Platelet-derived Growth Factor-specific Regulation of theJE Promoter in Rat Aortic Smooth Muscle Cells*

JE is a member of the family of “immediate early” genes induced by growth factors and cytokines.JE encodes a low molecular weight secretory glycoprotein analogous to the human monocyte chemoattractant protein, MCP-1. JE and MCP-1 proteins are thought to play an important role in inflammation and in the recruitment of monocyte/macrophages to the vessel wall during the development of atherosclerosis. We have previously reported that the induction of JE in rat aortic smooth muscle cells (SMC) was specific to platelet-derived growth factor (PDGF) and was not seen with other growth agonists. Using a luciferase reporter system and transient transfection assays of rat aortic SMC, we now report the identification of a region in the proximal rat JE promoter that is responsive to PDGF but not to other growth factors (angiotensin II and α-thrombin) or cytokines (interleukin 1-β and tumor necrosis factor-α). The full response to PDGF (∼6-fold) requires the cooperative activity of two potentially novel cis-acting elements, at positions −146 to −128 and −84 to −59. While each element produces a different pattern in electrophoretic mobility shift assays, they appear to bind the same PDGF-responsive species. Further analysis of these regions should provide important insights into PDGF-specific responses in vascular SMC.

Growth and migration of vascular smooth muscle cells (VSMC) 1 are critical events in the pathogenesis of atherosclerosis, hypertension, and angiogenesis (1). VSMC exhibit two types of growth: hyperplasia, characterized by increased DNA and protein synthesis as well as cell division, and hypertrophy, characterized by increased cell size and protein content without DNA synthesis or cell division (2). VSMC hyperplasia is an important feature of atherogenesis and involves the migration of VSMC from the vessel media to the intima and the proliferation of medial and intimal VSMC (1). VSMC hypertrophy is more typical of the vascular response to chronic hypertension, although hyperplasia is seen in hypertension as well (2). The growth responses of cultured VSMC appear to be agonist-specific. In adult rat aortic VSMC, platelet-derived growth factor (PDGF) and serum induce hyperplasia (3). In the same cells, angiotensin II (Ang) and ␣-thrombin induce hypertrophy (3)(4)(5)(6)(7)(8)(9).
There is a considerable body of literature involving the response of SMC to PDGF and Ang (reviewed in Refs. 3 and 10). Ang and PDGF activate many of the same intracellular signals, including the activation of phospholipase C, the induction of the mitogen-activated protein kinase cascade, and the activation of the Na ϩ -H ϩ antiporter, and induce similar sets of early response genes (reviewed in Refs. 3 and 10). Attempts by this laboratory to identify differences in gene expression in response to Ang and PDGF using differential screening or high resolution two-dimensional protein gels (11,12) have been unsuccessful, underscoring the similarities in signaling and gene induction between the two agonists. These studies suggest that there are a limited number of molecular events that distinguish the hypertrophic and hyperplastic responses of VSMC.
JE is a member of the family of "immediate early" genes (13) induced by growth factors in various cell types, including macrophages, endothelial cells, fibroblasts, and vascular SMC (14 -19). The JE product is a secretory glycoprotein of the "C-C" chemokine subfamily (20) that appears to be the analog of the human monocyte chemoattractant protein, MCP-1 (21). JE/ MCP-1 protein has been identified in early human atherosclerotic lesions (22) and has also been found in smooth muscle cells and macrophages of advanced human, primate, and rabbit atherosclerotic plaques (2,(23)(24)(25)(26). JE mRNA is also induced in the media within hours of experimental rodent balloon arterial injury (19). While numerous agents have been shown to have monocyte chemotactic activity, JE/MCP-1 appears to account for all of the monocyte chemotactic activity secreted by cultured rat (27) and human (28) SMC in response to PDGF or oxidized low density lipoprotein, respectively. JE/MCP-1 may thus play a particularly important role in SMC-mediated monocyte recruitment during the development of atherosclerosis and in acute vessel injury (29).
