c-Jun Regulates Shear- and Injury-inducible Egr-1 Expression, Vein Graft Stenosis after Autologous End-to-Side Transplantation in Rabbits, and Intimal Hyperplasia in Human Saphenous Veins*

Coronary artery bypass graft failure represents an unsolved problem in interventional cardiology and heart surgery. Late occlusion of autologous saphenous vein bypass grafts is a consequence of neointima formation underpinned by smooth muscle cell (SMC) migration and proliferation. Poor long term patency and the lack of pharmacologic agents that prevent graft failure necessitate effective alternative therapies. Our objective here was to evaluate the effect of targeted inhibition of the bZIP transcription factor c-Jun on intimal hyperplasia in human saphenous veins and vein graft stenosis after autologous end-to-side transplantation. DNAzymes targeting c-Jun attenuated intimal hyperplasia in human saphenous vein explants. Adenovirus-forced c-Jun expression stimulated SMC proliferation, proliferating cell nuclear antigen, and MMP-2 expression. c-Jun DNAzymes abrogated Adeno-c-Jun-inducible SMC growth and wound repair and reduced intimal thickening in jugular veins of New Zealand white rabbits 4 weeks after autologous end-to-side transplantation to carotid arteries. Conversely, in a DNAzyme-free setting, Adeno-c-Jun potentiated neointima formation in the veins compared with Adeno-LacZ. Inducible c-Jun expression is ERK1/2- and JNK-dependent but p38-independent. Injury- and shear-inducible c-Jun controls early growth response-1. These data demonstrate that strategies targeting c-Jun may be useful for the prevention of vein graft stenosis. Control of one important shear-responsive transcription factor by another indicates the existence of transcriptional amplification mechanisms that magnify the vascular response to cell injury or stress through inducible transcriptional networks.

Coronary artery bypass grafting (CABG) 2 remains one of the most widely used surgical means of treating coronary heart disease after heart attack (1,2). Venous graft occlusion is the major cause of recurrent ischemia in patients after CABG. Because of difficul-ties in obtaining arterial graft material and the relative ease of harvesting long vascular segments, the saphenous vein remains the most widely used graft. Fitzgibbon et al. (3) found that patency rates of saphenous vein grafts in patients undergoing CABG 1, 5, and 10 years after surgery were 78%, 65, and 57%, respectively. Although statins improve graft patency by ϳ24 -29% (4), poor long term patency, and the lack of other pharmacologic agents preventing graft failure, necessitates the search for effective alternative therapies and surgical techniques (5). The cellular processes underlying intimal hyperplasia and vein graft failure are smooth muscle cell (SMC) proliferation and migration. Nuclear transcription factors controlling SMC proliferation have become targets of interventional approaches (6). The bZIP protein and prototypic member of the AP1 dimeric family of transcription factors c-Jun is an attractive target for vascular therapeutic intervention. c-Jun is poorly expressed in uninjured arteries but activated by injury (7). Previous studies from our group have shown that catalytic oligodeoxynucleotides (DNAzymes) targeting c-Jun inhibit intimal thickening in carotid arteries following permanent ligation (7) and balloon injury (8). Here we evaluated the effect of targeted inhibition of c-Jun on intimal hyperplasia in human saphenous veins, and of vein graft stenosis after autologous end-to-side transplantation in rabbits. We also explored the role of c-Jun in the initiation of injury-and shear-inducible transcriptional networks.

