ERK2. Maximal phosphorylation was also observed at 10 min (Fig. 1C).

View larger version (17K): [in this window] [in a new window]   View larger version (43K): [in this window] [in a new window]   View larger version (21K): [in this window] [in a new window]   Fig. 1.   Phosphorylation of extracellular signal-related kinases (ERK) 1 and 2 by cyclic strain: time dependence. Cell lysates from synchronized endothelial cells (EC) exposed to 10% average strain (0, 2, 5, 10, and 30 min and 1, 2, and 4 h) or to 10% FCS for 5 or 10 min were used for immunoblotting studies. A: immunoblot with anti-active mitogen-activated protein (MAP) kinase antibody is shown. B: quantification by densitometry for ERK1/ERK2 phosphorylation expressed as means ± SE (n = 3) is shown. Phosphorylation of ERK1/ERK2 stimulated with 10% FBS was 547 ± 31 and 353 ± 81% of control, respectively. C: immunoblot with anti-ERK1/ERK2 antibody. Open arrows indicate unphosphorylated (ERK1 and 2) and filled arrows indicate phosphorylated (pERK1 and 2) isoforms of MAP kinase. Representative result of 3 independent experiments is shown.

To verify whether cyclic strain-induced phosphorylation is directly linked to the activation of ERK1/ERK2, an in-gel kinase assay using MBP as the substrate was performed. As shown in Fig. 2, the cyclic strain-induced time-dependent activation of ERK1/ERK2 was similar to that demonstrated in Fig. 1A. The 42- and 44-kDa proteins detected in the in-gel kinase assay were confirmed as ERK1/ERK2 by in-gel kinase assays with the samples immunoprecipitated with anti-ERK1/ERK2 antibody (data not shown). To determine whether the phenomenon was dependent on the amplitude of strain, cells were exposed to either 6 or 10% average strain for 10 min followed by immunoblotting with anti-active MAP kinase antibody and by in-gel kinase assay. Figure 3 shows that phosphorylation and activation of ERK1/ERK2 were detected at 6% average strain, but the increase with 10% average strain was higher than that with 6%. Under the regimen of 6% strain exposure, maximum phosphorylation of ERK1/ERK2 detected by immunoblotting with anti-active MAP kinase antibody was observed after 30-min exposure to strain (data not shown). However, maximum phosphorylation of ERK1/ERK2 by 6% average strain at 30 min was still lower than that by 10% average strain at 10 min.

View larger version (19K): [in this window] [in a new window]   View larger version (41K): [in this window] [in a new window]   Fig. 2.   Activation of ERK1 and ERK2 by cyclic strain: time dependence. Cell lysates from synchronized EC exposed to 10% average strain (0, 2, 5, 10, and 30 min and 1, 2, and 4 h) or to 10% FCS for 5 min were resolved on SDS-PAGE containing myelin basic protein (MBP). After denaturation and renaturation steps, the gel was incubated with [-32P]ATP. The gel was then washed, dried, and subjected to autoradiography. A: representative result from 3 independent experiments is shown. Activation of ERK1/ERK2 stimulated with 10% FCS was 575 ± 70 and 554 ± 151% of control, respectively. B: quantification by densitometry expressed as means ± SE (n = 3) is shown.

View larger version (29K): [in this window] [in a new window]   View larger version (33K): [in this window] [in a new window]   View larger version (41K): [in this window] [in a new window]   Fig. 3.   Phosphorylation and activation of ERK1/ERK2 by cyclic strain: strain dependence. Synchronized EC were subjected to either 6 or 10% average strain at 60 cycles/min for 10 min or 10% FCS for 5 min. A: immunoblot with anti-active MAP kinase antibody is shown. Shown is a representative result of 3 independent experiments. B: quantification by densitometry for ERK1/ERK2 phosphorylation expressed as means ± SE (n = 3) is shown. C: for in-gel kinase assay, cell lysates were resolved on SDS-PAGE containing MBP. After denaturation and renaturation steps, the gel was incubated with [-32P]ATP. The gel was then washed, dried, and subjected to autoradiography. Representative result from 3 independent experiments is shown. Cont, control.

