Molecular Mechanism of Fibronectin Gene Activation by Cyclic Stretch in Vascular Smooth Muscle Cells*

Fibronectin plays an important role in vascular remodeling. A functional interaction between mechanical stimuli and locally produced vasoactive agents is suggested to be crucial for vascular remodeling. We examined the effect of mechanical stretch on fibronectin gene expression in vascular smooth muscle cells and the role of vascular angiotensin II in the regulation of the fibronectin gene in response to stretch. Cyclic stretch induced an increase in vascular fibronectin mRNA levels that was inhibited by actinomycin D and CV11974, an angiotensin II type 1 receptor antagonist; cycloheximide and PD123319, an angiotensin II type 2 receptor antagonist, did not affect the induction. In transfection experiments, fibronectin promoter activity was stimulated by stretch and inhibited by CV11974 but not by PD123319. DNA-protein binding experiments revealed that cyclic stretch enhanced nuclear binding to the AP-1 site, which was partially supershifted by antibody to c-Jun. Site-directed mutation of the AP-1 site significantly decreased the cyclic stretch-mediated activation of fibronectin promoter. Furthermore, antisense c-jun oligonucleotides decreased the stretch-induced stimulation of the fibronectin promoter activity and the mRNA expression. These results suggest that cyclic stretch stimulates vascular fibronectin gene expression mainly via the activation of AP-1 through the angiotensin II type 1 receptor.

Extracellular matrix of the vascular wall plays an important role in pathophysiological changes including vascular remodeling and atherosclerosis in response to hypertension. Fibronectin (FN) 1 is an important component of the extracellular matrix and is implicated functionally in the regulation of several cellular processes, including cell adhesion, migration, transformation, and motility and wound healing. FN has been found to modulate the phenotype of vascular smooth muscle cells (VSMCs) and regulate VSMC growth (1). We previously found that angiotensin II (Ang II) enhances transcription of the FN gene through the Ang II type 1 receptor (AT1 receptor) in VSMCs, at least in part via activation of the rat FN promoter AP-1 binding motif (rFN/AP-1) (2). Although rFN/AP-1 may not be involved in the regulation of FN gene expression in cells other than VSMCs (3,4), this result proposes that rFN/AP-1 is functionally important for the regulation of vascular FN expression in response to various stimuli. Accumulated evidence suggests that hemodynamic forces (including stretch and shear stress) as well as endocrine factors (such as Ang II) are among the most important factors implicated in the physiology and pathophysiology of the vascular wall in vivo. Interactions between extracellular matrix proteins and cellular receptors can transduce signals that lead to changes in shape, motility, and growth of VSMCs. Thus, investigation of mechanical stressmediated regulation of the extracellular matrix and the tissue renin-angiotensin system in VSMCs may be important for the elucidation of a molecular mechanism of vascular remodeling and atherosclerosis. With respect to vascular FN regulation, several in vivo studies have reported that hypertension activates the vascular renin-angiotensin system and induces expression of vascular FN (5,6). Actually, we have recently examined expression of the tissue renin-angiotensin system and FN genes in inbred Dahl Iwai salt-sensitive (DS) and saltresistant (DR) rats (7). The expression of tissue angiotensinogen, AT1 receptor, and FN is regulated differently in DS and DR rats, and salt-mediated hypertension in DS rats stimulates the aortic FN gene, with activation of the tissue renin-angiotensin system in a tissue-specific manner.
Although our previous studies and others suggest that Ang II in vitro and hypertension in vivo increase vascular FN expression, the molecular mechanisms of mechanical stress-mediated regulation of vascular FN is still unclear. Therefore, in the present study, we examined the effects of cyclic stretch on gene expression of extracellular matrix components (FN and collagen) and renin-angiotensin system components (angiotensinogen, angiotensin-converting enzyme (ACE), and AT1 receptor in VSMCs. Cyclic stretch of VSMCs induced mRNA expression of FN, collagen, ACE, and AT1 receptor, and Ang II secreted in an autocrine/paracrine manner is involved in stretch-induced expression of FN. Furthermore, a specific promoter region, rFN/AP-1, may play an important role in the stretch-mediated increase in FN gene transcription.

