Suppression of Angiotensin-Converting Enzyme Expression and Activity by Shear Stress
Shear stress caused by the frictional forces of a fluid moving over a cell monolayer is an important regulator of gene expression. In this study, we investigated the effect of shear stress on angiotensin-converting enzyme (ACE) expression and promoter activity in vitro and on local vascular ACE activity in vivo. ACE activity measured in bovine pulmonary artery endothelial (BPAE) cells was reduced by 49.5% after exposure to a shear stress of 20 dyne/cm2 for 18 hours. Short-term shearing (2 hours) elevated ACE activity in BPAE cells, whereas long-term shearing produced a time-dependent reduction in ACE activity by 23.3%, 33.5%, and 48.9% at 8, 12, and 18 hours, respectively. Northern blot analysis revealed that shear stress (20 dyne/cm sup 2 for 18 hours) significantly reduced ACE mRNA expression by 82%. To determine the mechanism of ACE activity and message reduction, the effect of shear on transcriptionally related events was determined in a rabbit aortic endothelial cell line (W3LUC) stably transfected with 1.3 kb of a rat ACE promoter/luciferase construct. Different shear stress magnitudes (5 to 20 dyne/cm2) caused suppression of luciferase activity by an average of 40.7%. ACE promoter activity was suppressed by 2 hours of shear stress (24.7%) and was further inhibited at time periods >8 hours. In vivo elevations in shear stress were created by placing a stainless steel clip over a 12-mm region of the rat abdominal aorta. Restriction of vessel diameter increased blood flow velocity and caused reduction in vascular ACE activity by 40%. These studies suggest that elevations in the level of shear stress alter endothelial cell function by suppressing ACE gene and protein expression in vitro and in vivo. (Circ Res. 1997;80:312-319.)
The endothelial cells lining the inside of blood vessels act as a physical barrier and play a critical role in the release of physiological and metabolic regulators. Endothelial cells are unique in that their position inside the blood vessel allows them to act as a hemodynamic sensor of blood flow and adjust vessel caliber as required. Blood flow is controlled by short- and long-term responses of the blood vessel endothelium. [1,2] For example, short-term responses include the relaxation of the blood vessel in response to NO and epoxides of arachidonic acid, which are both generated within the endothelium. In vivo long-term changes in hemodynamics result in vascular remodeling that is mediated in part by the endothelial cells. 
Blood flowing through a vessel creates a frictional drag as it moves the fluid and suspended blood particulates across the endothelial surface, creating a shear stress at the endothelial-blood interface. The effect of shear stress on endothelial cells has been characterized by its effect on gross cellular alignment in the direction of the shear stress vector,  cytoskeletal restructuring,  and gene expression. The expression of various genes is known to be either upregulated [6–10] or downregulated [11,12] by the hemodynamic shear stress signal in vitro. In an attempt to elucidate the genetic mechanism of the shear stress response, a 6-bp consensus sequence response element was identified in the promoter of platelet-derived growth factor-B, which binds the transcription factor nuclear factor-kappa B and is selectively responsive to shear stress. This element was termed SSRE. [6,13] Several other genes have been found to contain this consensus sequence, as well as other sequences shown to confer shear sensitivity, [9,10] and may provide a common mechanism by which shear sensitivity of protein expression is regulated.
The renin-angiotensin system is an important regulator of vascular tone and sodium-electrolyte balance, and its product, Ang II, may also be important as a trophic factor for the maintenance of vascular structure. [14,15] Interestingly, the genes for human, rat, and rabbit ACE all contain a number of shear responsive elements in their promoter region. On the basis of this common feature, we reasoned that ACE, which is found in high concentrations in the endothelium, may be responsive to shear stress. In the present study, we hypothesized that shear stress suppresses ACE gene expression and leads to a long-term inhibition of ACE enzyme activity. In addition, the present study characterized the shear dependence of endothelial ACE expression, promoter activity, and enzyme activity of endothelial ACE using a series of in vitro and in vivo experiments. Our findings suggest that long-term exposure to elevated shear stress suppresses ACE promoter and gene expression, leading to a reduced enzyme activity.
