Patterns of gene expression differentially regulated by platelet-derived growth factor and hypertrophic stimuli in vascular smooth muscle cells: markers for phenotypic modulation and response to injury.

In vascular smooth muscle cells (VSMC), platelet-derived growth factor (PDGF) suppresses expression of multiple smooth muscle contractile proteins, useful markers of differentiation. Conversely, hypertrophic agents induce expression of these genes. The goal of this study was to employ genomic approaches to identify classes of genes differentially regulated by PDGF and hypertrophic stimuli. Changes in gene expression were determined using Affymetrix RAE-230 GeneChips in rat aortic VSMC stimulated with PDGF. For comparison with a model hypertrophic stimulus, a microarray was performed with VSMC stably expressing constitutively active Galpha(16), which strongly induces smooth muscle marker expression. We identified 75 genes whose expression was increased by exposure to PDGF and decreased by expression of Galpha(16) and 97 genes whose expression was decreased by PDGF and increased by Galpha(16). These genes included many smooth muscle-specific proteins; several extracellular matrix, cytoskeletal, and chemotaxis-related proteins; cell signaling molecules; and transcription factors. Changes in gene expression for many of these were confirmed by PCR or immunoblotting. The contribution of signaling pathways activated by PDGF to the gene expression profile was examined in VSMC stably expressing gain-of-function H-Ras or myristoylated Akt. Among the genes that were confirmed to be differentially regulated were CCAAT/enhancer-binding protein delta, versican, and nexilin. All of these genes also had altered expression in injured aortas, consistent with a role for PDGF in the response of injured VSMC. These data indicate that genes that are differentially regulated by PDGF and hypertrophic stimuli may represent families of genes and potentially be biomarkers for vascular injury.

Vascular smooth muscle cells (VSMC) 1 are the contractile component of blood vessels, and express a set of smooth muscle (SM)-specific genes, which are characteristic of their contractile, differentiated phenotype (1). In contrast to skeletal and cardiac myocytes, VSMC do not terminally differentiate, and they undergo phenotypic modulation in vivo and in vitro in response to environmental signals (2,3). This process involves changes in gene expression, which convert these cells from a nonproliferative contractile phenotype to a proliferating synthetic one (2)(3)(4)(5). Indeed, a well defined characteristic of vascular occlusive disease, arterial interventions in response to disease, and in vitro subculturing of SMC is the phenotypic modulation of SMC from a normally quiescent, contractile state to one of increased growth, migration, and matrix synthesis. Formation of neointima after vascular injury is largely a consequence of dedifferentiation, proliferation, and migration of medial SMC (6,7). During injury to the blood vessels, PDGF assists this process by triggering phenotypic changes in VSMC, which are presumably mediated through changes in patterns of gene expression (8). In cultured VSMC, PDGF increases cell proliferation (9). In contrast, vasoconstrictors such as vasopressin (AVP) and angiotensin II promote increased contractility and induce hypertrophy (10 -12). Previous studies from our laboratory have demonstrated that these two classes of agents have opposing effects on expression of smooth muscle markers (SM markers), specifically contractile proteins such as smooth muscle ␣-actin and SM22␣ (13)(14)(15), with AVP increasing and PDGF suppressing transcription of these genes.
The effects on SMC gene expression are mediated through distinct postreceptor signaling pathways. PDGF-induced suppression is mediated through activation of Ras and phosphatidylinositol 3-kinase/Akt pathways (14,15), whereas AVP-mediated regulation of SM markers involves activation of heterotrimeric G proteins of the G q family, leading to stimulation of c-Jun N-terminal kinase and p38 mitogen-activated protein kinases (12,16). However, the phenotypic changes mediated by these agents are probably not exclusively mediated through changes in expression of contractile proteins but will instead be a consequence of more global changes in gene expression. Dedifferentiated VSMC have increased production of matrix proteins and metalloproteinases, which are critical for their proliferative and migratory properties (17,18). Additionally, changes in the local environment of the cells or cell-matrix interactions can affect the phenotypic state of VSMC (19,20). With the advent of microarray technology, it is feasible to  assess global changes in gene expression to begin to define the repertoire of proteins regulating SMC phenotype. For example, gene array analysis has been used to identify changes in gene expression in angiotensin II-stimulated SMC in the presence and absence of inhibitors of signaling pathways (21), between the aorta and vena cava (22), in mechanically stimulated SMC (23), and between neointimal and aortic SMC (24). However, a comprehensive analysis of changes in gene expression induced by hypertrophic stimuli compared with PDGF has yet to be investigated. The goal of this study, therefore, was to use microarray analysis to define subsets of genes that are regulated in opposing directions by PDGF compared with hypertrophic stimuli. By using cell lines stably expressing constitutively active forms of either H-Ras or Akt, downstream effectors of PDGF signaling, we were able to dissect out the contributions of these individual pathways to the regulation of specific genes. We report here a novel set of genes, which are involved in modulation of SMC phenotype during balloon injury.