EXPERIMENTAL PROCEDURES
Materials-RPMI 1640 medium, fetal calf serum, penicillin, and streptomycin were obtained from Life Technologies, Inc. Ang II, actinomycin D (AMD), cycloheximide, and saralasin were purchased from Sigma. AT1 receptor-specific antagonist CV11974 and Ang II type 2 receptor (AT2 receptor)-specific antagonist PD123319 were supplied by Takeda Chemical and Parke Davis, respectively.
Cell Culture and Application of Cyclic Stretch to Cultured Cells-VSMCs were isolated and cultured as described previously (2). For the experiments, cells at approximately 80% confluence in dishes were made quiescent by incubation with serum-free RPMI 1640 medium for 24 h, unless otherwise stated.
Cells grown in 6-well silicone elastomer-bottomed culture plates (Flexcell Corp.) were subjected to cyclic mechanical stretch in a FX-3000 Flexercell Strain Unit with BioFlex Loading Stations (Flexcell Corp.). The strain unit, a modification of the device described by Banes et al. (8) and Gilbert et al. (9), consists of a computer-controlled vacuum unit and a base plate to hold the culture dishes on the Loading Stations. Vacuum is repetitively applied (1 Hz, 0.5 s on-time) at 60 cycles per min to the rubber-bottomed dishes via the base plate, which is placed in a standard CO 2 tissue culture incubator. When vacuum is applied to a flexible culture dish with the Flexercell Strain Unit and Loading Stations, the bottom deforms across the post face, creating uniform radial and circumferential strain to the bottom of culture dish, principally as described previously (10). This process allows cell culture in a uniform mechanically active environment so that all cells stretched over the loading station surface receive the same amount of strain. Previous studies showed that cyclic stretch of VSMCs with the Flexercell Strain Unit induced increases in DNA synthesis and total protein synthesis, thereby suggesting that this system is useful for in vitro analysis of the processes involved in vascular remodeling (11,12).
To determine possible cellular injury due to mechanical stretch, viability of stretched VSMCs was constantly monitored by either the trypan blue dye exclusion test or assay of the mitochondrial reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide to formazan (Chemicon International Inc.) (13). The attached cells showed a viability of 82 Ϯ 7, 84 Ϯ 7, and 87 Ϯ 6% in the 30% stretched, 20% stretched, and nonstretched regimens (12 h), respectively. The viability of stretched VSMCs did not differ from that of nonstretched VSMCs throughout the experiments in this study.
RNA Isolation and Northern Blot Hybridization-Northern blot analysis was performed essentially as described previously (2). Twenty micrograms of each total RNA sample were denatured, electrophoresed, and transferred to a nylon membrane. Hybridization proceeded in the solution containing denatured salmon sperm DNA and the 32 P-labeled probes. Membranes were washed and exposed to x-ray film or to the imaging plate of a FUJIX BIO-Imaging Analyzer BAS2000 (Fuji Photo Film). Probes used in this study were described previously (2).
Radioimmunoassay of Ang II-The conditioned medium was collected from culture dishes of VSMCs either nonstretched or stretched for 12 h at 20% elongation. Samples were applied to a Sep-Pak C18 cartridge (Waters). Peptides were eluted with 80% methanol, dried, and dissolved in Tris acetate buffer (0.1 M, pH 7.4) containing 2.6 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 0.1% bovine serum albumin. Concentration of immunoreactive Ang II was determined by a specific direct radioimmunoassay with an antibody to Ang II as described previously (14).