We have previously examined the expression of JE mRNA and protein in rat aortic SMC (19). In contrast to other cell types, where the induction of JE mRNA is common to a variety of growth agonists, the induction of JE mRNA and chemotactic activity in rat aortic SMC was specific to calf serum and PDGF and was not seen with other agonists, including Ang or ␣-thrombin (19). In addition, the induction of JE mRNA by PDGF appears to be independent of activation of protein kinase C, mobilization of intracellular calcium, or stimulation of the Na ϩ -H ϩ exchanger, signals shared by Ang and PDGF and often involved in the induction of immediate early genes in SMC (19). Accordingly, we have employed the rat JE gene as a model system for identifying potential pathways that differentiate the response to Ang and PDGF. We now report the identification of two sites in the rat JE promoter that act synergistically to regulate JE transcription in a PDGF-specific manner. These sites do not include previously identified PDGF-responsive elements and therefore may be part of a novel PDGF-specific pathway. Such pathways may be involved in differentiating the hypertrophic and hyperplastic phenotypes.

MATERIALS AND METHODS
Growth Factors and Other Reagents-Recombinant human PDGF (BB, AA, and AB isoforms) was obtained from Boehringer Mannheim. Recombinant murine IL-1␤ and TNF␣ were obtained from R&D Systems Inc. (Minneapolis, MN). Ang was obtained from Sigma. Purified ␣-thrombin was obtained from Dr. Peter Harpel (Mount Sinai School of Medicine, New York). Thermophilic DNA polymerase was purchased from Hoffman-La Roche (Nutley, NJ). TransCruz Gel Supershift reagents containing antibodies to c-Jun/AP-1 or Ets-1/Ets-2 (1 g/ml) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).
Cell Culture-SMC were isolated from thoracic aortas of Sprague-Dawley rats by enzymatic dissociation and grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated calf serum (Sigma) and 100 units/ml penicillin/streptomycin as described previously (30). Cells from spontaneously hypertensive rats were prepared and grown in the same fashion; rats were 16 weeks of age at the time of harvesting. For measurement of DNA or protein synthesis, SMC were plated in 12-well culture dishes at 10 4 cells/well and incubated for 48 h in serum-free medium supplemented with bovine serum albumin, insulin, transferrin, and ascorbate (31,32). SMC were then washed with phosphate-buffered saline and incubated with growth agonists in the presence of [ 3 H]thymidine or [ 3 H]leucine (1 Ci/well) for 24 h. Cells were then fixed for 60 min at 4°C with 10% trichloroacetic acid. Incorporation of the radiolabeled material was determined by liquid scintillation spectrometry of the trichloroacetic acid-precipitable material as described previously (4). Duplicate experiments were performed using three wells per treatment per experiment.
Nuclear Run-on Assays-60 -80% confluent SMC were incubated in fresh DMEM plus 10% calf serum for 24 h. After treatment with PDGF BB for varying times, nuclei were collected and used for run-on transcription assays as described previously (19). The resulting [ 32 P]UTPradiolabeled RNA transcripts were hybridized to linearized and denatured full-length rat JE cDNA immobilized on nitrocellulose filters using a Hybri-Dot blotter (Life Technologies, Inc.) as described (19). Glyceraldehyde-3-phosphate dehydrogenase cDNA and the pBluescript vector (Stratagene, La Jolla, CA) were used as positive and negative controls, respectively. Results were analyzed using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Levels of JE transcript were normalized to those of the constitutively expressed glyceraldehyde-3phosphate dehydrogenase after subtraction of background (pBluescript) and presented as -fold increase over basal activity Ϯ S.E. Studies were performed in triplicate.
