INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

BY ACTING DIRECTLY AND INDIRECTLY on the surface of endothelium, hemodynamic shear stress generated from blood flow plays a pivotal role in endothelial function (6). To maintain physiological levels of vessel wall shear stress and flow rate, vascular tissues respond to changes in shear stress with acute adjustments in vascular tone and with chronic structural remodeling when variations in shear stress persist (24, 30). Certain characteristics of fluid shear stress are associated with endothelial dysfunction as well as vascular disorders such as atherosclerosis, where lesions occur more readily at the lateral wall of vessel bifurcations where the mean wall shear stress is low but oscillating in direction (14, 17). Because of the complicated flow regimens in vivo and the implication in endothelial biology, much recent interest has focused on the endothelial responses and signaling pathways under different flow conditions.

Fluid shear stress presents two distinct mechanical stimuli to the endothelium: 1) the temporal gradient in shear (rapid change in magnitude) and 2) steady shear, which may have differential effects on endothelial function. This concept has been previously suggested in a number of studies (8, 12, 13, 27). Recently, three potential antiatherogenetic genes (manganese superoxide dismutase, cyclooxygenase-2, and endothelial nitric oxide synthase) have been demonstrated to be continuously upregulated by steady laminar shear but not turbulent shear (43). Our laboratory (1) has further shown that the temporal gradient in shear leads to enhanced and sustained expression of several atherosclerosis-related genes such as platelet-derived growth factor A (PDGF-A) and monocyte chemotactic protein 1 (MCP-1) and their respective transcription activators Egr-1 and nuclear factor-B (NF-B), whereas the presence of steady laminar shear stress reduces these responses. These results suggest distinct roles for the temporal gradient in shear and steady shear in atherogenesis in vivo. Whereas it is becoming apparent that these two distinct fluid shear stimuli contribute differently to vascular physiology and pathology, there is little, if any, knowledge on how endothelial cells (EC) transduce the temporal gradient in shear and steady shear differentially into biological signals, which ultimately lead to discrete endothelial responses.

In this report, we directly compare the effects of the temporal gradient in shear to steady shear on three indexes of mechanotransduction: extracellular signal-regulated kinases 1 and 2 (ERK1/ERK2), c-fos, and connexin43 (Cx43). To separate these two distinct fluid stimuli, three well-defined flow profiles were used: ramp, step, and impulse flow. The smoothly transitioned ramp flow eliminates the rapid large change (temporal gradient) in shear at flow onset, whereas a single 3-s impulse flow isolates this transient flow event. Step flow, which has been applied widely in in vitro flow studies, contains both the sharp increase and the steady flow component. By creating these simplified flow profiles, we can better characterize the responses of endothelial cells to physiological flows by deconstructing the complex flows that exist in vivo into the individual components of the flow in an in vitro system.

The cytoplasmic mitogen-activated protein kinase (MAPK) cascade is a process through which extracellular stimuli can be transmitted from the cell surface to the nucleus (3). c-fos is an immediate-early gene encoding a DNA-binding protein of the AP-1 family, which can function in the transcription activation of many other genes (4). Cx43 is a membrane-spanning gap junction protein complex that allows the direct movement of small molecules and ions between adjacent cells and is involved in cell-to-cell communication (28). MAPK, including ERK1/ERK2 and c-jun NH₂-terminal kinase (JNK), as well as c-fos and Cx43 genes, are all rapidly and transiently activated by fluid shear (5, 22, 31, 44). We report here that the activation of ERK1/ERK2 and the expression of c-fos and Cx43 mRNA are highly sensitive to the temporal gradient in shear (impulse flow and the onset of step flow) but not steady shear (ramp flow and the steady component of step flow). Surprisingly, ERK1/ERK2 was inactivated by ramp flow. By using a MAPK kinase inhibitor, the involvement of MAPK in the induction of c-fos and Cx43 gene expression by the temporal gradient in shear stress is also demonstrated. These results indicate the distinct signaling pathways induced by the temporal gradient in shear and steady shear.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell culture and experimental preparation. Primary human umbilical vein endothelial cell (HUVEC) isolation was performed as previously described (12). Briefly, cells were harvested from human umbilical veins by collagenase (Boehringer Mannheim) treatment and were then seeded onto glass microscope slides. The cells were grown to confluence within 3-4 days in medium 199 (M199) supplemented with 20% fetal bovine serum (FBS, Hyclone), 2 mM L-glutamine, 0.5 U/ml penicillin, and 0.05 mg/ml streptomycin (Sigma). Before shear exposure, the confluent cells were serum starved for 48 h in 0.5% FBS-supplemented M199 to establish quiescence in the monolayers. For the MAPK inhibitor study, cell monolayers were preincubated with PD-98059 (Calbiochem) in DMSO (Sigma) at concentrations of 5-50 µM for 1 h before being subjected to shear. All cell cultures were maintained in a humidified 5% CO₂-95% air incubator at 37°C.