PD-98059 prevents cyclic strain-induced phosphorylation and activation of ERK1/ERK2 but not cyclic strain-induced increased cell proliferation and cell orientation. To elucidate the role of the ERK1/ERK2 in the signal transduction pathway induced by cyclic strain, we employed PD-98059, a specific MEK inhibitor (14). Compared with the vehicle control, preincubation with 20 or 40 µM PD-98059 for 1 h inhibited the cyclic strain-induced ERK1/ERK2 phosphorylation in a dose-dependent manner (Fig. 4, A and B). Forty micromolar PD-98059 completely blocked cyclic strain-induced phosphorylation of ERK1/ERK2. Activation of ERK1/ERK2 revealed by the in-gel kinase assay also demonstrated the dose-dependent inhibition by PD-98059 (Fig. 4C).

View larger version (25K): [in this window] [in a new window]   View larger version (32K): [in this window] [in a new window]   View larger version (29K): [in this window] [in a new window]   Fig. 4.   PD-98059 prevents cyclic strain-induced phosphorylation and activation of ERK1/ERK2. EC were pretreated with 20 or 40 µM PD-98059 or 0.04% DMSO control for 1 h and subjected to 10% average strain for 10 min. A: for the immunodetection, cell lysates were probed with anti-active MAP kinase antibody. Arrows indicate the phosphorylated (pERK1/ERK2) isoforms of MAP kinase. Representative result from 3 independent experiments is shown. B: quantification by densitometry expressed as means ± SE (n = 3) is shown. C: for in-gel kinase assay, cell lysates were resolved on SDS-PAGE containing MBP. After denaturation and renaturation step, the gel was incubated with [-32P]ATP. The gel was then washed, dried, and subjected to autoradiography. Representative result from 3 independent experiments is shown.

To further investigate the role of ERK1/ERK2 activation in cyclic strain-induced cell proliferation, EC were maintained in the media containing 10% FCS and incubated with PD-98059 with or without cyclic strain. Under our conditions, as shown in Fig. 5, 20 and 40 µM PD-98059 effectively blocked the cell growth induced by 10% FCS. In the DMSO control group, the cell number with cyclic strain was significantly greater than that without cyclic strain after 3 days. Both 20- and 40-µM PD-98059 treatment decreased the cell number compared with the DMSO groups; however, there was still an incremental increase in proliferation induced by strain (Fig. 5). Similar results were obtained in serum-free media, where serum proliferation effect can be ignored (Fig. 5).

View larger version (38K): [in this window] [in a new window]   Fig. 5.   PD-98059 does not prevent the increased cell proliferation induced by cyclic strain (CS). Synchronized EC seeded on Flex I plate were subjected to 10% average strain or stationary condition with PD-98059 (20 and 40 µM) or DMSO control for up to 7 days in media containing 10% FCS or with serum-free condition. Inset: cell numbers of EC at day 5 without 10% FCS. Cell numbers were determined as described in MATERIALS AND METHODS. Basal EC proliferation was diminished by PD-98059, but the increased cell proliferation induced by cyclic strain was not prevented.

EC subjected to 24-h cyclic strain were elongated and aligned their long axes perpendicular to the vector force in the periphery of the well (Fig. 6B). Figure 6A shows the randomly aligned cells, which were not subjected to cyclic strain. Twenty and 40 µM PD-98059 did not block the strain-induced effect on cell elongation and alignment (Fig. 6, D and F). These results shown in Fig. 5 and 6 suggest that ERK1/ERK2 activation is not critically involved in the strain induced increase in cell proliferation and cell orientation.

View larger version (138K): [in this window] [in a new window]   Fig. 6.   PD-98059 does not prevent the cell orientation induced by cyclic strain. Synchronized EC seeded on Flex I plate were subjected to 10% average strain with DMSO control (B) or 20 (D) or 40 µM (F) PD-98059 for 24 h. Cells were fixed and stained with crystal violet and observed under phase-contrast microscopy (×50). Pictures were taken from the periphery of the culture well, and the dark areas in the lower left-hand side correspond to the edge of the flexible membrane. EC exposed to cyclic strain were elongated and aligned perpendicular to the force vector in the presence or absence of PD-98059, compared with the cells without cyclic strain (A, C, and E). Bar, 100 µm.