Transient Transfection Assays-SMC were plated at 10 6 cells/ 100-mm dish 24 h prior to transfections. Cells were co-transfected with 20 g of JE reporter constructs, 10 g of human growth hormone expression vector (pXGH5; Promega, Madison, WI), and 10 g of carrier DNA (pGEM4Zϩ, Promega). Transfections were carried out by CaCl 2 precipitation as described (33), except that chloroquine (final concentration of 10 mM) was added to the CaCl 2 /DNA mixture. Cells were exposed to the mixture for 12-14 h, washed with phosphate-buffered saline, and incubated in DMEM plus 10% calf serum for 8 h. The medium was then switched to DMEM plus 2% calf serum for an additional 38 -42 h. The medium was then collected and assayed for human growth hormone using a growth hormone detection kit (Nichols Institute, San Juan Capistrano, CA). Fresh DMEM supplemented with growth factors was then added for 4 -6 h, and protein extracts were collected and analyzed for luciferase activity (see below).
Plasmid Construction-A 1093-bp fragment containing the 5Ј region of the rat JE gene (corresponding to bases Ϫ1040 to ϩ40 of the published sequence (34), with an additional 13 bases at the 5Ј-end) was cloned from a rat genomic library (CLONTECH, Palo Alto, CA). This fragment was ligated into the luciferase reporter plasmid PXP2 (35) to produce construct pJE-1053. Constructs with serial deletions at the 5Ј-end of pJE-1053 were generated using the polymerase chain reaction or restriction digestion and were named according to the number of base pairs they contained upstream of the start site previously established for the rat JE gene (34). All plasmid constructions were confirmed by DNA sequence analysis using an Applied Biosystems DNA sequencer, model 373A.
Luciferase Assay-SMC were washed twice at 25°C with phosphate-buffered saline, lysed in luciferase cell culture lysis reagent (Promega) and assayed for luciferase activity in a BioOrbit 1251 luminometer (Wallac, Gaithersburg, MD) using luciferase assay reagent (Promega). Levels of luciferase activity were normalized to levels of growth hormone detected in the medium collected from the same plate. The normalized luciferase activity detected in stimulated (i.e. growth agonisttreated) SMC was expressed as -fold increase Ϯ S.E. over luciferase activity in unstimulated SMC. Each experiment was performed using duplicate plates and was repeated at least twice. Electrophoretic Mobility Shift Assay (EMSA)-SMC were grown and treated as described for nuclear run-on studies. Preparation of nuclear protein extracts was carried out as described by Andrews and Faller (36). Double-stranded DNA probes were end-labeled with [␥-32 P]ATP (3000 Ci/mmol; NEN Life Science Products) using T4 polynucleotide kinase (New England Biolabs, Beverly, MA). 20 -50,000 cpm (1-2 ng of DNA) were incubated with 6 g of nuclear extracts for 30 min on ice. To minimize nonspecific DNA binding, 2 g of poly(dI-dC) (Boehringer Mannheim) was added to each reaction. Binding reactions contained 20 mM HEPES, pH 7.9, 12% glycerol, 1 mM MgCl 2 , 1 mM EDTA, 1 mM dithiothreitol, and 60 mM KCl. Unlabeled competitor DNA was added 10 min prior to the initiation of reactions. For EMSA supershift analysis, 4 l of TransCruz Gel Supershift reagents (final concentration of 0.25 g of specific antibody/l) were added to the samples together with the labeled probe. Protein-DNA complexes were resolved by electrophoresis on nondenaturing 4% polyacrylamide (40:1) at 4°C in 50 mM Tris, 50 mM boric acid, 1 mM EDTA. Gels were transferred to polyacrylamide gel lift paper (Schleicher & Schuell), dried, and analyzed by overnight autoradiography.

PDGF BB Stimulates JE Transcription in Rat Aortic
SMC-To verify that the accumulation of JE mRNA in response to PDGF was due in part to stimulation of transcription, nuclear run-on assays were performed (Fig. 1). JE transcripts were noted in unstimulated SMC. This is consistent with the presence of detectable levels of JE mRNA by Northern blot analysis (19). 20 ng/ml PDGF BB caused a 2.8 Ϯ 0.3-fold induction of JE transcription at 30 min (p Ͻ 0.05). JE transcription returned to near base-line level (1.5 Ϯ 0.1, S.E. not significant) within 1 h.