Flow experiments. Confluent cell monolayers on glass slides were subjected to well-defined laminar flow for various times in a parallel-plate flow chamber as previously described (12), where perfusing medium (serum-free M199) was driven by a syringe pump (Pump 44, Harvard Apparatus). The cell-covered slide was placed gently on the flow chamber containing 1.5 ml of perfusing medium, and the flow experiment was started immediately. The computer-controlled syringe pump was programmed to generate the different flow profiles. The flow rate (and hence shear stress) was changed in 1-s microsteps with volume/flow increments determined from the slope of the overall increase in shear stress. Four well-defined laminar flow profiles were generated as described elsewhere (1): 1) step flow (instantaneous shear stress increase from 0 to 16 dyn/cm², followed by steady shear for a sustained period); 2) ramp flow [shear stress smoothly transitioned from 0 to 16 dyn/cm² over 2 (for ERK1/ERK2 activity) or 5 min (for gene expression) and then sustained for a desired period]; 3) impulse flow (a 3-s impulse of 16 dyn/cm²); and 4) reverse impulse flow (a step increase in shear to 16 dyn/cm², sustained for 3 s, followed by a ramped decrease to 0 dyn/cm² over 5 min).

The entire flow device was placed in an air curtain incubator and maintained at 37°C throughout the experiment. The fluid used to shear the cells was the same as the medium used to preincubate the cells before shear exposure and was placed in the incubator for 2-3 h before the flow experiments. After impulse flow exposure, slides were removed from the flow chamber, placed back into petri dishes, and kept in the incubator for the appropriate time-matched period. As sham controls, the glass slides were removed from the petri dish and placed on the flow chamber with perfusing medium, but without flow for 10 min, and then placed back into the petri dish and kept in the incubator for another 20 min. Time-matched stationary (no-flow) controls were performed in petri dishes.

RNA isolation and Northern blotting. The cell monolayers were washed twice with PBS and then lysed. Total RNA was then isolated using a RNAeasy Total RNA Kit (Qiagen). After quantification by measuring absorbance at 260 nm, equal amounts of total RNA were loaded and electrophoresed on 1% formaldehyde-agarose gels, transferred to nylon membranes (Micron Separation), and then fixed by ultraviolet irradiation. The following cDNA fragments were labeled with [³²P]dCTP (ICN Radiochemical) and used as hybridization probes: for c-fos, a 1.0-kb Pst I fragment of pfos-1 (ATCC); for Cx43, a 1.8-kb BamH I/Hind III fragment of HCGJ2 (kind gift of Dr. Glenn I. Fishman); for glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a 0.8-kb Xba I/Pst I fragment of pHcGAP (ATCC). Nylon filters with uniformly transferred RNA were prehybridized for 2-4 h and then hybridized at 42°C in hybridization buffer overnight with the appropriate ³²P-labeled DNA probes. The membranes were then washed and exposed to Kodak X-omat film at 70°C for autoradiography. For some membranes the bound probes were stripped off by washing the membranes in stripping buffer (1 mM Tris · HCl, pH 8; 1 mM EDTA, pH 8; and 0.1× Denhardts) at 75°C for 2 h. The same membranes were subsequently rehybridized with another probe. Densitometry was performed using a model IS-10000 image analyzer (Alpha Innotech).