Culture of Vein Grafts in Vitro and DNAzyme Intervention-
Segments of human saphenous vein ϳ8 -10 cm in length were obtained during bypass surgery from the Cardiothoracic Surgical Unit at Royal Prince Alfred Hospital, Sydney with informed consent. The vein was placed in sterile saline and kept at room temperature; it was promptly transported from the hospital to the University of New South Wales where the vein was cut into small segments, ϳ1 cm in length and cultured in RPMI 1640 (Invitrogen) supplemented with 20% FBS. Veins were randomly allocated to 4 groups: untreated (fresh isolated), vehicle alone, those treated with c-Jun DNAzyme (Dz13), or Dz13scr, in which the hybridization arms were scrambled but containing the same 10 -23 catalytic core as Dz13. Transfections with FuGENE 6 (Roche Applied Science) were performed with either 0.4 mol/liter DNAzyme, or mock (vehicle alone) on Day 0 and Day 7. Veins were grown for 14 days with a daily change of culture medium. On Day 14, veins were placed into 10% forma-lin and fixed at 4°C overnight. Veins were cross-sectioned and stained by Miller's elastin and van Gieson staining.
Construction of Adenovirus-c-Jun-Murine c-Jun cDNA was cloned from mouse mRNA by RT-PCR using primers (forward,  5Ј-ACG GAC TGT TCT ATG ACT GCA A-3Ј; reverse,  5Ј-TCT TCG TTG CCC CTC AGC-3Ј) and ligated into a shuttle plasmid (Qiagen PCR cloning kit). c-Jun cDNA was then transferred into the Ad-Easy expression system (Stratagene) (9), and Adeno-c-Jun was produced after digestion by PacI and amplification in HEK 293 cells.
SMC Culture Conditions, Proliferation, and Injury Assays-SMCs isolated from rat and rabbit aortae were obtained mycoplasma-free at P1 from Cell Applications, Inc. (San Diego, CA). Cells were grown in Waymouth media with 10% fetal bovine serum, 50 g/ml streptomycin, and 50 IU/ml penicillin at 37°C in a humidified atmosphere of 5% CO 2 . In proliferation assays, SMCs in 96-well plates were transfected with Adeno-c-Jun with or without DNAzyme (using FuGENE 6) and arrested for 24 h in medium, excluding serum, then exposed to medium containing 5% FBS at 37°C for 72 h. Cells were harvested by trypsin, and the suspension was counted by automated Coulter counter. In injury assays, SMCs were seeded into 6-well plates, transduced with Adeno-c-Jun and DNAzyme (multiplicity of infection of 10, unless otherwise indicated) in serum-free medium, then scraped using a P1000 pipette tip. After 48 h, the cells were washed with PBS prior to photomicroscopy.
Fluid Shear Stress Model and Immunocytofluorescence-SMCs were seeded on Flexcell slides precoated with 0.1% gelatin and poly-L-Lysine and cultured in Petri-dishes filled with 10% FBS culture media. After reaching 70% confluence, the cells were starved overnight and then transfected with Dz13, Dz13scr, ED5 (DNAzyme targeting early growth response-1), or ED5SCR with 0.4 mol/liter, respectively, following the protocol mentioned above. Meanwhile, the other SMCs were treated with JNKII (JNK pathway inhibitor), PD98059 (MEK/ ERK pathway inhibitor), and SB202190 (p38 pathway inhibitor) at the concentration of 20 mol/liter. After pretreatment with DNAzymes or pharmacologic inhibitors, the cells were transferred into the Streamer Shear Stress Device (Flexcell International, Hillsborough, NC) filled with Dulbecco's modified Eagle's medium/0.5% FBS, and shear stress was applied to the cells with the aid of a computer-controlled peristaltic pump. Immunocytofluorescence staining was performed to detect c-Jun, early growth response-1 (Egr-1), or MMP-2. Briefly, the cells were firstly washed by PBS twice and fixed by 3% formaldehyde for 30 min. The fixed cells were incubated with primary c-Jun (Santa Cruz Biotechnology) and Egr-1 (Cell Signaling) antibody overnight. The secondary fluorescein isothiocyanate (FITC)-conjugated antibody was added to the cells for 1 h. Finally cells were stained with DAPI for nucleus visualization. The results were carefully observed under fluorescence microscope. FITCϩ cell percentages were calculated based on the ratio of positively staining cells and the total cell number indicated by DAPI from at least five randomly chosen fields of view.
RT-PCR, Western Blotting, ELISA, and Zymography-For RT-PCR analysis, total RNA was extracted using the RNeasy Mini Kit (Qiagen). After assessment of RNA concentration, constant amounts of RNA were combined with RT-PCR rea-gent (Qiagen one-step RT-PCR kit) prior to amplification (10,11). Glyceraldehyde-3-phosphate dehydrogenase was used as an internal control. For Western blot analysis, cells were washed twice with PBS and lysed in radioimmune precipitation assay buffer (12,13). Lysates were resolved by SDS-PAGE on 4 -15% gels, and proteins were transferred to polyvinylidene difluoride membranes and blocked for 30 min in blocking buffer (Tris-buffered saline, pH 7.6, 0.05% Tween, and 5% nonfat dry milk). c-Jun (H-79, Santa Cruz Biotechnology) proliferating cell nuclear antigen (PCNA, PC10, Santa Cruz Biotechnology), MMP-2 (4D3, Santa Cruz Biotechnology), and Egr-1 (44D5, Cell Signaling) antibodies were added to the membranes for 1 h, prior to washing and addition of secondary antibody peroxidase conjugate. The membrane was washed and incubated with ECL chemiluminescence substrate (Pierce) prior to exposure. For MMP-2 ELISA, the culture media were harvested and centrifuged, and supernatants were assessed for soluble MMP-2 by commercial ELISA (R&D Systems, Minneapolis, MN). The supernatants containing equal amounts of protein were also run on SDS-PAGE gels, renatured with Triton X-100, and incubated in zymography developing buffer overnight at 37°C (14). Coomassie Blue was used to visualize the bands on the gel.