Growth Factor Specificity of the Rat JE Promoter in Aortic SMC-To examine the regulation of the JE promoter in rat SMC, a 1093-bp fragment, corresponding to bases Ϫ1053 to ϩ40 of the published rat JE genomic sequence (34) was isolated and ligated upstream of the luciferase reporter gene. The resulting construct, pJE-1053, was tested for luciferase activity in transient transfections. As shown in Fig. 2, PDGF BB upregulated the JE promoter construct, whereas 1 M Ang failed to significantly increase JE promoter activity. PDGF AB (20 ng/ml) and 10% calf serum had effects on JE promoter activity (not shown) identical to that shown for PDGF BB. In contrast, 1 M ␣-thrombin also failed to induce the JE promoter. To determine whether the responses to Ang and PDGF were unique to one type of SMC, cells were also isolated from the aorta of spontaneously hypertensive rats and used in transient transfection assays. PDGF induced similar levels of luciferase activity in these SMC, whereas Ang failed to elicit a response (data not shown).
To verify that the lack of responsiveness to growth agonists other than PDGF BB was not due to the absence of functionally coupled receptors, the effects of these agonists on SMC growth were evaluated. As shown in Fig. 3A, PDGF BB increased DNA synthesis by 820 Ϯ 85% and 752 Ϯ 224%, respectively. Typical of their hypertrophic action on adult rat aortic SMC (5, 6, 11), Ang and ␣-thrombin significantly increased protein synthesis (233 Ϯ 15% and 257 Ϯ 12%, respectively) but had no effect on DNA synthesis (Fig. 3B). Of note, PDGF AA failed to increase DNA or protein synthesis, suggesting a low level of functionally coupled PDGF ␣-receptors in these cells (27).
In addition to growth agonists, a variety of cytokines are also known to induce JE in fibroblasts and endothelial cells (37,38).
In contrast, neither TNF␣ nor IL-1␤ induce monocyte chemotactic activity in the adult rat SMC culture system used for these studies (27). As shown in Fig. 2, 20 ng/ml IL-l␤ or 20 ng/ml TNF␣ failed to up-regulate transcriptional activity of the JE-1053 reporter construct. Because IL-l␤ and TNF␣ do not substantially alter DNA or protein synthesis in aortic SMC, the activity of these agonists was verified by their ability to induce tissue factor (data not shown; methods described in Ref. 39).
Identification of PDGF-specific Response Elements in the Proximal JE Promoter-To locate the region(s) in the proximal JE promoter that confer regulation by PDGF BB, constructs containing serial deletions from the 5Ј-end of the JE promoter were tested for luciferase activity. As shown in Fig. 4, the response to PDGF remained unchanged for all deletions up to Ϫ146. Deletion of bases Ϫ146 to Ϫ128 (corresponding to bases 894 -912 of the published genomic sequence), containing the sequence TCCAAGGGCTCGGCACTTA (element 1), resulted in an ϳ50% loss of PDGF stimulation. Deletion of the region between Ϫ128 and Ϫ84 did not additionally affect the PDGF response. In addition, PDGF specificity was retained (see Fig.  4, inset). However, removal of bases Ϫ84 through Ϫ59 (corresponding to positions 956 -981 of the genomic sequence), containing the sequence TGATGCTGCTCCTTGGCACCAACCAC (element 2), eliminated most of the remaining response to PDGF, without changing the activity of the unstimulated promoter. Deletion of bases Ϫ59 through Ϫ42, containing an AP-1-like site (TGACTCC), resulted in a complete loss of basal promoter activity.