Materials and Methods
BPAE cells (American Type Culture Collection, No. 209-CCL) were grown in culture using minimum essential medium (Eagle) in Earle's balanced salt solution, supplemented with 20% fetal bovine serum, 100 U/mL of penicillin, 100 micro g/mL streptomycin, and 250 micro g/mL geneticin. Cells were grown in 100x20-mm Falcon Primeria culture dishes in 7 mL of medium in a 95% humidified air/5% CO2 environment at 37 degrees C. Cells were passaged by exposure to trypsin (Sigma Chemical Co) for 10 minutes, removed from culture dishes, spun at 600g for 5 minutes, resuspended in medium, and passaged at 1:5.
Rabbit aortic endothelial cells transfected with a rat ACE promoter/luciferase construct were cultured in F-12 Coon's medium (Sigma) containing 10% fetal bovine serum and 100 U/mL penicillin, 100 micro g/mL streptomycin, and 250 micro g/mL geneticin. Cells were passaged by exposure to pancreatin for 30 minutes, removed from culture dishes, spun at 600g for 5 minutes, resuspended in medium, and passaged at 1:5. All cells used were between passages 4 and 8.
Cell Transfection With ACE Promoter
Rabbit aortic endothelial cells  maintained on F-12 Coons medium (Sigma) supplemented with 10% fetal bovine serum were stably transfected by the calcium phosphate method with a construct containing 1.3 kb of sequence upstream from the transcription initiation site from the rat ACE gene and fused to the luciferase gene (pGL2 vector, Promega). The half-life of the luciferase gene product is 3 hours in mammalian cells,  making it ideal for studies of stably transfected cells. Cells were cotransfected with a neomycin-resistant plasmid at a ratio of 1:20 (pSV7Neo/ACELUC) and selected with G418 (250 micro g/mL, Life Technologies BRL) over several passages, from which several cell clones were selected on the basis of luciferase production and maintenance of cell appearance. In these experiments, clone W3LUC was used.
All cells were sheared using a modified cone plate viscometer as described by Sdougos et al,  adapted for 100x20-mm cell culture dishes. The cone had a fixed 0.5 degrees angle and was rotated at a constant speed to create defined levels of shear stress. All experiments were carried out with cells in a buffered medium (pH 7.4) in a humidified environment with 5% CO2 at 37 degrees C. The pH was monitored periodically and remained unchanged throughout the experiments. In the first experiment, BPAE cells were sheared at 20 dyne/cm2 for 18 hours within the 2- or 4-day period after confluence and processed for measurement of ACE activity (described below). In BPAE and W3LUC cells, the effect of varying levels of shear stress magnitude from 0 to 20 dyne/cm2 for 18 hours was determined. The minimum values of shear stress that produced suppression (10 dyne/cm2 for W3LUC and 20 dyne/cm2 for BPAE cells) were used in all subsequent experiments. The time-course response of these cells to shear stress was also determined for 2-, 4-, 8-, 12-, and 18-hour time periods at 10 and 20 dyne/cm2 in W3LUC and BPAE cells, respectively. Experiments were designed so that all protocols finished exactly 4 days after the cells became confluent. Three groups of cells were used for each time point: (1) experimental cells sheared for the appropriate time period (2, 4, 8, 12, or 18 hours), (2) cells sheared for 18 hours, and (3) static (unsheared) cells containing a nonrotating cone apparatus. All data were expressed as the percentage of the static control endothelial cells. Any experiments in which cells did not display normal morphology or did not align with the shear stress vector were discarded. At the end of the experiment, the cells were processed for measurement of ACE or luciferase activity as described below.
NOS activity was measured by detection of arginine conversion to citrulline.  BPAE cells were homogenized in 0.5 mL of homogenization solution (mmol/L: sucrose 250, monobasic potassium phosphate 10, dibasic potassium phosphate 10, EDTA 1, and phenylmethylsulfonyl fluoride 0.1). Total cell homogenate (50 micro g) was incubated in 40 mmol/L CaCl2, 30 mmol/L NADPH, 25 micro g/mL calmodulin, 200 micro mol/L BH4, 500 micro mol/L flavin adenine dinucleotide, and [2,3,4,5-sup 3 H]L-arginine for 5 minutes. Conversion of labeled arginine to citrulline was determined by assaying 50 micro L of sample using HPLC.