Plasmid Construction, DNA Transfection, and Chloramphenicol Acetyltransferase (CAT) Assay-Plasmid rFN-CAT was constructed by insertion of a 2044-base pair PstI/PstI rat FN promoter fragment (Ϫ1908 to ϩ136 relative to the major transcription start site) into the BglII/HindIII sites upstream of pUCSV0CAT as described previously (2). rFN-CAT was used as a template to construct mutations in the rFN/AP-1 element (rFN/m[AP-1]), the rFN/Sp1 element (rFN/m[Sp1]), and the rFN/CRE element (rFN/m[CRE]) by oligonucleotide-directed mutagenesis (17). Once the mutations were obtained and confirmed by sequencing, the altered 2044-base pair (Ϫ1908 to ϩ136) fragments were subcloned into the BglII/HindIII sites of pUCSV0CAT to construct rFN/m[AP-1]-CAT, rFN/m[Sp1]-CAT, and rFN/m[CRE]-CAT. Construction of TK-CAT, containing the herpes simplex virus-thymidine kinase promoter upstream of the CAT coding gene, was described previously (2).
The FN promoter-CAT chimeric construct (5 g) and a ␤-galactosidase expression plasmid pCH110 (1 g), which is used to normalize transfection efficiency, were transiently co-transfected into VSMCs by the CaPO 4 precipitation method as described previously (2). Media were replaced with fresh serum-free media 12 h after the transfection, and cells were incubated for 36 h. VSMCs were then pretreated with Ang II receptor antagonists or signaling inhibitors for 30 min, and cells were subjected to 9 h of cyclic stretch (20%). Aliquots of cell extracts containing equal amounts of total protein (60 g) were used for CAT assay. CAT assay was performed as described previously (2), and results were normalized on the basis of protein concentration or ␤-galactosidase activity to correct for differences in transfection efficiency. The conversion ratios of [ 14 C]chloramphenicol were measured with BAS2000.
Preparation of Nuclear Extracts and Electrophoretic Mobility Shift Assay (EMSA)-VSMCs were pretreated with Ang II receptor antagonists or signaling inhibitors for 30 min and subjected to 6 h of cyclic stretch (20%). Nuclear extracts from VSMCs were prepared with a modification of the protocol of Dignam et al. (18) and Swick et al. (19). EMSA was performed essentially as described previously (2). For the supershift experiments, antibodies with epitopes specific to c-Jun, JunB, and conserved regions of the Fos family (Santa Cruz Biotechnology) were used.
DNase I Footprint Analysis-The DNase I footprinting was performed essentially as described previously (20). The FN promoter fragment from Ϫ746 to Ϫ414 relative to the transcriptional start site was end-labeled with [␥-32 P]ATP, followed by digestion with RsaI (Ϫ656) to generate a probe.

Effect of Cyclic Stretch on Extracellular Matrix and Renin-
Angiotensin System mRNA in VSMCs-We first examined whether cyclic stretch (20%) modulated gene expression of extracellular matrix and renin-angiotensin system components in VSMCs. VSMCs were plated on silicone elastomer-bottomed dishes coated with type I collagen, unless otherwise stated. Expression of FN and type I collagen started to increase 1 h after cyclic stretch, and peaked at 12 h (the maximum induction ratios were 4.2-fold for FN, 3.5-fold for type I collagen, and 1.8-fold for ACE); at 48 h these mRNA levels were still elevated compared with control levels (Fig. 1). The levels of ACE mRNA also began to increase 1 h after cyclic stretch, peaked at 6 h, and remained higher than control levels at 48 h. The increase in AT1 receptor mRNA level was first detected 12 h after cyclic stretch, and the mRNA levels were still rising at 48 h. On the other hand, the angiotensinogen mRNA level did not show any significant change in stretched VSMCs, at least by Northern blot analysis. These results indicate that cyclic stretch of VSMCs increases extracellular matrix and renin-angiotensin system mRNA expression, although the responsible mechanism may differ between them.
Next, VSMCs were incubated with various degrees of cyclic stretch. A significant increase in FN mRNA levels was observed at 5% cyclic stretch compared with control, and a %stretchdependent increase in expression of FN mRNA by cyclic stretch was observed between 5 and 20% stretch ( Fig. 2A). Maximum activation of FN mRNA was obtained by 20% cyclic stretch, and the levels of FN mRNA stimulated by 25 and 30% cyclic stretch were almost the same as that stimulated by 20% stretch. Therefore, subsequent experiments were performed by 20% cyclic stretch of VSMCs in serum-free medium.