Cyclic strain-induced activation of ERK1/ERK2 is independent of intracellular Ca²⁺ mobilization and extracellular Ca²⁺ influx. We have previously shown that GdCl₃, a nonselective cationic blocker of activated Ca²⁺ channels, and BHQ, a depleter of inositol trisphosphate (IP₃)-sensitive Ca²⁺ pools, effectively inhibited cyclic strain-induced extracellular Ca²⁺ influx and Ca²⁺ mobilization in aequorin-loaded BAEC, respectively (32). Pretreatment with GdCl₃ or BHQ for 10 min did not inhibit the cyclic strain-induced phosphorylation and activation of ERK1/ERK2 (Fig. 7).

View larger version (26K): [in this window] [in a new window]   View larger version (29K): [in this window] [in a new window]   View larger version (27K): [in this window] [in a new window]   Fig. 7.   EC were pretreated with 10 µM GdCl3, 25 µM 2,5-di-tert-butyl-1,4-benzohydroquinone (BHQ), or methanol control for 10 min and subjected to 10% average strain for 10 min. A: cell lysates were probed with an anti-active MAP kinase antibody. Arrows indicate phosphorylated ERK1/ERK2 (pERK1 and pERK2). BHQ, a depleter of the inositol trisphosphate-sensitive calcium storage, or GdCl3, a nonselective cationic-blocker of activated Ca2+ channels, did not inhibit the cyclic strain-induced phosphorylation of MAP kinase. Representative result of 3 independent experiments is shown. B: quantification by densitometry expressed as means ± SE (n = 3) is shown. C: for in-gel kinase assay, cell lysates were resolved on SDS-PAGE containing MBP. After denaturation and renaturation steps, the gel was incubated with [-32P]ATP. The gel was then washed, dried, and subjected to autoradiography. Representative result from 3 independent experiments is shown.

PKC and tyrosine kinase(s) regulate cyclic strain-induced activation of ERK1/ERK2. Calphostin C (100 nM) activated by light inhibited the strain-induced phosphorylation of ERK1/ERK2 by 49.0 ± 28.7% and 60.7 ± 19.5%, respectively (n = 3). Although genistein (100 µM) suppressed the basal phosphorylation of ERK1/ERK2, as reported previously (22), it also completely inhibited the cyclic strain-induced phosphorylation and activation of ERK1/ERK2 (Fig. 8). We also examined the effect of another tyrosine kinase inhibitor, herbimycin A. Pretreatment with 3 µM herbimycin for 1 h completely inhibited strain-induced phosphorylation of ERK1/ERK2 [ERK1: 102.2 ± 40.1% (before strain) and 85.1 ± 44.7% (strain); ERK2: 79.3 ± 6.1% (before strain) and 72.2 ± 9.3% (strain, % of DMSO control, n = 2)].

View larger version (26K): [in this window] [in a new window]   View larger version (32K): [in this window] [in a new window]   View larger version (34K): [in this window] [in a new window]   Fig. 8.   EC were pretreated with 100 nM calphostin C for 3 h, 100 µM genistein for 1 h, or 1% DMSO control for 3 h and subjected to 10% average strain for 10 min. A: cell lysates were probed with an anti-active MAP kinase antibody. Arrows indicate phosphorylated ERK1/ERK2 (pERK1 and pERK2). Representative result of 3 independent experiments is shown. B: quantification by densitometry expressed as means ± SE (n = 3) is shown. C: for in-gel kinase assay, cell lysates were resolved on SDS-PAGE containing MBP. After denaturation and renaturation steps, the gel was incubated with [-32P]ATP. The gel was then washed, dried, and subjected to autoradiography. Representative result from 3 independent experiments is shown.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

The major finding of this study is that cyclic strain stimulates time- and strain-dependent phosphorylation and activation of ERK1/ERK2 in EC. Maximum activation was observed at 10 min after cyclic strain. The strain-induced activation of ERK1/ERK2 occurs at the same time as the shear stress-induced activation in EC (22, 47) and mechanical stretch-induced activation in cardiac myocytes (48). However, the exact signal transduction of these stimulations leading to ERK1/ERK2 activation may differ.