Northern Blot Analysis
Total cellular RNA was isolated from each plate of endothelial cells (100 mm2) using phenol-chloroform extraction (TRIzol reagent, GIBCO BRL Life Technologies). For electrophoresis, 15 micro g total RNA was loaded into each lane and run on a denaturing 1% agarose gel containing 5% formaldehyde for 4 hours at 75 V. The gel was illuminated by UV transmittance, and RNA integrity was verified by the existence of distinct 18 and 28S rRNA bands. The RNA was then transferred to a positively charged nylon membrane (maximum strength Nytran Plus, Shleicher & Schuell) in 20x SSC solution, UV cross-linked, and prehybridized in 20% formamide, 4x SSPE, 5x Denhardt's solution, 5% SDS, 10% dextran sulfate, 200 micro g/mL yeast RNA, 200 micro g/mL denatured salmon sperm DNA, and 25 mmol/L HCl for 6 hours at 62 degrees C. A full-length bovine ACE cDNA (a gift from K. Bernstein, Emory University, Atlanta, Ga) was random-primed (Pharmacia-Ready To Go) and labeled using [alpha-sup 32 P]dCTP (NEN Research). The denatured probe was added to the prehybridization solution for 18 hours at 60 degrees C. After hybridization, the membrane was quickly rinsed twice in 1x SSC/0.1% SDS mixture at room temperature, followed by two 20-minute rinses in 0.1x SSC/0.1% SDS at 55 degrees C. The membrane was exposed using a Phosphor Imager plate (Image-Quant) for 2 hours and scanned. The digital image was quantified and normalized to GAPDH to determine the relative levels of mRNA expression in each group.
In Vivo Aortic Constriction
Intravascular increases in shear stress were created by constriction of the abdominal aorta. Seven-week-old male Sprague-Dawley rats weighing [nearly =]200 to 220 g (SASCO, Madison, Wis) were housed in an American Association for Accreditation of Laboratory Animal Care-approved animal care facility. Surgical procedures were performed under aseptic conditions, and the animals were anesthetized with a ketamine (50 mg/kg) and acepromazine (5 mg/kg IM) mixture. A frontal midline incision was made to expose the abdominal aorta and inferior vena cava. A specially designed cylindrical stainless steel clip (12 mmx1.04 mm ID) was constructed and carefully placed on the aorta to cause a mild constriction. The clip was designed to provide uniform constriction along a 12-mm length of aorta. Built into the clip were two ultrasonic Doppler crystals. The proximal crystal was mounted on an arm that extended 3 mm upstream from the clip entrance and faced upstream at a 45 degrees angle. The distal crystal was mounted in the center of the clip. The mean unclipped diameter of the aorta in the region of the clip was 1.12 +/- 0.04 mm. In control animals, a nonconstricting stainless steel clip (12 mmx2.04 mm ID) was used. After 5 days, the animals were killed by sodium pentobarbital overdose, and the clipped aortic region was removed, rinsed in physiological saline solution, cleared of connective tissue, and sectioned into three 3-mm rings, which were then assayed for ACE activity as described below. All rings were weighed, and ACE activity was normalized to total tissue wet weight.
Determination of Aortic Constriction
Changes in midline velocity in the rat aorta were verified using a clip with two ultrasonic Doppler crystals implanted in the clip wall. One crystal was mounted at a 45 degrees angle facing upstream on a flange arm extending 3 mm upstream from the clip entrance, and a second crystal was located in the middle (6 mm from the entrance) of the clip to minimize entrance effects in the measured velocity profile. The pair of ultrasonic Doppler crystals were calibrated using blood perfused through polyethylene tubing of various IDs (0.58 to 1.5 mm ID). The clip was fixed parallel to the tubing and submerged in water for maximal acoustic coupling. This arrangement allowed for each crystal to detect the same peak Vshift, which was an indicator of midline flow velocity. From the measurement of Vshift, a calibration factor (middle crystal Vshift/end crystal Vshift) for the pair of crystals was determined to be 0.66 +/- 0.03 (n=5 diameters). To verify constriction in vivo, the constricting clip with the embedded Doppler crystals was tested in rats acutely anesthetized (n=5). For each rat, the Vshift from both crystals was determined, ratioed, and compared with the calibrated value (0.66). An increase in the crystal ratio would indicate an increased velocity measured from at the middle of the clip. The peak velocity ratio from these rats was 1.04 +/- 0.04, indicating an increased midline velocity over the clipped region due to constriction. Using this technique, the change in midline flow velocity during aortic constriction could be determined, and changes in intravascular shear stress were calculated.