EXPERIMENTAL PROCEDURES
Materials-Trypsin/EDTA and Eagle's minimal essential medium were from Invitrogen. Antibodies to C/EBP␦, GKLF, and alkaline phosphatase-labeled IgGs were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Antibody to heme oxygenase-1 was from Stress-Gen Biotech, antibody to nexilin was from BD Biosciences, antibody to Ezrin was from Upstate Biotechnology, Inc. (Lake Placid, NY), and antibody to versican was from Affinity Bioreagents (Golden, CO).
Affymetrix Microarray Analysis-VSMC stimulated with AVP or PDGF for 6 or 24 h and VSMC stably expressing H-Ras (14), Myr-Akt (15), or G␣ 16 (12) along with their control cells were subjected to total RNA isolation by using RNeasy miniprep (Qiagen). Labeling of the probes for microarray was performed as described previously (27). Briefly, cDNA was synthesized by using the Superscript Choice System (Invitrogen) and purified by phenol/chloroform extraction and ethanol precipitation. In vitro transcription of labeled cRNA was performed from purified cDNA using a Bioarray high yield RNA transcript labeling kit (Enzo Diagnostics) in the presence of biotinylated UTP and CTP. Labeled cRNA was purified and then fragmented using fragmentation buffer (40 mM Tris acetate, pH 8.1, 10 mM KOAc, 30 mM MgOAc). Hybridization of the probes to the Affymetrix RAE-230 rat Genechip was performed according to the manufacturer's recommendations (Affymetrix, Santa Clara, CA) at the University of Colorado Health Sciences Center Microarray Core Facility. DNA chips were scanned (6-m resolution) with an HP Gene Array scanner, and results were analyzed with GeneChip Suite 4.0 analysis software from Affymetrix and Genespring software from Silicon Genetics.
Transient Transfection-For transient transfections measuring promoter activity, plasmid encoding 765 bp of the rat SM-␣-actin promoter ligated into a promoterless luciferase vector (PA3-Luc) (13) was used. VSMC were transiently transfected with Lipofectamine PLUS (Invitrogen) in 35-mm dishes using 1 g of the SM-␣-actin promoter-luciferase construct together with 1 g of a plasmid encoding cytomegalovirus-␤galactosidase vector (Clontech) for normalization of transfection efficiency. The transfections also included 0.5 g of other plasmids (C/ EBP␦, or empty vector). The plasmid encoding C/EBP␦ is a generous gift of Steven L. McKnight from University of Texas South Western Medical Center (Dallas, TX). Cells were incubated for 16 h in Eagle's minimal essential medium with 10% fetal calf serum and then placed in Eagle's minimal essential medium with 0.2% fetal calf serum with or without agonists for 72 h. The transiently transfected VSMC were then washed twice with ice-cold phosphate-buffered saline and harvested in luciferase reporter lysis buffer (Promega, Madison, WI). Cell lysates were centrifuged, and the supernatants were assayed for luciferase and ␤-galactosidase activities as previously described (16).