Rabbit Vein Graft Studies-Adult male New Zealand white rabbits (ϳ3-3.5 kg, n ϭ 10 rabbits per group) on a normal diet were anesthetized with an intramuscular injection of ketamine (25 mg/kg) and xylazine (Ilium Xylazil-20, 5 mg/kg, Parnell Laboratories, Australia). Anesthesia was maintained via mask throughout the procedure with 2% isofluorane mixed with 2% O 2 . Cephazoline (300 mg/kg), an antibiotic, was administered intravenously 15 min prior to incision. A 2-cm segment of the right external jugular vein was carefully exposed after midline incision, flushed, and kept moist in heparinized saline (100 units of heparin/ml, David Bull Laboratories, Australia). Rabbits were heparinized (1000 units/kg), and the right common carotid artery was dissected and clamped at the proximal and distal ends. An incision was made along the side of the artery and flushed with heparinized saline containing 1% lignocaine. The vein was dissected free of surrounding tissue, flushed with 5 ml of heparinized saline, and transferred to a 500-l tube containing 200 l of vehicle solution with 40 l of FuGENE 6 and 5 l of MgCl 2 in sterile PBS, pH 7.4, and the following groups: 1) vehicle only; 2) 500 g of Dz13; and 3) 500 g of Dz13scr. Veins were incubated with DNAzyme or adenovirus (10 10 pfu/ml) at 37°C for 30 min and attached end-to-side to the artery in reversed fashion (proximal end of the vein to the distal part of the artery and vice versa) using 8-0 Proline uninterrupted suture under 3.5ϫ magnification (Zeiss, Germany). The segment of the artery between vein attachments was ligated and cut to ensure unidirectional blood flow. After graft anastomosis, the incision was closed using a 4-0 silk suture (Silkam, Johnson & Johnson). Patency of anastomosis was confirmed by palpable pulsatile blood flow through the vein upon completion of surgery. Fifty ml of normal saline was given intravenously prior to the removal of the cannula. Each rabbit received 2 mg/kg Rimadyl (Pfizer Animal Health) intramuscularly, at the end of surgery for analgesia. Animals were placed on their sides under warming light and given oxygen until complete recovery from anesthesia. To demonstrate DNAzyme transfection in veins, conduits were incubated in 200 l of transfection solution containing 500 g of FITC-Dz13 at 37°C for 30 min. The vein was rinsed in PBS and snap frozen with OCT in liquid nitrogen. Blocks were sectioned and fixed in 95% ethanol. Following DAPI nuclear staining, the slides were imaged under fluorescence microscopy. At 28-day post-transplantation, the proximal and distal carotid segments were cannulated and the vein was perfusion-fixed under distending pressure with 10% formalin in PBS, pH 7.4. Grafts were fixed in 10% formalin (Ambion) overnight and stored at Ϫ80°C until immunohistochemical analysis.
Immunohistochemistry-Antibodies to c-Jun (H-79 and KM-1, Santa Cruz Biotechnology), PCNA (PC10, Santa Cruz Biotechnology), MMP-2 (4D3, Santa Cruz Biotechnology), and alpha-SM actin (M0851, Dako) were used with consecutive paraffin-embedded sections of formalin-fixed tissue. Prior to staining, deparaffinized sections were treated with 3% hydrogen peroxide (peroxidase blocking) and boiled in citrate buffer, pH 6.0, to retrieve antigenicity. After washing in PBS, sections were incubated with primary antibody for 60 min, the appropriate secondary antibody (horse anti-mouse, vector BA-2000 or goat anti-rabbit, vector BA-1000) for 20 min, and finally with avidinbiotin complex (Elite Vector PK-6100) for 30 min. Immunogenicity was visualized by treatment in 3,3Ј-diaminobenzidine solution for 2 min, which produced brown coloration. Sections were counterstained with Mayer's hematoxylin and Scot blue. As negative controls, the primary antibody was omitted, or the sections were treated with the immunoglobulin fraction of suitable non-immune serum as a substitute for the primary antibody.
Statistical Analysis-A test for unequal variance was performed on the data with a two-sample F-test (i.e. no assumption). Student's t test was used to statistically evaluate data from in vitro experiments. In vivo experiments were analyzed by one-way analysis of variance followed by Bonferroni/Dunn post hoc analysis. Standard errors of the mean are shown (*, p Ͻ 0.05; ϩ , p Ͻ 0.01). Animal experiments were approved by the University of New South Wales Animal Care and Ethics Committee.