[Ca²⁺]_i has been shown to play an important role in mediating extracellular stimulation and regulation of cell function as a second messenger in EC (18). IP₃ production, which leads to mobilization of [Ca²⁺]_i, is promoted by cyclic strain (33, 34) as well as by shear stress in EC (4). Elevation of [Ca²⁺]_i is induced by cyclic strain in EC (32) and by shear stress in EC (38). Therefore, it is likely that hemodynamic and mechanical forces act like a chemical agonist and stimulate similar signaling pathways. Tseng et al. (47) proposed a calcium-independent pathway in ERK1/ERK2 activation in BAEC subjected to shear stress. On the other hand, Yamazaki et al. (48) demonstrated that ERK2 activation is partially dependent on the Ca²⁺ signaling pathway in cardiac myocytes subjected to mechanical loading. On the basis of our previous results that 10 µM GdCl₃ effectively reduced the magnitude of the initial [Ca²⁺]_i transient and abrogated the sustained phase by blocking extracellular Ca²⁺ influx and that 25 µM BHQ abolished the initial [Ca²⁺]_i spike induced by repetitive stretch, presumably by blocking the IP₃-sensitive Ca²⁺ mobilization (32), we investigated the effect of GdCl₃ and BHQ on cyclic strain-induced ERK1/ERK2 activation. Treatment with GdCl₃ and BHQ had no effect on strain-induced ERK1/ERK2 activation, suggesting that strain-induced ERK1/ERK2 activation is independent of extracellular Ca²⁺ influx and internal Ca²⁺ mobilization. These results demonstrate the existence of different signaling pathways in the same cell type responding to different hemodynamic forces (shear stress and cyclic strain) and in different cell types (EC and cardiac myocytes) responding to similar forces such as cyclic strain. This kind of cell-specific signaling pathway has also been reported in astrocytes and glioma cells responding to hypo-osmolarity (37). ERK1/ERK2 activation by hypo-osmolarity is calcium dependent in astrocytes but calcium-independent in glioma cells.

In various types of cells, a requirement of PKC activation in the signal transduction pathway to ERK1/ERK2 activation has been reported. The importance of PKC activation has been also shown in EC subjected to shear stress (47) and in cardiac myocytes subjected to mechanical loading (48). Tseng et al. (47) demonstrated that activation of ERK1/ERK2 by shear stress is dependent on PKC activation, and Yamazaki et al. (48) reported that ERK2 activation by mechanical loading is partially dependent on PKC activation. Because our previous results demonstrated that cyclic strain increases IP₃ and diacylglycerol (34), which subsequently activates PKC in EC (33), it is likely that PKC activation may be involved in ERK1/ERK2 activation. In the present study, pretreatment with 100 nM calphostin C inhibited strain-induced ERK1/ERK2 activation by ~50-60%, suggesting the existence of PKC-dependent and -independent pathways to ERK1/ERK2 activation. It is likely that certain isoenzymes of PKC play a role in the EC response to cyclic strain. By immunoblotting or immunohistochemical studies, major PKC isoenzymes that have been identified in BAEC are PKC-, -, -, -, and -. PKC- and - are Ca²⁺dependent and PKC-, -, and - are Ca²⁺ independent. On the basis of our results that strain-induced ERK1/ERK2 activation is independent of Ca²⁺ flux and partly dependent on PKC activation, we speculate that the Ca²⁺-independent PKC isoenzymes such as , , and  are important in strain-induced ERK1/ERK2 activation. Consistent with this hypothesis, Ishida et al. (21) proposed the possible involvement of PKC- in activation of ERK1/ERK2 by shear stress. Likewise, recent studies in our laboratory on mechanical loading of keratinocytes demonstrated translocation of the PKC- and - (but not ) isoforms (45). However, further study is needed to determine the specific role of PKC isoenzymes in EC response to cyclic strain.