Protein preparation and Western blotting. The cells were washed twice with PBS, lysed at 4°C with SDS sample buffer [62.5 mmol/l Tris · HCl (pH 6.8), 2% wt/vol SDS, 10% glycerol, 50 mmol/l DTT, and 0.1% wt/vol bromphenol blue]. Cellular lysate was scraped, transferred into a microtube, and sonicated for 10 s, followed by centrifugation (15 min, 4°C, 14,000 rpm). Protein concentration in the supernatant was determined by Bio-Rad colorimetric protein assay method (Bio-Rad Laboratories). Twenty micrograms of total cellular protein were size fractionated in a 10% SDS-polyacrylamide gel and electroblotted to nitrocellulose membranes. The membrane was stained with Ponceau S (Sigma) to confirm equal amounts of protein in each lane and homogeneous transfer. After the membrane was washed with 20 mmol/l Tris · HCl, 0.5 mol/l NaCl, pH 7.4 (TBS) to remove the stain, the filters were blocked with 5% nonfat dry milk in TBS overnight at 4°C, incubated with the following diluted primary antibodies for 1 h: rabbit anti-human phosphospecific MAPK (ERK1 and ERK2) antibody (New England Biolabs); rabbit polyclonal c-fos antibody (Santa Cruz), and goat polyclonal Cx43 antibody (Santa Cruz), and incubated with horseradish peroxidase-conjugated goat anti-rabbit (New England Biolabs) or monkey anti-goat (Santa cruz) antibody for 1 h. The filters were then washed twice with 0.05% Tween 20 in TBS, pH 7.4 (TTBS) for 5 min each and once with TBS. Protein-antibody conjugates were detected by chemiluminescence (super signal CL-HRP, Pierce Chemical).

Statistics. The results are presented as means ± SE compared with controls and among separate experiments. Numerical data consisting of multiple comparisons were first analyzed by one-way ANOVA; if ANOVA indicated a significant difference between groups, the difference between each control and experimental group was then evaluated with the Student-Newman-Keuls test. For all of these tests, a value of P < 0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effects of temporal gradient in shear and steady shear on activation of ERK1/ERK2 and mRNA expression of c-fos and Cx43. ERK1/ERK2 has been shown to be activated by fluid shear in a rapid and transient manner, with the peak activity at 10 min followed by sustained downregulation (39, 44). To directly compared the effects of the temporal gradient in shear and steady shear on the activation of ERK1/ERK2 in HUVEC, confluent cells were subjected to ramp, step, or impulse laminar flow at a maximum shear stress of 16 dyn/cm² for 10 min. We found that compared with static control, ramp flow reduced the phosphorylation of ERK1 and ERK2 by 52 ± 18% and 57 ± 20%, respectively. In contrast, step flow increased the phosphorylation of ERK1 and ERK2 by 332 ± 173% and 294 ± 129%, respectively (P < 0.05, when combining the data of ERK1 and ERK2 together), and impulse flow increased the activation of ERK1 and ERK2 by 756 ± 278% (P < 0.01) and 651 ± 278% (P < 0.05) (Fig. 1), respectively. Moreover, impulse flow-induced enhanced activation of ERK1/ERK2 was observed even at 30 min (data not shown).

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Effect of ramp, step, and impulse flow on activation of extracellular signal-regulated protein kinases 1 and 2 (ERK1/ERK2) in human umbilical vein endothelial cells (HUVECs). Cells were serum deprived for 48 h and then subjected to various flow regimes as indicated for 10 min (see text for details). Total cellular proteins were harvested for Western blotting assay phosphorylation of 44- and 42-kDa mitogen-activated protein kinases (MAPK) (ERK1 and ERK2). Relative levels of ERK1/ERK2 activity (as shown on graph) are expressed as a percentage of level in static cells. Data are presented as means ± SE (n = 4 experiments). *P < 0.05 vs. static control. **P < 0.01 vs. static control. #P < 0.05 vs. step flow.