Measurement of Aortic Ring Viability
Abdominal aortas from a separate group of rats with both constricting and nonconstricting clips were removed, sectioned into three 3-mm rings beginning at the distal end of the clipped region, and placed in a tissue bath containing PSS and suspended between two triangular clips attached to a force transducer. Each ring was preloaded with 2 g of passive tension and allowed to equilibrate in PSS aerated with 95% O2/5% CO2 for 1 hour. The rings were constricted with three concentrations of norepinephrine (10 sup -9 to 10 sup -7 mol/L). In rings preconstricted with the final dose of norepinephrine (10 sup -7 mol/L), a cumulative dose-response curve to acetylcholine was performed over a range from 10 sup -7 to 10 sup -3 mol/L. Any rings not responding to norepinephrine or acetylcholine were not used (n=2 with constricting clip, n=2 with nonconstricting clip, and n=3 with no clip).
Immediately after the shearing, culture medium was removed, and the cells were rinsed twice with borate buffer (0.5 mol/L boric acid and 1.125 mol/L NaCl, pH 8.3). Five hundred microliters of borate buffer plus 0.1% Triton X-100 was placed in each culture dish, and the cells were scraped and drawn several times through a 1-mL syringe with a 25-gauge needle to homogenize the cells. ACE activity was measured using a modified fluorometric technique by incubating 30 micro L of total cell homogenate at 37 degrees C for 60 minutes according to the method described by Santos et al.  Vascular ACE activity was determined by immersion of whole aortic rings, without homogenization, into the assay solution. The amount of His-Leu product formed was measured by adding 50 micro L (20 mg/mL, 74.5 mmol/L) of o-pthaldialdehyde reagent (Sigma). Fluorescence of the samples was read with 365-nm excitation and 485-nm emission using a scanning fluorometer (F-2000 fluorescence spectrophotometer, Hitachi). ACE activities were expressed as mU/mg, where 1 U=1x10 sup -6 mol His-Leu product produced per minute at 37 degrees C. Cell protein levels were determined using the Bradford protein assay (Bio-Rad).
Luciferase Activity Measurement
Luciferase activity was measured using a commercially available kit (Promega). Briefly, cells were rinsed in PBS (mmol/L: NaCl 137, KCl 2.7, Na2 HPO4 4.3, and KH2 PO4 1.4), lysed in 500 micro L of lysis solution, scraped, and homogenized as described for the measurement of ACE activity. Cell homogenate was spun in a microcentrifuge to pellet large debris. The supernatant was removed and assayed (50 micro L) for luciferase activity using a luminometer (AutoLumat LB 953, EG&G Berthold). Supernatant protein concentrations were determined as described above. Luciferase activity was reported as RLU/mg protein.
Differences in ACE activity, ACE mRNA expression, and NOS activity between sheared and control groups were determined using an unpaired t test. Time- or magnitude-dependent changes in ACE activity and luciferase activity by shear stress were analyzed using a one-way ANOVA. Significant differences between groups were determined using Dunnett's post hoc test. Differences in aortic ring reactivity were analyzed using a two-way ANOVA (drug dosexclip type) with repeated measures (drug dose). Significant differences between groups were determined using Bonferroni's post hoc test. All data are expressed as mean +/- SEM. Values of P<.05 were considered significant.
As previously reported, ACE activity increased after the BPAE cells reached confluence. [5,21] Figure 1 shows how ACE activity changes as a function of time past confluence. Control BPAE cells grown in culture had an 85% increase in ACE activity (from 0.17 to 0.31 mU/mg) from 2 to 4 days after confluence. Exposure of these cells at 2 days after confluence to a shear stress of 20 dyne/cm2 for 18 hours had no effect on ACE activity; however, at 4 days after confluence, shear stress inhibited ACE activity by 27%.
The effect of shear stress on upregulating NOS activity was used as a positive control for the cells exposed to shear stress and to disprove the hypothesis that the suppression of ACE activity was due only to a general downregulation of enzyme activity. Figure 2 demonstrates that BPAE cells exposed to the same level of shear stress that caused the ACE activity suppression (20 dyne/cm2 for 18 hours) also caused an 82% increase in NOS activity, as measured by conversion of arginine to citrulline (0.78 +/- 0.10 versus 1.43 +/- 0.13 micro mol/L citrulline [centered dot] mg protein sup -1 [centered dot] min sup -1). BPAE cells exposed to this level of shear stress (20 dyne/cm2 for 18 hours) aligned in the direction of the shear stress vector, as has been previously reported. 