Elements 1 and 2 are located 43 bp apart. To investigate the importance of this 43-bp sequence, element 1 was ligated in the 5Ј 3 3Ј and 3Ј 3 5Ј orientations directly upstream of element 2. As shown in Fig. 5, these ligations (constructs pJE-84(1S) and pJE-84(1AS)) did not restore PDGF-induced JE transcription to wild-type levels. Element 1 also failed to restore wildtype levels of JE transcription when placed in its proper position relative to element 2 but in the opposite (3Ј 3 5Ј) orientation (pJE-128(1AS)). Thus, the cooperativity between elements 1 and 2 may be positionally dependent and require the presence of at least some of the intervening 43 bases. Element 1 had no significant activity when placed directly in front of the AP-1-like site (pJE-59(1S)) in the absence of element 2.
To identify region(s) of element 2 critical for PDGF regulation, four reporter constructs containing 7-8-bp cluster mutations in element 2 in the context of a fully active JE promoter containing wild type element 1 (Fig. 6, constructs A-D) were analyzed. All constructs had ϳ50% of the activity of the wild type promoter, suggesting that the entire length of element 2 is required for the wild-type response to PDGF. Complete mutation of element 2 (construct E) also resulted in ϳ50% residual activity, demonstrating that in its appropriate context within the promoter, element 1 retained activity. This construct also failed to respond to Ang (not shown).
Elements 1 and 2 Exhibit Different Binding Patterns on EMSA-To demonstrate that regions conveying PDGF inducibility bind proteins, EMSAs were performed using nuclear extracts from rat aortic SMC. As shown in Fig. 7A, a probe derived from element 1 bound several protein fractions in a constitutive fashion. In addition, the binding of one fraction (marked by an arrow) was rapidly induced by PDGF. Binding of this fraction was not seen early after treatment with Ang but was noted 60 min after Ang treatment. Element 2 also bound several protein fractions (Fig. 7B). The slower mobility fraction was bound constitutively. However, binding of the faster moving species (marked by an arrow) was rapidly induced by PDGF but not by Ang. All fractions were successfully competed using a 300-fold molar excess of cold probe. Similar binding patterns were observed using nuclear extracts from SMC derived from spontaneously hypertensive rats (e.g. see Fig. 8).
To further ascertain the relationship between the proteins binding to each element, competitions were performed. As shown in Fig. 8A, an excess of cold element 2 competed away the PDGF-regulated species binding to element 1 but had little effect on the unregulated species. As expected, cold element 1 competed away all binding. Similarly, element 1 competed away the PDGF-regulated species binding to element 2, but had no effect on the binding of the unregulated species (Fig.  8B). These data suggest that both elements bind the same PDGF-regulated protein; in addition, each element constitutively binds a distinct set of proteins. Members of AP-1 and Ets families of transcription factors have been previously shown, respectively, to up-regulate JE/MCP-1 transcription (40) and respond to PDGF stimulation in vascular smooth muscle cells (41). To investigate whether these proteins bind to either element, supershifts were performed using antibodies to c-Jun/ AP-1 or Ets-1/Ets-2 (Fig. 8). No supershift bands or disruption of existing complexes was observed, suggesting that the nuclear fractions do not contain AP-1 or Ets proteins. DISCUSSION Previous studies from this laboratory have demonstrated that the accumulation of JE mRNA and the induction of chemotactic activity in rat aortic SMC is specific for PDGF and is not induced by a variety of other growth factors or cytokines Reporter constructs pJE-84(1S) and pJE-84(1AS) contain element 1 directly upstream of element 2 in sense and antisense orientations, respectively. pJE-128(1AS) contains element 1 in its correct location but in the antisense orientation. pJE-59(1S) contains element 1, in the sense orientation, directly upstream of the AP-1 site and in the absence of element 2. The constructs were transiently expressed in SMC and stimulated with PDGF as described in Fig. 4. Luciferase activity from triplicate experiments performed on two different SMC strains is normalized to co-transfected human growth hormone and expressed as -fold increase over levels in unstimulated SMC. The reporter construct pJE-1053 was used in each transfection assay as a control for the full response to PDGF. The construct pJE-59, containing the AP-1 site, was used as a control for the minimal promoter; the PDGF-induced -fold increase in normalized luciferase activity for pJE-59 is represented by the dotted line. Error bars show S.E. p Ͻ 0.01 between unstimulated and PDGF-stimulated constructs pJE-84(1S), pJE-84(1AS), and pJE-128(1AS). p Ͻ 0.05 between the same constructs and PDGF-stimulated pJE-59. S.E. value was not significant between PDGF-stimulated pJE-59(1S) and pJE-59. known to be active in these cells (27). This paper describes the first analysis of the JE promoter in SMC and reports the identification of a PDGF-responsive area (Ϫ146 to Ϫ59) in the proximal rat JE promoter. In addition, it demonstrates that this area has high specificity for PDGF and suggests that this may be responsible for the PDGF-specific accumulation of JE mRNA seen in rat SMC.