Immunofluorescense-Adult thoracic aortas from sham-injured or 7-day and 4-week post-balloon catheter injured animals were collected and fixed overnight at 4°C in 10% buffered formalin, embedded in paraffin, and sectioned as described previously (28) and were immuno-stained with specific antibodies per the manufacturers' recommendations. Prior to labeling with primary antibodies for 1 h at room temperature in 5% goat serum in phosphate-buffered saline, tissue sections were deparrafinized and then heated for 20 min at 105°C in a decloaking chamber (Biocare) for antigen retrieval and then blocked with 5% goat serum for 45 min. Following incubations with primary antibodies for C/EBP␦ (1:50), versican (1:10), and nexilin (1:10), antigen-antibody complexes were visualized using rhodamine (Alexa Fluor-568)-coupled or fluorescein isothiocyanate (Alexa Fluor-488)-coupled secondary antibodies (Molecular Probes, Inc., Eugene, OR). Some sections were double-labeled with anti-SM ␣-actin. Coverslips were mounted with VectaShield medium containing DAPI (Vector Laboratories), and cells were visualized using a Nikon inverted fluorescence microscope equipped with Metamorf software.

Changes in Gene Expression Induced by PDGF in Rat
VSMC-The goals of this study were to compare the effects of PDGF and hypertrophic stimuli on patterns of gene expression in VSMC. Microarray analysis was performed as described under "Experimental Procedures," using the Affymetrix Rat RAE-230 GeneChip. The RAE-230 GeneChip contains 31099 genes and expressed sequence tags; the first of the two chips were used for the experiments presented here. Each experiment was done in duplicate and compared with its control. Our criteria for changes in gene expression were either a p value of 0.0001 or a 2-fold change in value, and concordance of 4 of 4 comparisons. VSMC were stimulated with PDGF-BB for 6 or 24 h, and gene expression was compared with control cells  Tables A and B).
Our laboratory previously showed that PDGF-BB-mediated suppression of SM-specific gene expression is partly mediated by activation of both the H-Ras and phosphatidylinositol 3-kinase/Akt pathways (14,15). To investigate the contribution of these pathways to the changes in the global gene expression profile induced by PDGF, we compared the increases and decreases induced by PDGF stimulation to the changes seen in VSMC stably expressing H-Ras (14) or constitutively active Myr-Akt (15) (see Supplementary Tables A and B). In general, VSMC constitutively expressing H-Ras had similar changes as the ones stimulated with PDGF for 6 h. On the other hand, VSMC constitutively expressing Myr-Akt had the gene expression profile similar to the cells expressing PDGF for 24 h.
Changes in Gene Expression Induced by Hypertrophic Stimuli-We have previously demonstrated that AVP signaling in VSMC is mediated through coupling of the V1 receptor to G q , a member of the family of trimeric G proteins. This was assessed by transfection of constitutively active G-proteins, in which the intrinsic GTPase activity has been abolished by a point mutation. Since the constitutively active form of G␣ q (␣ q Q209L) was cytotoxic in VSMC, expression of another member of the G q family, G␣ 16 (␣ 16 Q212L) was used instead (12). Stable expression of G␣ 16 in VSMC gives a stronger, more consistent hypertrophic response compared with VSMC exposed to vasoconstrictors such as AVP (12). To compare the effects of PDGF and its downstream signaling pathways with changes in gene expression induced by hypertrophic stimuli, we examined VSMC stimulated with AVP for 6 or 24 h, and VSMC stably expressing constitutively active G␣ 16 (G␣ 16 Q212L). Similar to the study by Campos et al. (21) with angiotensin II, VSMC stimulated with AVP for either 6 or 24 h induced minimal changes in gene expression. AVP stimulation increased 11 (p Ͻ 0.0001; 3 of 4) or 3 (2-fold; 3 of 4) of the genes that were decreased by 24-h PDGF stimulation. However, of the genes increased by PDGF, only two (p Ͻ 0.0001; 3 of 4) were decreased by AVP. Whereas expression of G␣ 16 caused large scale changes in gene expression, we sought to focus on genes that were differentially regulated by G␣ 16 compared with PDGF. Tables I and II are lists of genes, regulated in opposing directions by PDGF at either time point compared with G␣ 16 . We identified 75 genes whose expression was increased by PDGF at either time point and decreased by expression of G␣ 16 (Table I) as well as 97 genes whose expression was decreased by PDGF at either time point and increased by expression of G␣ 16 (Table II).