Previous studies showed that specific extracellular matrix proteins modulated the mechanical strain-induced responses of VSMCs (11,21). To examine a possible effect of extracellular matrix proteins on the induction of FN mRNA by cyclic stretch in VSMCs, we coated the silicone elastomeric bottom of dishes with type I collagen, elastin, FN, or laminin (Type I collagen was used as a control, as stated earlier). However, there were no significant differences in the degree of activation of FN mRNA expression induced by 20% cyclic stretch among these extracellular matrix proteins (Fig. 2B).

Effects of Transcriptional Inhibitor and Protein Synthesis
Inhibitor on Cyclic Stretch-mediated Increase in FN mRNA-To determine whether de novo RNA or protein synthesis was required for a cyclic stretch-induced increase in FN mRNA, VSMCs were pretreated with AMD or cycloheximide for 30 min and stretched (20%) for 6 h. The RNA synthesis inhibitor AMD almost abolished the cyclic stretch-mediated increase in FN mRNA, whereas the induction of FN mRNA by cyclic stretch was not significantly altered by cycloheximide (Fig. 3A). None of these inhibitors alone had any influence on the expression of FN mRNA in control VSMCs. These results indicate that de novo mRNA transcription is required for the induction of FN mRNA expression by cyclic stretch, but that the effects of cyclic stretch do not require de novo protein synthesis to stimulate the expression of FN mRNA.
Putative Role of Ang II Receptor in Cyclic Stretch-mediated Increase in FN mRNA-A previous study using an organ culture model of the aorta showed that the effect of transmural pressure was mediated by stimulation of the vascular reninangiotensin system (21), and the results of the present study indicated that cyclic stretch of VSMCs up-regulated mRNA expression of the local renin-angiotensin system. Thus, to examine a possible involvement of autocrine/paracrine Ang II in the cyclic stretch-mediated increase in FN mRNA and, if so, to determine the type of Ang II receptor(s) involved in mediating the enhanced expression of FN mRNA in response to cyclic stretch, the effects of Ang II receptor antagonists were analyzed. VSMCs were incubated for 30 min with Ang II receptor antagonists and were subjected to 20% cyclic stretch for 6 h. Treatment of VSMCs with the AT1 receptor-specific antagonist CV11974 (10 M) significantly decreased the stimulatory effect of cyclic stretch (Fig. 3B). In contrast, incubation of cells with the AT2 receptor-specific antagonist PD123319 (100 M) did not affect the response to cyclic stretch at all. None of these Ang II receptor antagonists alone had any influence on the expression of FN mRNA in control VSMCs. It is possible that the lack of inhibition by PD123319 was due to an insufficient dose of the inhibitor. However, preincubation of VSMCs with increasing doses of PD123319 (100 and 500 M) did not affect the response to cyclic stretch at all (data not shown). These results indicate that cyclic stretch activates FN mRNA expression, at least partly, through an AT1 receptor-dependent pathway in VSMCs.
Effects of Cyclic Stretch on Ang II Secretion into Culture Medium-A previous study using the Flexercell strain system showed that concentration of immunoreactive Ang II in the conditioned medium was elevated by cyclic stretch of VSMCs (12), and in the present study cyclic stretch of VSMCs caused increases in ACE and AT1 receptor mRNA expression. To determine whether the autocrine/paracrine-secreted Ang II actually contributes to the increase in FN mRNA expression in response to cyclic stretch, incubation medium obtained from VSMCs that were stretched for 12 h was transferred to VSMCs cultured on regular culture dishes. The addition of conditioned media for 12 h significantly enhanced FN mRNA levels in recipient VSMCs (Fig. 4A). Pretreatment of the VSMCs with the AT1/AT2 receptor antagonist saralasin or CV11974 completely blocked the increase induced by the addition of conditioned media, whereas pretreatment with PD123319 or BQ123, an endothelin-1 type A receptor antagonist, did not affect the increase in FN mRNA levels. Furthermore, the culture medium from stretched VSMCs showed a significant increase in immunoreactive Ang II compared with that from nonstretched control VSMCs (Fig. 4B). Thus, Ang II secreted from stretched VSMCs is suggested to be the major molecule in conditioned medium that up-regulates FN mRNA levels.