The full response to PDGF (ϳ6-fold) appears to require the cooperative activity of two cis-acting elements at positions Ϫ146 to Ϫ128 (element 1) and Ϫ84 to Ϫ59 (element 2). Either element alone is sufficient to confer PDGF-specific inducibility on a luciferase reporter construct, although the level of luciferase activity is ϳ50% of that seen with constructs containing both elements. Element 1 has no activity in the antisense orientation when placed immediately upstream of element 2 or when placed upstream of the AP-1-like site in the absence of element 2. Its ability to confer full activity to element 2-containing constructs thus appears to be dependent upon its position and orientation. Element 2 generates two major bands on EMSA. The amount of the slower moving species does not change significantly in response to PDGF, suggesting that the observed DNA-protein interaction is constitutive. In contrast, a faster moving species appears rapidly in response to PDGF but not Ang. As suggested by the deletion analyses, this species is likely to be of paramount importance in the PDGF-specific induction of JE. Element 1 also binds several species, one of which is responsive to PDGF and may be the same PDGF-responsive species bound by element 2. The appearance of the PDGF-responsive species has a time course consistent with the nuclear run-on experiments. Of note, no early binding is seen in response to Ang, but some binding is seen 1 h after Ang treatment. This later appearance FIG. 6. Mutation analysis of element 2. Luciferase reporter constructs containing a 646-bp JE promoter sequence (bases Ϫ606 to ϩ40) with a wild type element 2 (WT), 7-8-bp cluster mutations (sequences underlined) in overlapping regions of element 2 (A-D), or a mutation of the entire element 2 (E) were transiently expressed in SMC and stimulated with PDGF as described in Fig. 4. Luciferase activity from triplicate experiments is normalized to co-transfected human growth hormone and expressed as a fraction of the activity of the wild type construct induced by PDGF. Error bars show S.E. p Ͻ 0.01 for all constructs relative to the wild type. FIG. 7. EMSA analysis of elements 1 and 2: identification of PDGF-responsive complexes. EMSA was performed using nuclear extracts from unstimulated SMC (0) or stimulated with 20 ng/ml PDGF or 1 M Ang for the times indicated. Extracts were incubated with radiolabeled double-stranded oligonucleotides corresponding to element 1 (A) or element 2 (B). The first lane of each gel contains the radiolabeled probe (Probe) incubated under identical conditions without nuclear extracts. The last lane represents a competition (Comp) in which the radiolabeled probe was incubated with nuclear extracts derived from SMC stimulated with PDGF for 15 min in the presence of a 300-fold molar excess of unlabeled specific competitor DNA. Arrows represent PDGF-inducible binding complexes. Gels were run under identical conditions. may reflect secondary, rather than direct, effects of Ang receptor stimulation and apparently is not sufficient for inducing the JE promoter. Further studies will be necessary to fully characterize the mechanism(s) by which elements 1 and 2 cooperate in regulating JE promoter activity.