Characterization and Verification of the Genes-As expected, a number of the genes differentially regulated by PDGF and G␣ 16 included contractile proteins, which have been used as markers for differentiated VSMC. These included SM22␣, SMmyosin heavy chain, tropomysin 1␣, and caldesmon 1 (Table  III), all of whose expressions were decreased by PDGF and increased by G␣ 16 . In addition, all of these genes were downregulated by expression of both H-Ras and Myr-Akt, except for caldesmon 1, which was only down-regulated by expression of H-Ras. In addition to SMC-specific genes, a number of transcription factors were differentially regulated by PDGF and G␣ 16 . These included C/EBP␦, retinol-binding protein-1, and Prx-2, which were induced by PDGF in an Akt-dependent manner, and decreased by expression of G␣ 16 (Table IV). As verification of the microarray data, we observed increased expression of C/EBP␦ by Western blotting in VSMC stimulated with PDGF for 6, 24, or 48 h as well as VSMC stably expressing H-Ras, and Myr-Akt, compared with their appropriate controls (Fig. 1A). In contrast, but consistent with its role in opposing PDGF, stable expression of G␣ 16 decreased expression of C/EBP␦ (Fig. 1B). To determine a functional link between increased C/EBP␦ expression and regulation of SMC-specific gene expression, the effects of overexpressing C/EBP␦ on SM ␣-actin promoter activity were determined. Overexpression of C/EBP␦ decreased basal SM ␣-actin promoter activity and markedly inhibited induction seen with AVP in VSMC (Fig.  1C), suggesting its involvement in regulation of SMC-specific gene expression. In contrast to C/EBP␦, PDGF stimulation  (Table IV). Consistent with the array data, Western blotting showed that GKLF (KLF4) protein expression was decreased at 24 h of PDGF stimulation and was also decreased in cells expressing Myr-Akt or H-Ras (Fig. 1A), whereas expression of G␣ 16 had the opposite effect, resulting in increased expression of GKLF (Fig. 1B).
PDGF and G␣ 16 had opposing effects on the expression levels of several cytoskeletal proteins ( Table V). Expression of nexilin, an F-actin-binding protein, was decreased at both 6 and 24 h of PDGF stimulation as well as by expression of H-Ras or Myr-Akt, whereas its expression was increased with G␣ 16 . Western blotting verified the array data showing decreased protein expression of nexilin following 24 or 48 h of PDGF stimulation, as well as decreased expression in cells expressing H-Ras and Myr-Akt compared with the appropriate empty vector controls ( Fig. 2A). Although not determined to be statistically significant by microarray analysis, Western analysis showed that expression of nexilin was increased by AVP at all time points examined, as well as in VSMC expressing G␣ 16 (Fig. 2B). In addition to nexilin, PDGF and H-Ras decreased and G␣ 16 increased expression of Rho-associated kinase-1 and four and a half LIM domain proteins 1 and 2 (Table V). Interestingly, although array analysis showed decreased expression of ezrin, another actin-binding protein, by 24-h PDGF treatment, Western blotting showed that PDGF increased protein levels of Ezrin at all time points examined ( Fig. 2A). Consistent with the array data, Western analysis also showed increased protein levels of ezrin by constitutive expression of H-Ras and Myr-Akt but decreased levels by constitutive expression of G␣ 16 (Fig. 2,  A and B).
Additionally, PDGF and G␣ 16 regulated the expression of several extracellular matrix proteins, proteoglycans, and matrix enzymes ( Table V). Expression of collagen type X ␣1, previously shown to be associated with cell hypertrophy (29,30), was decreased by 6 and 24 h of PDGF stimulation, as well as by expression of Myr-Akt, but was increased by 24 h of AVP stimulation as well as by expression of G␣ 16 . Real time quantitative PCR for ColX␣1, used to verify the array results, showed that PDGF and Akt suppressed, whereas AVP and G␣ 16 increased, ColX␣1 mRNA levels (Fig. 3A). No change was observed in clones expressing H-Ras as predicted by array data. In contrast, versican, an extracellular matrix proteogly-can of blood vessels, was increased following 24 h of PDGF and by constitutive activation of both H-Ras and Akt but decreased by constitutive G␣ 16 ; real time quantitative PCR was used to confirm the array data (Fig. 3B). Although it was not predicted, both 6-and 24-h AVP stimulation decreased the mRNA levels of versican in VSMC (Fig. 3B).