Cyclic Stretch-mediated Activation of FN Promoter-We examined whether cyclic stretch activates transcription directed by the FN promoter. We transfected a rat FN promoter (Ϫ1908 to ϩ136 of the transcriptional start site)-CAT chimeric gene (rFN-CAT) into VSMCs and stimulated VSMCs with cyclic stretch (20%) for 6 h. Cyclic stretch increased CAT activity of rFN-CAT by 4.1-fold (Fig. 5). We also examined what type of Ang II receptor(s) was involved in mediating the enhanced CAT activity through the FN promoter in response to cyclic stretch. Incubation of VSMCs with 10 M saralasin or CV11974 significantly decreased the stimulatory effect of cyclic stretch, whereas incubation of cells with 100 M PD123319 did not affect the response to cyclic stretch at all. Thus, the AT1 receptor appears to play a role in the cyclic stretch-mediated increase in FN promoter activity in VSMCs.
Cyclic Stretch-mediated Increase in AP-1 Binding Activity-The results of a DNA transfection study showed that cyclic stretch stimulated transcription directed by the FN promoter in VSMCs. Previous studies showed that the 5Ј-flanking region of the rat FN gene contained the motifs for transcription factors E4TF1, AP-1, AP-2, PEA2, Sp1, and CRE (24). Among them, the AP-1-binding motif is known to be one of the targets of Ang II-mediated transcriptional activation (22), and we have recently shown that a rat FN promoter region from Ϫ473 to Ϫ447 of the transcriptional start site (rFN/AP-1), which contains an AP-1 binding motif, is involved in the Ang II-mediated transcriptional activation of the FN gene (2). To examine the effects of cyclic stretch on binding of nuclear factors to this element, we first performed DNase I footprint analysis. VSMCs were subjected to cyclic stretch (20%) for 6 h. The results showed that a sequence from Ϫ458 to Ϫ439 in the rFN/AP-1 element was protected from digestion by DNase I (lanes 2 and 3, denoted by  the hatched box, Fig. 6A). The pattern of DNase I footprinting disclosed increased protection upon 20% cyclic stretch I (Fig. 6A,  lanes 4 and 5). In footprinting competition assay, this protection was inhibited by the addition of nonlabeled rFN/AP-1 element to the reaction mixture (Fig. 6A, lane 6).
We next carried out EMSA using rFN/AP-1 as a probe. VSMCs were subjected to cyclic stretch (20%) for 6 h. Incubation of VSMC-nuclear extracts with the 32 P-labeled rFN/AP-1 produced a single shifted band, and cyclic stretch significantly increased the intensity of this band (Fig. 6B, lanes 1 and 2). In electrophoretic mobility shift competition assay, the shifted band was specifically competed out by the unlabeled rFN/AP-1 element, but not by the rFN/m[AP-1] element which contained substitution mutations interrupting the AP-1 binding motif (Fig. 6B, lanes 3-6).
Incubation of VSMCs with 10 M saralasin or CV11974 significantly decreased the stretch-induced nuclear binding to rFN/AP-1, whereas incubation of cells with 100 M PD123319 did not affect the nuclear binding (Fig. 7, lanes 3-8). Thus, AT1 receptor, but not AT2 receptor, is suggested to be involved in the cyclic stretch-mediated increase in rFN/AP-1 binding activity in VSMCs.