Fluid shear stress has been demonstrated to induce significant increases in c-fos mRNA level in HUVECs, with the maximal response at 30 min (22). To differentiate the effects of the temporal gradient in shear stress versus steady shear stress on endothelial c-fos expression, confluent cells were subjected to ramp, step, or impulse laminar flow at a maximum shear stress of 16 dyn/cm² for 30 min. Compared with static control, ramp flow was unable to induce significant expression of c-fos mRNA (6.1 ± 1.6-fold increase), whereas step flow produced a 27.5 ± 2.6-fold increase in c-fos mRNA level (P < 0.01). Impulse flow resulted in a further 1.8-fold increase in c-fos expression relative to that induced by step flow (P < 0.01) (Fig. 2).

View larger version (35K): [in this window] [in a new window]   Fig. 2.   Effect of ramp, step, and impulse flow on c-fos mRNA expression in HUVECs. Confluent cells were subjected to various flow regimes as indicated for 30 min (see text detail), and total RNA was isolated from cells for Northern blot analysis of c-fos and GAPDH. Bar graph demonstrates relative levels of c-fos mRNA (normalized by GAPDH mRNA relative to static control) under various flow regimes. Data points are presented as means ± SE (n = 4). *P < 0.01 vs. static control. #P < 0.01 vs. step flow.

To account for the possible effect of small differences in the duration of steady flow due to linear ramp portion of the flow profile on the activity of ERK1/ERK2 and the expression of c-fos mRNA, HUVECs were subjected to two different ramp flows with the shear stress smoothly transitioned from 0 to 16 dyn/cm² over 2 (for ERK1/ERK2) and 5 (for c-fos) min and then continuing for 8 and 25 min (ramp 1) or 10 and 30 min (ramp 2) at 16 dyn/cm². No significant differences in ERK activity and c-fos mRNA expression were observed between these two ramps (Figs. 1 and 2).

To account for possible stimulation of cells by the manipulation during assembly of the cell-covered slide on the chamber, sham controls were utilized in addition to static controls. Here, the slide was placed on the flow chamber with perfusing medium but without flow for 10 min and then immediately lysed for analysis of ERK1/ERK2 activity or placed back in the petri dish for another 20 min for evaluation of c-fos mRNA expression. This procedure of manipulating slides did not induce significant ERK1/ERK2 activation and c-fos mRNA expression (Figs. 1 and 2).

The expression of gap junction gene Cx43 has been found to be modulated by mechanical forces, including shear stress (5). To examine the hypothesis that induction of Cx43 is highly sensitive to the temporal gradient in shear, cells were exposed to ramp and step flow for 1.5 h or exposed to impulse flow followed by static incubation for 1.5 and 4 h. Relative to static control, ramp, step, and impulse flow induced 2.8-, 7-, and 8-fold increases of Cx43 mRNA level at 1.5 h, respectively. The level of Cx43 mRNA induced by impulse flow at 4 h was significantly higher than the expression level at 1.5 h (2.5-fold, P < 0.05), indicating sustained and increasing expression (Fig. 3). To confirm that this sustained gene induction by temporal gradient in shear is not unique to Cx43, the expression of c-fos gene was tested again by rehybridizing the same membrane with c-fos cDNA probe. We found that after 1.5 h, c-fos mRNA expression induced by step and ramp flow returned to the basal level, but impulse flow-induced c-fos mRNA level remained high for at least 4 h (Fig. 3).

View larger version (33K): [in this window] [in a new window]   Fig. 3.   Sustained expression of connexin43 (Cx43) and c-fos mRNA induced by impulse flow in HUVECs. Confluent cells were exposed to ramp or step flow for 1.5 h. Cells subjected to reverse impulse and impulse flow were statically incubated for 1.5 or 4 h as indicated. Total RNA was isolated from cells for Northern blot analysis of Cx43, c-fos, and GAPDH. Bar graph demonstrates relative levels of Cx43 and c-fos mRNA under various flow regimes. Data points are presented as means ± SE (n = 3).

Impulse flow can be further decomposed into two stimuli: a step increase and step decrease in shear with negligible net flow. To elucidate the possible effect of a step decrease on gene expression, the responses to reverse impulse flow (see MATERIALS AND METHODS for details) were analyzed. The levels of c-fos and Cx43 mRNA induced by reverse impulse flow were equivalent to those induced by impulse flow at 1.5 h (Fig. 3). Thus the impulse flow appears to be an equivalent stimulus to a single step increase, suggesting that either the stimulus of the step increase saturated the response or the ramp decrease in flow had no effect.