In cells exposed to shear stress, ACE mRNA expression was measured. Figure 3 shows the effect of shear stress (20 dyne/cm2 for 18 hours) on confluent BPAE cells. Shear induced a suppression of ACE mRNA levels compared with control unsheared plates. As shown in Figure 3, normalized ACE mRNA levels in sheared cells were reduced by 86% (1.0 +/- 0.13 to 0.14 +/- 0.06) from control levels.
The response of both BPAE and W3LUC cells to shear stress (0 to 20 dyne/cm sup 2) is summarized in Figure 4. ACE activity in the BPAE cells was not increased by 5 to 15 dyne/cm2 shear stress but significantly decreased by 20 dyne/cm2. In W3LUC cells, luciferase activity, an indicator of ACE promoter activity, was significantly reduced at all levels of shear stress tested (5 to 20 dyne/cm2) compared with unsheared control cells (0 dyne/cm2). The average luciferase reduction across all levels of shear stress was 40.7%.
(Figure 5) summarizes the result of BPAE cell and W3LUC cell lines after exposure to varying periods of shear stress for 2, 4, 8, 12, and 18 hours at a constant shear stress magnitude of 10 and 20 dyne/cm2, respectively. In these time course experiments, all shearing periods were initiated so that the end of shearing was exactly at the same time, 4 days after the cells were determined to be confluent. ACE activity in the BPAE cells had a biphasic response: 2 hours of shear caused a significant increase in ACE activity, whereas 8, 12, and 18 hours of shear caused a dose-dependent suppression. In the W3LUC line, luciferase activity was suppressed after 2 hours of shear stress and at all times thereafter. Luciferase activity was decreased to a greater extent after 8, 12, and 18 hours of shearing compared with short-term shearing of 2 and 4 hours.
Implantation of the constricting clip for 5 days led to a suppression of local ACE within the vessel region constricted by the clip (Figure 6). Using ultrasonic Doppler velocimetry, the effect of abdominal aortic constriction by clipping was verified. An increased Vshift ratio indicated a constriction of the aortic diameter, corresponding to an increase in midline velocity of 58.3 +/- 0.1% and a 2.4-fold increase in shear stress (from [nearly =]12 to 26.8 dyne/cm2) over the unclipped region. In chronic experiments with the Doppler clip implanted for the 5-day duration of the experiment, no change was seen in the Vshift of the middle crystal, indicating that no chronic remodeling of the internal diameter occurred during the period (Figure 7).
The viability of the aortic segment after clipping was tested by determining the reactivity of the aortic ring segment to norepinephrine and acetylcholine. Norepinephrine caused a dose-dependent increase in aortic ring constriction and tension generated and was reversed by acetylcholine, as shown in Figure 8. The response of aortic segments exposed to clipping was not different from segments receiving a nonconstricting clip.
The present study reports the suppression of endothelial ACE activity by fluid shear stress both in vitro and in vivo. Physiological levels of shear stress in vitro (20 dyne/cm2) were shown to cause a reduction in cellular ACE activity and mRNA expression. Further experiments determined that the ACE gene promoter activity was suppressed at all magnitudes of shear stress tested. Both promoter and enzyme activity were suppressed in a time-dependent manner. Studies done in vitro were confirmed by in vivo results demonstrating that local vascular ACE activity was suppressed in aortic regions exposed to elevated shear stress.
The effect of shear stress on enzyme systems has not been extensively studied except for the effect of shear stress on NOS (upregulation).  The effect of shear stress on the expression and activity of enzymes, which act at critical junctions in catalytic pathways, would provide a sensitive means for controlling these pathways. In the present study, we observed a reduction in ACE activity, which is important in the production of the potent vasoconstrictor Ang II. Shear stress regulation of these pathways may provide a stable negative-feedback loop by which fluid shear stress is regulated in vivo  through the inhibition of vasoconstrictor substances (eg, endothelin-1) and upregulation of vasodilators (eg, NOS) by increased shear stress.