Functional Importance of rFN/AP-1 Element in Stretch-mediated Activation-From the above results, the rFN/AP-1 element seems to exert a major influence on cyclic stretch-mediated transcriptional activity of the FN gene in VSMCs. Thus, to evaluate the functional significance of the rFN/AP-1 element in stretch-mediated FN promoter activity, we first fused the rFN/ AP-1 or rFN/m[AP-1] element in 5Ј to 3Ј orientation upstream of a herpes simplex virus-thymidine kinase (TK) promoter-CAT hybrid gene. As shown in Fig. 8A, rFN/AP-1/TK-CAT elicited stretch-induced expression of the CAT-reporter gene (3.6-fold activation). On the other hand, the rFN/m[AP-1] element (rFN/ m[AP-1]/TK-CAT) did not confer the stretch-mediated activation of CAT expression.
To further establish the functional roles of the rFN/AP-1 element in directing cyclic stretch-induced CAT expression in the native FN promoter context, we assayed effects of a mutation that disrupted binding of nuclear factors to this element. The FN promoter (Ϫ1908 to ϩ136 of the transcriptional start site)-CAT hybrid gene with the mutated rFN/AP-1 element (rFN/m[AP-1]-CAT) showed a significant decrease in stretchmediated promoter activity (Fig. 8B). Because previous studies showed that Sp1 binding motif and CRE site were involved in the regulation of FN expression in several cultured cells other than VSMCs (3,4), we also examined effects of mutations that disrupted binding of nuclear factors to these elements. However, site-directed mutations of these elements (rFN/m[Sp1]-CAT and rFN/m[CRE]-CAT) in the native FN promoter context did not affect the stretch-mediated activation of CAT expression. These functional assays further suggest that the rFN/ AP-1 element, but not the rFN/Sp1 or rFN/CRE element, is important for cyclic stretch-mediated activation of the endogenous FN promoter in VSMCs.
To determine further whether the factors binding to the rFN/AP-1 element were directly related to the components of

FIG. 5. Effects of Ang II receptor antagonists and effects on cyclic stretch-induced FN promoter activity.
VSMCs were transfected with a plasmid rFN-CAT (5 g) containing the FN promoter (Ϫ1908 to ϩ136) linked upstream of the CAT coding gene. Forty-eight hours after transfection, VSMCs were pretreated for 30 min with Ang II receptor antagonists, followed by cyclic stretch (20%) for 6 h, and promoter activity was estimated by CAT assay. CAT activities were measured with an Imaging Analyzer BAS2000 and expressed relative to those achieved with cell extracts from control VSMCs (the mean CAT activity of control VSMCs is expressed as 100%). Bars represent mean Ϯ S.E. of four independent transfection experiments. *, p Ͻ .05 versus stretch (ϩ) without antagonist.
AP-1 family of transcription factors, a supershift assay was performed. An antibody recognizing conserved region epitopes of the Fos family and antibodies recognizing specific region epitopes of c-Jun or JunB were used. The protein-DNA complex formed by the rFN/AP-1 element probe was not supershifted by the antibodies to either Fos or JunB, whereas the shifted complex was partially supershifted in the presence of the antibody to c-Jun antibody (Fig. 9A). These results indicate that the c-Jun protein is one of the components of the protein-DNA complex, which is formed by the binding of nuclear factors to rFN/AP-1 in response to cyclic stretch of VSMCs.