It should be stressed that the studies presented above do not establish the precise boundaries of the PDGF-responsive elements. As shown in Fig. 6, mutations in any region of element 2 abolished the activity. This suggests that the entire sequence may be necessary for binding. Preliminary gel shifts (data not shown) with the mutations examined in Fig. 6 showed that disruption of any part of the molecule resulted in a loss of the PDGF-responsive band. Given the complexity of the binding patterns for both elements and the presence of several species that are bound constitutively, it is highly likely that the PDGF effect requires the interaction of several proteins bound to different regions of the elements. The lack of effect of antibodies to c-Jun/AP-1 or Ets-1/Ets-2 on the binding of the PDGFresponsive species raises the possibility that the protein(s) involved may be different from those previously shown to mediate PDGF responses or to regulate the JE or MCP-1 genes.
Searches of GenBank TM (release 90.0, 8/95), EMBL (release 43.0, 6/95), and Eukaryotic Promoter (release 43, 6/95) data bases with both putative elements and the entire fragment containing both elements suggest that the sequences conveying PDGF specificity may be novel. Element 1 has high homology (Ͼ70% identity) to a 5Ј-flanking region of Mus musculus cathepsin B (accession number X76621) and to areas in the promoter regions of Aspergillus niger PX18 (accession number M90701), Homo sapiens gastrin (accession number EDP 25015), and M. musculus keratin (accession number EDP 32005). Element 2 has high homology only with sequences in the 5Ј region of the murine macrophage inflammatory protein 2 gene. Macrophage inflamatory protein 2 protein is a cytokine with potent chemotactic activity for human polymorphonuclear leukocytes (42,43). Interestingly, none of the genes identified contain areas of high homology with both elements, and searches of the literature did not provide evidence that these genes were PDGF-responsive. No regions highly homologous to either element exist in the human MCP-1 gene, suggesting that they may be species-specific. It should be noted that while PDGF and MCP-1 have been co-localized in human plaques (23,26), the regulation of MCP-1 by PDGF in human SMC has not as yet been described.
Regulatory elements in the promoter regions of JE/MCP-1 genes have been identified by several groups of investigators (34, 44 -46). Timmers et al. (34) established that the Ϫ70/Ϫ38 region of the rat JE promoter was necessary for its basal activity in transiently transfected 3T3 cells and established that the region Ϫ141/Ϫ88 was essential for the response to the phorbol ester TPA. Our study also suggests that the AP-1-like site located within the Ϫ70/Ϫ38 region is necessary for basal promoter activity in rat aortic SMC. Li and Kolattukudy (44) established that two elements (at positions Ϫ156 to Ϫ150 and Ϫ128 to Ϫ123) in the promoter region of human MCP-1 were necessary and sufficient to confer inducibility by TPA onto a CAT reporter construct transiently expressed in human glioblastoma cells. Activation of human MCP-1 by TNF␣ occurred via different cis-acting elements, because TNF␣ did not activate the TPA-inducible reporter constructs.