In addition to its reported effects on VSMC phenotypic modulation, PDGF is a strong mitogen inducing proliferation of cultured VSMC. Several proliferation-associated genes were identified from the analysis of the microarray data (Table VI). One of these, growth-related oncogene-1 (Gro), a chemoattractant chemokine, was increased following 6-and 24-h stimulation with PDGF. Real time quantitative PCR of Gro verified the effect of PDGF on mRNA expression, showing that PDGF stimulation increased the mRNA levels of Gro in an Akt-mediated manner (Fig. 4). Expression of G␣ 16 decreased Gro mRNA levels. Tissue inhibitor of metalloproteinase 1-collagenase ϩ1.7 ϩ1.7 NM_133537 Expi Extracellular proteinase inhibitor ϩ21 Ϫ100

FIG. 2. Regulation of cytoskeletal proteins in VSMC.
A, VSMC were serum-restricted for 24 h and stimulated with PDGF (20 ng/ml) for the indicated lengths of time, or VSMC stably expressing Myr-Akt or H-Ras were serum-restricted for 24 h. Cell extracts were prepared and immunoblotted with antibodies specific for nexilin and ezrin. B, VSMC were serum-restricted for 24 h and stimulated with AVP (10 Ϫ6 M) for the indicated lengths of time or VSMC stably expressing G␣ 16 were serumrestricted for 24 h, and extracts were immunoblotted with nexilin and ezrin.

FIG. 3. Regulation of extracellular matrix protein expression in VSMC.
A, VSMC were serum-restricted for 24 h and stimulated with PDGF (20 ng/ml) or AVP (10 Ϫ6 M) for the indicated lengths of time, or VSMC stably expressing Myr-Akt, H-Ras, or G␣ 16 were serum-restricted for 24 h and then total RNA isolated. Real time quantitative PCR for collagen type X 1␣ was performed as described under "Experimental Procedures." Ribosomal RNA was used for the normalization of cDNA from each sample. Results represent the mean and S.E. of two independent samples performed in at least triplicate. B, real time quantitative PCR for versican was performed as described under "Experimental Procedures." Ribosomal RNA was used for the normalization of cDNA from each sample. Results represent the mean and S.E. of two independent samples performed in at least triplicate.

Correlation of in Vitro Expression Patterns of Select Gene Products with in Vivo Expression Patterns in Injured Blood
Vessels-Vascular injury-induced VSMC phenotypic modulation is associated with the down-regulation of SMC-specific marker genes, such as SM ␣-actin and SM22␣, and contributes to neointima formation. Immunofluorescent staining was used on tissue sections from sham-injured, 7-day, and 4-week postballoon-injured rat thoracic aortas to determine whether the changes in gene expression observed in cultured SMC in response to PDGF and/or hypertrophic stimuli correlate with in vivo changes in SMC phenotype. We found barely detectable VSMC expression of C/EBP␦ in sham-injured vessels but increased accumulation of C/EBP␦ in the nuclei of neointimal VSMC at 7 days after injury (Fig. 5A). By 4 weeks postinjury, C/EBP␦ was restricted to the most luminal neointimal VSMC. Consistent with a functional role in suppressing ␣-actin promoter activity (Fig. 1C), high in vivo levels of C/EBP␦ were associated with low levels of ␣-actin expression in neiointimal VSMC (Fig. 5A). A similar, but cytosolic in vivo expression pattern was observed for versican ( Fig. 5B), another gene identified as being up-regulated by PDGF in an H-Ras-and Akt-dependent manner. Versican was only detectable in the endothelium of sham-injured vessels, but was significantly upregulated in the developing neointima at 7 days postinjury. In contrast, nexilin, down-regulated by PDGF, H-Ras, and Akt and up-regulated by AVP and G␣ 16 , was abundantly expressed in sham-injured vessels but significantly down-regulated in injured aortas (Fig. 5C). DISCUSSION Plasticity of VSMC resulting in phenotypic modulation is important in the physiologic and pathophysiologic state of the vascular wall. During the early stages of development or following injury to the blood vessels, VSMC present a dedifferentiated phenotype, which is characterized by low expression levels of SM-specific