Protein tyrosine kinases such as Src and Pyk2 have been shown to mediate the pathway from G-protein-coupled receptors to ERK1/ERK2 activation (13). Jo et al. (22) demonstrated that genistein blocks shear-induced ERK1/ERK2 activation. However, Yamazaki et al. (48) reported that ERK2 activation by mechanical loading in cardiac myocytes is independent of the protein tyrosine kinase pathway. In this study, 100 µM genistein completely blocked the strain-induced ERK1/ERK2 activation, suggesting that strain-induced ERK1/ERK2 activation is dependent on tyrosine kinase pathway. In a previous study, we have shown that the tyrosine phosphorylation of pp125^FAK and paxillin is increased by cyclic strain and regulates the morphological change and migration induced by cyclic strain (49). These events are also tyrosine kinase dependent, because tyrosine kinase inhibitors such as genistein and tyrphostin A25 inhibited tyrosine phosphorylation of pp125^FAK and paxillin. Because genistein is a nonspecific inhibitor of tyrosine kinases, the exact identities of tyrosine kinases influencing the ERK1/ERK2 activation and pp125^FAK and paxillin phosphorylation remain unclear. However, our previous results suggest that ₅₁ and ₂₁ integrins play an important role in transducing cyclic strain stimulation into intracellular signals (50). Because integrins induce ERK1/ERK2 activation (8) and pp125^FAK and paxillin phosphorylation (35), it is likely that ERK1/ERK2 activation and pp125^FAK phosphorylation share a common signal transduction pathway. In fact, focal adhesion kinase overexpression induces ERK2 activation via the ras-dependent pathway (36), and Ishida et al. (20) demonstrated the involvement of pp125^FAK activation in shear-induced ERK1/ERK2 activation.

Recent accumulating findings suggest that integrins may function as mechanotransducers and have an important role in the signal transduction of both mechanical stress and adhesion (39). Takahashi and Berk (44) demonstrated that shear stress enhances the activation of ERK1/ERK2 by cell adhesion to fibronectin, suggesting the importance of integrin reaction with extracellular matrix (ECM) for ERK1/ERK2 activation. To compare the effect of different ECMs, we also examined the time-dependent phosphorylation of ERK1/ERK2 in BAEC grown on fibronectin-coated plates. Phosphorylation of ERK1/ERK2 in cells grown on fibronectin also peaked at 10 min, and there was no significant difference between collagen and fibronectin (data not shown).

To elucidate the physiological role of ERK1/ERK2 activation in EC exposed to cyclic strain, we employed a specific MEK inhibitor, PD-98059, in cell proliferation and morphological studies. Although PD-98059 inhibited both ERK1/ERK2 phosphorylation and activation induced by cyclic strain, increased cell proliferation and cell alignment induced by strain were not inhibited, suggesting that ERK1/ERK2 activation is not critically involved in these responses. Our previous study with a specific PKC inhibitor, calphostin C, demonstrated that strain-induced increased cell proliferation was abolished, suggesting that PKC activation is necessary for strain-induced cell proliferation (33). We also previously reported that pp125^FAK and paxillin phosphorylation regulates the cell orientation and migration induced by strain; however, in the present study ERK1/ERK2 activation is not necessary for strain-induced cell orientation (49). What then is the physiological role of the strain-induced ERK1/ERK2 activation? ERK1/ERK2 are known to phosphorylate transcription factors after translocation into the nucleus, such as Elk-1, p62^TCF (ternary complex factor), c-fos, and c-myc (10). Because we have shown that cyclic strain induces c-fos and c-jun (41), one possibility is that ERK1/ERK2 may have an important role in regulating gene expression through modulating these transcription factors.

In summary, cyclic strain induces ERK1/ERK2 phosphorylation and activation time dependently. This activation is independent of Ca²⁺ flux but partly regulated by PKC activation and completely regulated by tyrosine kinase(s). ERK1/ERK2 activation is required for basal proliferation of BAEC but does not appear to be involved in the strain-induced increase in cell proliferation and cell orientation. Further experiments involving ERK1/ERK2 phosphorylated substrates (i.e., Elk-1, p62^TCF) may provide a role for the significance of strain-induced ERK1/ERK2 activation.