The suppression of ACE by elevated wall shear stress could be regulated by at least two potential mechanisms. The activity of cellular membrane-bound ACE measured in the present study is controlled by a balance between ACE synthesis and proteolytic cleavage from the membrane. Suppression of ACE mRNA expression or increased mRNA degradation would lead to a decreased synthesis of enzyme. If this occurs along with an unchanged cleavage rate of its protein from the cell membrane, a resultant reduction in cellular ACE would be measured. Support for this mechanism is given by our in vitro experiments, in which shear stress reduced ACE promoter activity. These studies, in cells stably transfected with 1.3 kb of the ACE promoter, suggest that the reductions in ACE mRNA levels and subsequent reductions in enzymatically active cellular protein measured in BPAE cells were caused by suppression at the gene level.
Membrane-bound ACE could also be regulated by changes in proteolytic cleavage and release into the media. Shear stress may upregulate cleavage and reduce the cellular fraction of ACE, resulting in overall suppression of activity as measured using our technique. Our experiments did not directly test this potential mechanism; however, cleavage of ACE from the cell membrane has been reported to have t1/2 (time for half of the ACE to be removed from the cell membrane) ranging from 8 to 16 hours. [23,24] In our experiments, ACE activity was reduced at 8 hours after the onset of shearing and may be the result of suppressed ACE mRNA levels and a normal cleavage rate.
The BPAE cells used in the present study were found to have an upregulation in ACE activity as the time past confluence increased. BPAE cells after confluence exhibited an 85% increase in ACE activity (0.17 to 0.31 mU/mg). These results are similar to reports that ACE activity increases after contact inhibition occurring after confluence. [21,25] In contrast, subjecting BPAE cells to 20 dyne/cm2 shear stress for 18 hours reduced ACE activity 4 days but not 2 days after confluence (Figure 1). BPAE cells exhibit low-level ACE activity and gene expression until at least 1 day after confluence ; therefore, exposing these cells to shear stress 18 hours before the time point when ACE activity was measured (2 days after confluence) presumably did not result in a suppression of ACE activity due to low basal levels of ACE activity. Measurements of ACE activity at the later time point (4 days after confluence), when ACE activity and expression were elevated, produced a measurable suppression.
To verify our system's ability to generate experimental shear stress and to test the hypothesis that ACE suppression was caused by a generalized downregulation of cellular enzyme systems, we determined the effect of increased shear stress on NOS activity in the BPAE cell line. The same levels of shear stress that produced reductions in ACE activity (20 dyne/cm2 for 18 hours) caused an upregulation of NOS activity (82% above control levels). Previously, NOS mRNA expression and protein levels have been shown to be markedly increased by a shear stress level of 15 dyne/cm sup 2 applied for 24 hours.  This increase in NOS activity verified that these BPAE cells exposed to shear stress are not suppressing ACE activity as a result of an overall cellular reduction in enzymatic proteins. In addition, measurement of LDH activity, an index of gross cellular trauma, was unchanged in cells exposed to shear stress. All BPAE cells exposed to shear stress in the present study also aligned in the direction of the shear stress vector, as has been previously reported. 
We further investigated the effect of shear stress on ACE mRNA expression at the 4-day confluence time point. As presented in Figure 3, normalized ACE mRNA expression was inhibited by 86% (1.0 +/- 0.13 to 0.14 +/- 0.06) using the same level of shear stress (20 dyne/cm2 for 18 hours) as was shown to inhibit ACE activity (Figure 1). This finding demonstrates that ACE expression is inhibited at the genetic level. However, reduction in mRNA levels may be affected by changes in promoter activity, increases in mRNA degradation, or decreased mRNA stability. The suppression of ACE mRNA levels is suggestive of ACE expression being regulated at the promoter level. Previously, it has been reported that numerous genes shown to be responsive to shear stress contain a consensus element (GAGACC) within their promoter, termed SSRE.  The SSRE was first shown to have positive regulatory effects; however, other genes containing the SSRE are downregulated.  It has recently been shown that other regulating elements are capable of conferring shear sensitivity. For example, both TRE and the egr-1 site are positive shear regulatory elements. Downregulation of the endothelin-1 gene, which contains the classic SSRE, has recently been shown to require a portion of the endothelin-1 promoter that lacks the SSRE. Thus, it appears that the regulation of genes by shear stress most likely involves the recruitment and interaction of both positive and inhibitory regulatory regions within the promoter of these genes.