Effect of temporal gradient in shear on c-fos and Cx43 protein expression. To further confirm the role of the temporal gradient in shear in endothelial function, the levels of c-fos and Cx43 protein expression in endothelial cell monolayers exposed to impulse flow were compared with matched no-flow control cultures. There was no baseline c-fos protein expression in static control. Impulse flow dramatically increased c-fos protein expression after 1.5 h. Impulse flow induced a twofold increase in Cx43 protein expression relative to the static level after 4 h (Fig. 4).

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Effect of impulse flow on level of c-fos and Cx43 protein expression. Cells were serum deprived for 48 h and then subjected to impulse flow for 1.5 h for c-fos and 4 h for Cx43. Total cellular proteins were harvested for Western blotting assay of protein expression. Relative levels of c-fos (open bars) and Cx43 (filled bars) protein expression are expressed as a proportion. Data are presented as means ± SE (n = 3).

Involvement of MAPKs in c-fos and Cx43 mRNA expression induced by temporal gradient in shear. MAPKs have been implicated in the activation of c-fos and Cx43 (16, 45). The data presented in this study have demonstrated the similar activation pattern of ERK1/ERK2, c-fos, and Cx43 by the temporal gradient in shear, suggesting that MAPKs are the upstream regulators of impulse flow-induced c-fos and Cx43 mRNA. PD-98059 has been reported to specifically block the activation of MEK (MAP kinase kinase) and thus prevent the activation of MAP kinase in many cell types (10). Treatment with the MEK inhibitor PD-98059 (5, 25, and 50 µM) decreased the expression of c-fos mRNA induced by impulse flow (25, 62.5, and 62.5%, respectively), whereas 25 µM PD-98059 reduced the impulse flow-stimulated Cx43 mRNA expression by 39% (Fig. 5, P < 0.01). Thus these results indicate that the activation of MAPKs is involved in the gene expression of c-fos and Cx43 induced by the temporal gradient in shear.

View larger version (32K): [in this window] [in a new window]   Fig. 5.   Effect of MAPK activity on impulse flow-induced c-fos and Cx43 gene expression in HUVECs. Confluent cells were pretreated with PD-98059 (a specific MAPK kinase inhibitor) at different concentrations 1 h before being subjected to impulse flow. Total RNA was isolated from cells after 0.5 or 1.5 h for Northern blot analysis of c-fos (A) or Cx43 (B), respectively. Relative levels of c-fos and Cx43 mRNA (as demonstrated in C) are expressed as percentages of those induced by impulse flow alone. Data are presented as means ± SE (n = 3).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The regulation of vascular tone, vessel remodeling, and inflammatory response involves the release of the vasodilators [e.g., nitric oxide (NO), prostacyclin I₂] and vasoconstrictors [e.g., endothelin-1 (ET-1)], as well as the expression of genes that affect cell proliferation [e.g., PDGF and transforming growth factor -1 (TGF-1)] and cell adhesion [e.g., intracellular adhesion molecule 1 (ICAM-1) and MCP-1]. With the development of in vitro fluid shear systems, these vasoactive compounds and inflammatory-related genes have been demonstrated to be modulated by fluid shear (see Ref. 6 for review). The induction of some, and perhaps the majority, of these responses exhibits a biphasic pattern, with an initial transient upregulation or production burst followed by sustained downregulation or reduced production. This biphasic response may reflect the nature of the fluid shear to which cultured EC are subjected to: a large rate of change (temporal gradient) in shear at the onset of flow, followed by steady shear. Combined with other observations (1, 36, 38, 43), this biphasic phenomenon suggests that the temporal gradient in shear and steady shear represent distinct biomechanical stimuli that differentially regulate local endothelial function. The present study further confirms that these two distinct fluid stimuli regulate endothelial function via independent biomechanical pathways and directly compares the effect of temporal gradient in shear and steady shear on the activation of MAP kinase, as well as the gene expression of c-fos and Cx43 in EC.