To examine whether c-Jun was actually involved in the cyclic stretch-mediated transcriptional regulation of FN gene, stretched VSMCs were exposed to antisense oligodeoxynucleotides complementary to the c-jun mRNA translation initiation sites, thereby inhibiting c-Jun protein synthesis. This approach has previously been shown to be successful in inhibiting c-Jun protein synthesis in cultured cells (15). VSMCs were transfected by rFN-CAT and subjected to cyclic stretch (20%) for 9 h in the presence of sense or antisense c-jun oligodeoxynucleotides. The addition of 30 M antisense c-fos oligodeoxynucleotides or sense c-jun oligodeoxynucleotides, which were used as controls, did not affect the stretch-induced increase in CAT expression, whereas the presence of 30 M antisense c-jun oligodeoxynucleotides significantly inhibited the induction of CAT activity directed by rFN-CAT (Fig. 9B). Furthermore, antisense c-jun oligodeoxynucleotides also inhibited stretchmediated increase in FN mRNA expression, while antisense c-fos oligodeoxynucleotides or sense c-jun oligodeoxynucleotides did not affect the induction (Fig. 9C). These antisense experiments indicate that c-Jun is functionally involved in cyclic stretch-induced activation of the FN gene in VSMCs. DISCUSSION Previous results suggest that elevation of blood pressure and activation of the vascular renin-angiotensin system may be synergistically perceived by vascular cells as a signal that is transduced to increase the expression of aortic FN (21). Ang II stimulates VSMC growth and enhances the production of FN and collagen in VSMCs, mesangial cells, and renal fibroblasts (23)(24)(25). In addition, a recent study examined the effects of AT1 receptor blockade on the gene expression of immediate-early response genes, including c-jun and c-fos, and FN after endothelial denudation of the carotid artery by balloon catheter in Harlan Sprague-Dawley rats (26). The results showed that blockade of the AT1 receptor inhibited the induction of AP-1 and FN in an injured rat artery.
In this study, we showed that cyclic stretch of VSMCs enhanced the mRNA expression of ACE and AT1 receptor as well as FN and demonstrated that an AT1 receptor antagonist but not an AT2 receptor antagonist inhibited the cyclic stretchinduced expression of the FN gene. Furthermore, we have found that an AP-1-like element of the FN promoter plays a role in the stretch-mediated activation of the vascular FN gene. We have recently shown that Ang II activates transcription of the FN gene through the AT1 receptor by activation of AP-1 in VSMCs (2). Accumulated evidence suggests that mechanical stretch of cardiovascular cells induces the secretion of Ang II and evokes these hypertrophic responses in an autocrine/paracrine manner (12,27).
With respect to the regulation of FN gene expression in VSMCs, although similar mechanisms may be responsible for activation of vascular FN both in cyclic stretching and in Ang II stimulation via the AP-1 element of the FN promoter, there are several differences between them. First, ACE and AT1 receptor mRNA levels are increased by cyclic stretch, whereas AT1 receptor mRNA expression was decreased and ACE mRNA did not change significantly upon Ang II stimulation. Second, blockade of the AT1 receptor completely inhibits Ang II-stimulated as well as conditioned media-induced increase in FN gene expression, whereas AT1 receptor antagonism can partially decrease cyclic stretch-induced activation of the FN gene. The concentration of Ang II secreted into the medium from VSMCs exposed to cyclic stretch was about 100 nM (Fig. 4B). Since 100 nM CV11974 was sufficient for the complete inhibition of the effects of 100 nM Ang II both in isolated rabbit aorta (28) and in cardiac myocytes (29,30), 100 nM saralasin and 100 nM CV11974 abolished the maximum activation of extracellular signal-regulated kinases induced by 100 nM Ang II (31), and 100 nM PD123319 completely suppressed release of arachidonic acid from cardiac myocytes induced by 10 nM Ang II (32). Thus, the concentration of the Ang II receptor antagonists used in this study should be able to completely block the Ang II receptor-mediated effects of Ang II secreted from VSMCs in response to cyclic stretch, and incomplete inhibition of the stretch-mediated increase in mRNA expression and promoter activity of the FN gene suggests two possibilities. One possibility is that factors other than Ang II may also be involved in stretchmediated transcription of the FN gene in VSMCs, as we and others have reported with respect to the stretch-mediated activation of extracellular signal-regulated kinase and other protein kinases in cardiac myocytes (31,33).