In order to test the possibility of promoter level regulation of the ACE gene, we stably transfected rabbit aortic endothelial cells with [nearly =]1300 bp of the rabbit ACE promoter, containing the SSRE and other significant regulating elements, coupled to the luciferase reporter gene (W3LUC). Of particular interest was the occurrence in both the natural promoter and the 1300-bp construct of multiple TRE-like and egr-1 sites in addition to the SSRE. In the present study, we determined both the responsiveness of ACE promoter activity, as measured by luciferase expression, and ACE activity suppression over shear stress levels from 5 to 20 dyne/cm2 for a duration of 18 hours. Although the stability of the luciferase protein in the cells was not tested in these studies, it has previously been reported to have a half-life of 3 hours in mammalian cells.  Thus, we interpreted reductions of luciferase activity as an indication of reductions in ACE promoter activity. As shown in Figure 4, ACE promoter activity in W3LUC was very sensitive to low-level shear stress (5 dyne/cm2) and was suppressed by an average of 40.7% over all levels of shear stress from 5 to 20 dyne/cm2. These results confirmed that the ACE promoter was directly sensitive to shear stress and that ACE gene expression may be suppressed by a reduction in promoter activity in the BPAE cell line. Furthermore, this promoter construct responded to shear stress at all magnitudes tested.
ACE activity was not suppressed over the same range of shear stress magnitudes as the W3LUC luciferase activities. ACE activity in BPAE cells was suppressed by only the maximal level of shear stress tested (20 dyne/cm2). This difference between suppression of ACE promoter activity may be due to several factors. First, a different species ACE promoter (rat) in the W3LUC cell line was used to determine promoter activity. It is possible that the rat ACE promoter or the rabbit aortic endothelial cell is more sensitive to lower shear stress levels. In addition, this cell line was transfected with only a portion of the promoter and may lack some other upstream regulatory elements that offset this suppression at low shear levels.
From the determination of shear stress magnitude on ACE activity and promoter expression, we chose 20 and 10 dyne/cm2, respectively, to characterize the time course response of ACE activity and promoter suppression. ACE activity in BPAE cells produces a biphasic response with a transient increase after 2 hours of shear stress and a long-term suppression from 8 to 18 hours. Promoter activity was significantly decreased by 2 and 4 hours of shearing, with further suppression caused by long-term shearing (8 to 18 hours). Early changes in promoter activity preceded reductions in ACE activity by several hours and probably contributed to the reduction in ACE protein synthesis.
Although we did not measure ACE mRNA expression at each time point, ACE activity and expression have been shown to be tightly correlated in nonproliferating confluent BPAE cells.  However, the transient increase in ACE activity at 2 hours after the onset of shear stress would seem unlikely to result from the de novo production of ACE protein. Although the exact mechanism of this short-term response remains unknown, shear stress is a potent stimulator of numerous immediate cellular events (eg, K sup + channel activation and intracellular Ca2+ changes) and activation of second messenger pathways (eg, cGMP, inositol tris-phosphate, diacylglycerol, and mitogen-activated protein kinase). Previous reports indicate that platelet-activating factor stimulates membrane-bound ACE activity by increasing cytoplasmic Ca2+.  ACE activity was found to be directly upregulated by this mechanism. Increases in cytosolic Ca2+ may also occur after short-term exposure to shear stress. The increased ACE activity after 2 hours of shear stress could be explained by this mechanism; however, activation of other intracellular pathways also may act to change ACE activity through as-yet-unidentified mechanisms, since some authors have demonstrated an increase in cytosolic Ca2+ after short-term exposure to shear stress, [27,28] whereas others have not. [29,30]
In order to determine if these effects of shear stress on ACE activity are relevant to in vivo situations, we used constriction of the abdominal aorta of the rat to produce elevated levels of shear stress over defined vascular segments. Application of a stainless steel constricting clip over a 12-mm region of rat aorta increased mean blood velocity as measured by ultrasonic Doppler flowmetry by [nearly =]58.3 +/- 0.1%, which caused local shear stress over this region to be increased by 2.4 times. Five days of aortic constriction produced a suppression of vascular ACE activity by 39.8% compared with clipped but nonconstricted controls. Although not measured in these studies, implantation of the aortic clip likely caused an increase in aortic wall rigidity. This increased rigidity could increase the pulsatile pressure within the aortic segment, providing an alternative signal for the suppression of ACE activity. Further testing of aortic regions from rats with both constricting and nonconstricting aortic clips showed no significant differences between their response to constriction by norepinephrine or relaxation by acetylcholine (Figure 8), indicating that the clipping procedure had no damaging effect on the aorta. Interestingly, the constricted aortas did display a smaller degree of constriction to the highest dose of norepinephrine; however, this difference was not significant.