The intracellular signaling pathways through which EC transduce fluid shear to biochemical signals are complex and include G proteins, intracellular Ca²⁺, cAMP, cGMP, inositol trisphosphate, protein kinase C, MAPKs such as ERKs and JNKs, and small GTPases such as Ras (see Refs. 2 and 6 for review). We have previously identified differences stimulated by the temporal gradient in shear and steady shear. The temporal gradient in shear-induced NO production burst is dependent on pertussis toxin-insensitive G proteins as well as intracellular Ca²⁺ and calmodulin, whereas the steady shear-induced sustained low-rate NO production is independent of these factors (13, 27). Both the p44 and p42 MAPKs (ERK1 and ERK2) investigated in this study function in a protein kinase cascade that plays a critical role in the regulation of cell growth and differentiation. Activation of MAPKs occurs through phosphorylation of threonine and tyrosine by upstream MAPK kinases, which are in turn activated by MAPK kinase kinase. ERKs have been found to be activated by shear stress in a rapid and transient manner (39, 44). Here, we found that single 3-s impulse flow with virtually no time-mean component of steady shear dramatically activated ERK1/ERK2 (4.6-fold/5.0-fold); in contrast, ramp flow inhibited the activation of ERK1/ERK2 (50% reduction) (Fig. 1). This finding indicates that the temporal gradient in shear stimulates, whereas steady shear suppresses, the activation of MAPKs and thus provides another difference in signaling cascades induced by theses two fluid shear stimuli.

Downstream to the aforementioned signaling cascades, multiple transcriptional factors such as AP-1, NF-B, Sp1, and Egr-1 are also activated by shear stress (25, 29, 32, 35). A previous study suggests that the temporal gradient in shear but not steady shear stimulated the activation of NF-B and Egr-1 (1). The present study investigated the differential induction of c-fos by these two flow stimuli. c-fos, complexing with c-jun, belongs to the AP-1 family and can mediate transcriptional stimulation through their interaction with the TPA (12-O-tetradecanoylphorbol 13-acetate)-responsive element (TRE). Many genes including MCP-1, PDGF, ICAM-1, ET-1, and TGF-1 contain this cis-acting element, which has been identified as a shear stress response element (42). The phosphorylation of p62^TCF by MAP kinase is involved in the serum induction of c-fos gene (16). Our results reveal that impulse flow stimulates a larger increase of c-fos gene expression than step flow and even more pronounced increase than ramp flow (Fig. 2). Moreover, impulse flow was shown to further increase expression of c-fos gene after 30 min as monitored at 1.5 and 4 h after shear stimulation (Fig. 3). This is in sharp contrast to the effect of step flow at 1.5 h, where gene expression returned to basal levels. Both step and ramp flow contain a uniform time-mean shear stress, but a component present in both step and impulse flow, a large sharp gradient in shear, is absent in ramp flow. Therefore, the contrasting abilities of step and impulse flow and the inability of ramp flow to elicit significant c-fos gene expression suggest that the temporal gradient in shear stimulates c-fos gene expression, whereas steady shear suppresses the expression. Furthermore, a specific inhibitor of MAPK kinase kinase (PD-98059) dose-dependently inhibited impulse flow-induced expression of c-fos mRNA (Fig. 5), indicating the MAPK-c-fos pathway is involved in the temporal gradient in shear-induced signaling.