Freter et al. (45) identified two cis-acting regions that were necessary for the activation of the murine JE gene by serum, PDGF, IL-1, and double-stranded RNA in transiently transfected BALB/c 3T3 fibroblasts. That study did not demonstrate activation of reporter constructs and utilized RNase protection assays with a series of tagged genomic JE probes. The distal 240-bp sequence in the 5Ј-flanking JE region (positions Ϫ2537 to Ϫ2298 relative to the start site) had weak transcriptional enhancer activity. The other element identified in the study was a novel 7-mer, TTTTGTA, located in the 3Ј region of the murine JE after the stop codon. Its activity was position-dependent and orientation independent. The 7-mer is present in 3Ј regions of a number of transcription factors and cytokines belonging to the immediate early gene class, including human MCP-1. In follow-up studies (46,47), four PDGF-regulated elements were identified within the region located in the FIG. 8. Evidence for relatedness of PDGF-responsive complexes. EMSA was performed as described in Fig. 7 with radiolabeled double-stranded oligonucleotides corresponding to element 1 (A) or element 2 (B). Nuclear extracts were derived from SMC stimulated with PDGF for 30 min. Lanes E1 and E2 represent competition experiments in which a 300fold molar excess of unlabeled element 1 or element 2 was used, respectively. Lanes Jun Ab and Ets Ab represent supershift experiments in which supershift antibody reagents were used at a final concentration of 0.25 g/l incubation mix. The arrows represent the PDGF-inducible binding complexes. Gels were run under identical conditions. 240-bp 5Ј-flanking murine JE sequences. Two of these elements (I and IV) are similar in sequence and bind several forms of NF-B. The other two elements (II and III) were previously uncharacterized sequences that bind a PDGF-activated serinethreonine phosphoprotein found in the nucleus of PDGFtreated 3T3 fibroblasts.
In the rat JE gene, murine elements I and IV are not present. There is a region in the rat JE promoter (positions Ϫ472 to Ϫ449) with a 72% homology to murine element II in reverse orientation; however, our data indicate that the presence of this region is not required for activation of JE in SMC by PDGF. The rat JE gene also possesses a region with 69.2% homology to murine element III in reverse orientation; this region is located 15 bp downstream of the initiation codon. Thus, the pathways that lead to transcriptional induction by PDGF of JE in mouse appear to be different from those in rat and may also be different in SMC from other cell types.
We have previously reported that changes in mRNA stability were responsible for part of the increase in JE mRNA levels seen after PDGF treatment (19). In that study, nuclear run-on experiments, performed at 1 and 2 h, did not suggest a major transcriptional component. These time points were originally chosen because of the delayed peak in JE mRNA levels. The current study employed considerably earlier time points (15 and 30 min) to examine JE transcription. Using these time points, PDGF was found to cause a marked but very transient increase in JE transcription. Taken together, these studies suggest that the initial response to PDGF is an elevation in transcription, which is short lived. This is accompanied by an increase in mRNA stability (half-life of 2.5 versus 0.5 h), resulting in a prolonged elevation in JE mRNA levels beyond that seen with other immediate early genes.
As noted in the Introduction, the intracellular signaling pathways responsible for JE mRNA accumulation in rat aortic SMC remain to be determined. Initial studies employing RNA blot analyses and a variety of inhibitors failed to demonstrate a role for protein kinase C, mobilization of intracellular calcium, activation of the Na ϩ -H ϩ exchanger, and changes in cyclic AMP (19). While these studies were by no means exhaustive and do not rule out participation of any of these signals, they do suggest that the induction of JE mRNA in rat aortic SMC by PDGF involves signaling pathways distinct from those involved in the induction of a number of other immediate early genes in rat SMC, including c-fos, KC, and tissue factor (30,48,49). The current study, demonstrating potentially novel PDGFspecific elements in the rat JE gene, gives further credence to the concept that the regulation of JE in rat aortic SMC may involve an unusual set of signaling pathways.
PDGF and Ang have been implicated in the development of atherosclerosis, in the response of the vessel wall to injury, and in the vascular changes associated with hypertension (1, 2, 50). While considerable information exists about the signals induced in VSMC by PDGF and Ang, the induction of JE remains one of the few phenomena described that distinguish the early response of adult rat aortic VSMC to these two agonists. JE/ MCP-1 protein appears to function chiefly as a monocyte chemoattractant and has not been shown to be a mitogen. It is therefore unlikely that JE/MCP-1 is directly responsible for the difference between PDGF-induced hyperplasia and Ang-induced hypertrophy. Dissection of the rat JE promoter in SMC and identification of the proteins that bind to the PDGF-responsive elements may help elucidate PDGF-specific pathways. This may provide powerful tools for identifying downstream signals that differentiate PDGF-and Ang-induced transcription and may be involved in differentiating the growth responses.