Because ACE is a critical regulator of the renin-angiotensin system, the control of local ACE levels may be important in determination of active tone within the vessel due to the production of the vasoconstrictor Ang II. In addition, Ang II has been shown to have growth-promoting effects. [14,31] In previous studies, it has been shown that suppression of ACE by administration of the ACE inhibitor captopril also causes a reduction in microvascular density in the rat.  Reductions in microvascular density also are seen in experimental and human hypertension. It has been hypothesized that changes in blood flow through different tissue beds may be regulated during hypertension by long-term restructuring of the microvasculature, reducing microvessel density and increasing tissue resistance and returning flow to normal levels.  The in vivo experiments in the present study were performed in a largevessel region and demonstrate the potential effect of shear stress acting on the endothelial cells. The shear stress mechanism of ACE regulation may play an important role in the control of the local renin-angiotensin system for short-term changes in blood flow and long-term changes in vascular structure.
In conclusion, we have used a series of in vitro experiments to demonstrate the suppression of ACE activity by shear stress. Furthermore, this effect seems to be mediated by a reduction in ACE promoter activity and subsequent mRNA levels. We have also confirmed the upregulation of ACE after confluence and shown it to be suppressed by shear stress. When an in vivo aortic constriction model that causes increased intravascular shear stress was used, similar suppression of vascular ACE activity was seen. These results suggest that the ACE component of local vascular renin-angiotensin systems may be regulated by shear stress and alter its effects on vascular control and growth.
Dr Rieder was supported by an American Heart Association Predoctoral Grant. Dr Greene was supported in part by National Heart, Lung, and Blood Institute grant HL-29587. Dr Krieger was funded in part by grants from the Fundacao Amparo Pesquisa Estado Sao Paulo (95/4668-6), Conselho Nacional Desenvolvimento Cientifico e Tecnologico (CNPq-520696-95), Fundacao EJ Zerbini, and Fundacao ProSangue Sao Paulo. The authors would like to gratefully acknowledge Drs Elizabeth Jacobs and Rick Birks for their assistance in the shearing experiments, Dr David Mattson for the measurement of NOS activity, and Harold and David Eick for their technical expertise and construction of the aortic clips and Doppler clip.
Received June 24, 1996; accepted November 27, 1996.
- Selected Abbreviations and Acronyms
- angiotensin-converting enzyme
- Ang II
- angiotensin II
- BPAE cell
- bovine pulmonary artery endothelial cell
- inner diameter
- NO synthase
- relative light unit(s)
- shear stress response element
- tissue plasminogen activator response element
- velocity shift
- © 1997 American Heart Association, Inc.
- Koller A,
- Sun D,
- Kaley G
- Miller VM,
- Vanhoutte PM
- Languille L,
- O'Donnell F
- Resnick N,
- Collins T,
- Atkinson W,
- Bonthron DT,
- Dewey CF,
- Gimbrone MA
- Lin MC,
- Almus F,
- Shyy YJ,
- Chien S
- Shyy JY,
- Lin MC,
- Han J,
- Lu Y,
- Petrime M,
- Chien S
- Malek AM,
- Izumo S
- Ando J,
- Tsuboi H,
- Korenaga R,
- Takada Y,
- Toyama-Sorimachi N,
- Miyasaka M,
- Kamiya A
- Hernandez I,
- Cowley AW,
- Lombard JH,
- Greene AS
- Wang DH,
- Prewitt RL
- Santos RAS,
- Krieger EM,
- Greene LJ
- Shai S-Y,
- Fishel RS,
- Martin BM,
- Berk BC,
- Bernstein KE
- Keller A,
- Kaley G
- Ching S,
- Hayes LW,
- Slakey LL
- Ramaswamy R,
- Sen GC,
- Misono K,
- Sen I
- Helminger G,
- Berk BC,
- Naerem RM
- Shen J,
- Luscinskas FW,
- Connolly A,
- Dewey CF,
- Gimbrone MA
- Mo M,
- Eskin SG,
- Schilling WP
- Munzenmaier DH,
- Greene AS
- Hutchins PM,
- Darnell AE