The other possibility is that Ang II is produced intracellularly and activates intracellular (e.g. nuclear) Ang II receptors (34). A recent study showed that the intracellular Ang II effect is totally inhibited by the concomitant injection of CV11974 but that extracellular CV11974 does not influence the intracellular effects (35). Therefore, it is possible that cyclic stretch of VSMCs produces Ang II intracellularly at a sufficient concentration to activate nuclear Ang II receptors, which are insensitive to CV11974 in the medium. Actually, cyclic stretch-mediated activation of ACE and AT1 receptor gene expression in this study implies that stretch of VSMCs functionally enhances the action of vascular Ang II produced in an autocrine/paracrine manner. A similar phenomenon was reported in a previous study in which mechanical stretch of cardiomyocytes upregulated the expression of Ang II receptor subtypes to facilitate the AT1 receptor-mediated action of locally produced Ang II (26). Therefore, these results support one possible mechanism for the mechanical stress-mediated induction of vascular FN through the activation of the local renin-angiotensin system in the pathogenesis of vascular hypertrophy. They may also provide a rationale for the use of AT1 receptor antagonists, which seem to be effective in the inhibition of extracellular matrix production, in the management of vascular remodeling associated with hypertension.
Another finding of this study was that cyclic stretch increased transcriptional activity through the FN promoter and enhanced the binding of nuclear factors to rFN/AP-1. The results of the present study showed that mutation of the rFN/ AP-1 site decreased not only nuclear binding to rFN/AP-1 but also transcriptional activity of the FN promoter in response to cyclic stretch. Furthermore, the result of the supershift assay showed that the rFN/AP-1 element-binding factor contained c-Jun, one of the AP-1 family of transcription factors, and the antisense c-jun oligodeoxynucleotides inhibited stretch-induced transcriptional activity directed by rFN/AP-1 and expression of FN mRNA in VSMCs. Therefore, it is suggested that cyclic stretch activates transcription of the FN gene, at least partly, through activation of the rFN/AP-1-binding pro- tein, in particular c-Jun.
Blockade of the AT1 receptor significantly inhibited cyclic stretch-mediated stimulation of rFN/AP-1 cis-element activity, whereas blockade of AT2 receptor did not interfere with the stretch response, thereby indicating that the AT1 receptor mainly mediates the cyclic stretch-induced activation of AP-1 binding activity of the FN promoter in VSMCs. We have previously shown that Ang II treatment activates FN gene transcription through rFN/AP-1 in VSMCs (2). In this context, rFN/AP-1 may be an Ang II-responsive element. In a previous study, the GC-rich region of the platelet-derived growth factor-A gene promoter, which can bind Egr-1 and Sp1, is reported to be a stretch-responsive element in VSMCs (36). EMSA showed that a new Egr-1 binding activity was induced after exposure to cyclic stretch, while an Sp-1 probe disclosed constitutive shifted bands that did not change in response to stretch (36). Thus, the stretching seems to turn on the Egr-1 pathway to activate the platelet-derived growth factor-A gene promoter. With respect to the stretch-induced activation of FN gene, there seems to be some basal FN expression at the levels of nuclear binding, promoter activity, and mRNA expression. Therefore, it is likely that the stretching amplifies the rFN/ AP-1 pathway through the enhanced autocrine/paracrine production of vascular Ang II.
Concerning the regulation of FN gene expression in various cells other than VSMCs, Sp1 and CRE elements play important roles in basal and stimulated transcription (3,4). However, site-directed mutation of the Sp1 or CRE site does not affect stretch-mediated FN gene transcription (Fig. 8). In addition, previous studies showed that the AP-1 sites in the promoter regions were relevant to the pressure overload-mediated upregulation of cardiac AT1 receptor and atrial natriuretic factor genes (37,38). Furthermore, Ang II is a major factor in cardiovascular FN gene activation in vitro and in vivo, which seems to be different from FN gene regulation in other tissues. Therefore, the identification of rFN/AP-1 as a key regulatory and a pathogenetic stimuli-responsive element of the FN gene, which can efficiently direct the activation of the vascular FN gene in response to increases in locally produced Ang II and wall tension in the vascular wall, should be important for gene therapy of cardiovascular diseases at the level of transcription factors in the near future.