genes, including SM ␣-actin, SM22␣, calponin, and SM-myosin heavy chain (31,32). Previously, our laboratory and others showed that changes in SM-specific gene expression in cultures of adult VSMC could be modulated by hormones and growth factors (13-16, 33, 34). In general, PDGF suppressed expression of these proteins, whereas hypertrophic stimuli, such as AVP, increased expression. Although changes in contractile proteins have been widely used to define VSMC phenotype, the functional responses to these specific agonists will undoubtedly be a consequence of changes in expression of broad families of genes. The advent of microarray technology allows us to begin to define these changes, with the ultimate goal of defining critical events that impact on the physiologic properties of these cells. Since PDGF and hypertrophic stimuli have many similar effects on VSMC, including mobilization of intracellular Ca 2ϩ , activation of MAP kinase family members, and increases in protein synthesis, it is likely that there will be significant numbers of genes whose expression will be similarly regulated by both classes of agents. However, in this study, we focused on identifying genes whose expression was regulated in opposing directions by PDGF and hypertrophic stimuli. We hypothesized that these genes would be markers of phenotypic modulation and would be altered during vascular injury.
As a validation of our approach, we identified a number of contractile proteins, which were differentially regulated by the two classes of stimuli, many of which have previously been reported. The majority of these genes contain multiple CArG boxes in their promoter regions, and expression is likely to be controlled through SRF-dependent pathways (see Ref. 35). In addition, we identified a large number of differentially regulated genes that have not been reported to be controlled through SRF-dependent pathways. Many of these encode proteins functionally involved in cytoskeletal and cell-matrix interactions (see Tables V and VI). Nexilin, which is downregulated by PDGF, is an F-actin-binding protein localized to cell-matrix junctions (36) and may be involved in focal contacts. Adducin 3 ␥ is a cytoskeletal assembly protein up-regulated by PDGF and down-regulated by hypertrophic stimuli, involved in actin-driven migration (37). Another cytoskeletal organizer protein ezrin was up-regulated by PDGF and down-regulated by G␣ 16 . A reduction in the expression of nexilin combined with increases in ezrin and adducin would be predicted to increase cell migration, consistent with the effects of PDGF. We also demonstrated that levels of nexilin are decreased in blood vessels after injury. This novel finding suggests a critical connection between PDGF-induced gene expression and neointimal formation.
The opposing effects of PDGF and hypertrophic stimuli on cytoskeletal related proteins are likely to lead to changes of F-actin/G-actin ratio in VSMC. Numerous studies have demonstrated that alterations in F-actin/G-actin are critical in controlling SRF-dependent genes (38 -41). This is mediated at least in part through translocation of coactivators of the myocardin/MKL family (reviewed in Ref. 42). Our data show that

FIG. 4. Regulation of proliferationrelated protein expression in VSMC.
A, VSMC were serum-restricted for 24 h and stimulated with PDGF (20 ng/ml) or AVP (10 Ϫ6 M) for the indicated lengths of time or VSMC stably expressing Myr-Akt, H-Ras or G␣ 16 were serum-restricted for 24 h and then total RNA isolated. Real time quantitative PCR for Gro was performed as described under "Experimental Procedures." Ribosomal RNA was used for the normalization of cDNA from each sample. Results represent the mean and S.E. of two independent samples performed in at least triplicate.
Rho-associated kinase-1, a downstream effector of RhoA, was down-regulated by PDGF and up-regulated in response to expression of G␣ 16 . Decreased ROCK expression would be predicted to inhibit RhoA signaling and block SRF-dependent transcription (43) and induction of SM-specific genes in VSMC (44).