Cell-to-cell communication mediated through gap junctions is known to be essential to the morphogenesis and developmental remodeling of many tissues such as blood vessels (33, 40) and is thought to play a critical role in regulating vasotone (11, 41), vascular reactivity (18), as well as cell growth and differentiation (46). For example, the propagation of localized vasomotor stimuli to sites far distant from the initial stimulus is mediated by gap junctional communication (GJC) (41). Downregulation of GJC has been noted in many tumor cells (46). Conversely, the upregulation of GJC in the tumor cells has been associated with a decrease in cellular growth rates (21). Mechanical forces including both mechanical strain and fluid shear stress have been found to modulate the expression of the Cx43 gap junction gene (5, 9). We demonstrated here that the expression of Cx43 mRNA in EC is more sensitive to the temporal gradient in shear than to steady shear, because impulse flow induces a more enhanced and sustained expression of Cx43 mRNA compared with ramp and step flow (Fig. 2). It is not clear, however, if this increased Cx43 mRNA expression leads to enhanced GJC. The regulation of Cx43 gene expression by mechanical forces is controlled at the transcriptional level and does not depend on de novo protein synthesis (5). Disruption of GJC followed by the increased serine phosphorylation on Cx43 is known to be mediated by MAP kinase (45). In the present study, the temporal gradient in shear potently activated MAPK (ERK1/ERK2), which may in turn phosphorylate gap junction protein Cx43 and lead to the disruption of GJC. Therefore, we further investigated the level and location of Cx43 protein expression induced by impulse flow and found that the temporal gradient in shear slightly increased the Cx43 protein expression (Fig. 4). Because the antibody used in this study recognizes both phosphorylated and unphosphorylated Cx43, the effect of flow on GJC function could not be determined. In our immunohistochemical study, we observed punctate staining of Cx43 protein localized to cell-to-cell junctions in static control, whereas cells exposed to impulse flow after 4 h stained principally the cytoplasm (data not shown). This is consistent with the observation by DePaolo et al. (9). The translocation of Cx43 immunoreactivity from cell borders into the cytoplasm suggests that the temporal gradient in shear disrupts intercellular communication. Thus upregulation of Cx43 transcription by temporal gradient in shear may simply be a response to increased Cx43 protein turnover.

The present study utilized a single 3-s impulse flow as a simplified model to dissect out the temporal gradient in shear-induced signaling pathways in EC. There is uncertainty as to whether the endothelial responses induced by impulse flow are comparable to the biological counterpart in vivo, where EC experience continuous pulsatile flow. Previous studies found that certain endothelial responses are sensitive to flow pulsatility, including c-fos (22), PDGF-A (23), and NO production (37). These observations are consistent with the results presented here, because pulsatile flow (or square-wave flow) is essentially a series of step increase/decrease in shear stress. Conversely, sinusoidally reversing flow, which is essentially a sustained smoothly transitioned flow, is unable to activate intracellular Ca²⁺ mobilization (20), as well as ET-1 mRNA expression (34), and thus provides additional supportive evidence for the potency of temporal gradients in shear. Nevertheless, further investigation of the endothelial response to pulsatile flow (continuous impulses) will clarify the role of temporal gradients in shear.

The vascular endothelium in a disturbed flow area experiences two kinds of shear stress gradients: 1) temporal gradients in shear, where individual cells are subject to time-varying shear stresses due to the cardiac cycle; and 2) spatial gradients in shear, where there is a distribution of shear stresses between cells due to vessel geometry. The relative role of temporal gradients and spatial gradients in shear in regulating endothelial cell biology is not clear (7). For example, some studies link the atherogenesis and anastomotic intimal hyperplasia to the large temporal gradients in shear stress due to the change of shear direction (36, 38), whereas others relate this to the different distribution of the time-averaged wall shear stress gradients (19, 26). In the current report, we found that the temporal gradient in shear but not steady shear stimulates the gap junction gene Cx43 expression in EC. However, DePaola and co-workers (9), using a disturbed flow shear stress model, suggested that increased Cx43 mRNA expression and translocated Cx43 protein also correlate with the presence of spatial gradients in shear stress. By using histochemical staining in an in vivo rat model, Cx43 protein has been demonstrated to be highly localized to sites of disturbed flow (15), where both temporal and spatial gradients in shear are expected to exist. The flow profiles utilized in this report allow us to isolate the effect of temporal gradient in shear from that of steady shear. It is important to further differentiate the effects of temporal and spatial gradients in shear stress with respect to their roles in endothelial function in general and vascular disease such as atherosclerosis in particular.

In summary, our studies are beginning to define distinct signal transduction pathways for the temporal gradient in shear and steady shear in cultured human vascular endothelial cells. The temporal gradient in shear activates ERK1 and ERK2 and induces gene expression of c-fos and Cx43. Conversely, steady shear inhibits MAPK as well as suppresses c-fos and Cx43 gene expression. Thus it is becoming clearer that the temporal gradient in shear stress is a potent mechanical stimulus that may link disturbed flow to endothelial dysfunction. These findings shed new light on the role and mechanism of hemodynamic shear stress in the physiology and pathology of vascular endothelium.