Alterations in expression of matrix-related proteins were also observed. Expression of collagen type X ␣ 1 was decreased by PDGF and strongly up-regulated by AVP stimulation, cor-relating with a hypertrophic response (30). Versican (CSPG2), an extracellular matrix proteoglycan, was up-regulated by PDGF and down-regulated by G␣ 16 and AVP. Versican is secreted by VSMC in normal blood vessels, and its expression is increased in atherosclerosis and restenosis and in cells isolated from neointima (24,45,46). Coordinated down-regulation of hypertrophy-related matrix proteins (ColX␣1) and up-regulation of migratory and proliferative matrix proteins (versican) could lead to a PDGF-mediated dedifferentiated, migratory FIG. 5. Correlation of in vitro expression patterns of C/EBP␦, versican, and nexilin with in vivo expression patterns in injured blood vessels. Adult thoracic aortas from sham-injured, 7-day and 4-week post-balloon catheter-injured animals were processed for immunofluorescence analysis of C/EBP␦ (A), versican (B), and nexilin (C) as described under "Experimental Procedures." A, thoracic aorta sections were double-stained with C/EBP␦-specific (red; Alexa Fluor-568), and SM ␣-actinspecific (green; Alexa Fluor-488) antibodies. Nuclei of VSMC were stained with DAPI (blue). B, thoracic aorta sections were double-stained with versican-specific (red; Alexa Fluor-568), and SM ␣-actin-specific (green; Alexa Fluor-488) antibodies. Nuclei of VSMC were stained with DAPI (blue). A, thoracic aorta sections were stained with nexilin (green; Alexa Fluor-488). Nuclei of VSMC were stained with DAPI (blue). Lines in each image indicate the tunica media, and brackets indicate neointima. phenotype of VSMC, an important step in neointima formation.
A change in the expression of a transcription factor can regulate the transcriptional activity of several genes through cascades of transcriptional regulations. Increased expression of C/EBP␦, observed in our experiments, has been correlated with increased PDGF␣R expression and vascular remodeling (47). Overexpression using molecular transfection approaches demonstrated that C/EBP␦ decreased SM-specific gene expression (Fig. 1C), and increased expression was observed in neointima formation in vivo (Fig. 5A). C/EBP␦ could induce the transcription of a number of other genes involved in SMC phenotypic modulation. Preliminary analysis of the promoter region of human versican (accession number U15963) showed several putative C/EBP␦ binding sites (48), suggesting C/EBP␦dependent up-regulation of versican.
Activation of the PDGF receptor leads to the engagement of multiple downstream signaling cascades. By correlating changes in gene expression in response to PDGF with changes induced by expression of either gain-of-function H-Ras or Akt, it appears that these effectors have distinct roles in mediating the response to PDGF. Activation of H-Ras was more critical for the early changes in gene expression mediated by PDGF, whereas activation of Akt was better correlated with changes observed at longer time points (Table VII). We would therefore propose that PDGF rapidly activates the Ras pathway, which is critical for initial changes in gene expression. This is followed by a more sustained activation of the phosphatidylinositol 3-kinase/Akt pathway (15), which is a major determinant of the delayed changes in gene expression. It should be noted that expression of either H-Ras or Myr-Akt altered expression of several genes not shown to be regulated by PDGF. We attribute this to the fact that these cells are expressing constitutively active forms of these enzymes, whereas PDGF stimulation leads to a transient activation of these pathways. Furthermore, there were significant numbers of genes whose expression was altered by PDGF but not by either H-Ras or Akt. This would indicate that PDGF engages additional effectors, as has been well demonstrated. In particular, we have not examined the role of phospholipase C␥ in this study.
Phenotypic modulation of VSMC by environmental factors requires coordinated action of common signaling pathways leading to large scale changes in gene expression, which ultimately result in the distinct biology of VSMC in disease states. The balance between mitogenic and hypertrophic stimuli leading to specific gene expression profiles determines the physiological or pathological states of the arterial vessels. We therefore propose that genes differentially regulated by these two classes of stimuli provide potential biomarkers for the phenotypic state of VSMC both in vitro and in vivo. The data pre-sented here will increase our understanding of the molecular pathways mediating this response in normal development and in pathophysiologic states.