c-Jun DNAzymes Attenuate Intimal Hyperplasia in
Human Saphenous Veins-Saphenous vein graft failure after CABG is strongly associated with SMC proliferation and migration (15). To determine the effect of c-Jun knockdown on intimal hyperplasia in human vein grafts, we incubated freshly isolated human saphenous vein explants with Dz13, a catalytic oligodeoxynucleotide (DNAzyme) targeting c-Jun (16). In this explant model, the intima of the vessel thickens  spontaneously after several weeks in culture (9). Dz13 attenuated neointima formation within 2 weeks compared with the vehicle control (Fig. 1, A and B). Veins treated with Dz13scr, a DNAzyme containing an intact catalytic domain in which the RNA-hybridizing arms are scrambled, had no effect (Fig. 1, A and B). Immunohistochemical analysis revealed that Dz13, but not Dz13scr, suppressed both c-Jun and PCNA expression in these vessels, which stained positive for alpha-SM actin, suggesting c-Jun may be associated with intimal hyperplasia in human saphenous veins (Fig. 1C).

SMC Proliferation Is Increased after Transduction with Adeno-c-Jun-
We next determined the effect of overexpressing, rather than knocking down, c-Jun on SMC growth. c-Jun overexpression was achieved using the replication-defective adenovirus Ad-Easy system. c-Jun protein levels were barely detectable in growth-quiescent rat aortic SMCs ( Fig. 2A), consistent with poor c-Jun expression in uninjured carotid arteries (7). Within 24 h, however, Adeno-c-Jun transduction increased c-Jun expression in SMCs without affecting levels of ␤-actin ( Fig. 2A). c-Jun overexpression also induced PCNA expression ( Fig. 2A), complementing our findings with Dz13 in Fig. 1C. We assessed levels of MMP-2 in transduced SMC cells. Exogenous c-Jun stimulated MMP-2 expression within 24 h (Fig. 2A). SMCs transduced with adenovirus-c-Jun also underwent accelerated proliferation within 3 to 4 days compared with non-transduced cells, although the effect was not dosedependent at the multiplicity of infection examined (Fig. 2B).
DNAzyme Targeting c-Jun Suppresses Adeno-c-Jun-inducible SMC Gene Expression and Proliferation-To determine whether the mitogenic effect of c-Jun overexpression could be reversed with c-Jun DNAzyme, we transfected Adeno-c-Jun-transduced SMCs with Dz13 or Dz13scr and performed mRNA and protein expression analyses, or proliferation assays. Adeno-c-Jun increased levels of c-Jun mRNA (Fig.  3A) and protein (Fig. 3B) expression by RT-PCR and Western blotting after 24 h, respectively, in the absence of DNAzyme. The induction in c-Jun expression was virtually abolished by Dz13 (Fig.  3, A and B). The DNAzyme also inhibited MMP-2 and PCNA mRNA and protein expression (Fig. 3, A and B). Because MMP-2 is a secreted metalloproteinase, ELISA and zymographic analyses, performed with supernatants of SMCs transduced with adeno-c-Jun and transfected with Dz13, revealed that the DNAzyme reduced MMP-2 levels to basal levels (Fig. 3C). Dz13scr, by contrast, had no effect on MMP-2 mRNA or protein expression, or proteolytic activity (Fig. 3, A-C). Dz13, but not Dz13scr, also inhibited SMC growth that was potentiated by Adeno-c-Jun, regardless of whether the SMCs were of rat (Fig. 3D) or rabbit (Fig. 3E) origin.

DNAzyme Targeting c-Jun Suppresses Adeno-c-Jun-inducible SMC Wound
Repair-A single linear scrape of SMC monolayers in culture initiates a reparative response involving proliferation and migration from wound edge and regrowth in the denuded zone. We hypothesized that Adeno-c-Jun would accelerate this response to injury and that Dz13 would inhibit the effect of exogenous c-Jun. Adeno-c-Jun indeed stimulated this reparative response to injury (Fig. 4, A and B). Moreover, Dz13 inhibited this increase, whereas the Dz13scr had no effect (Fig. 4, A and B).

c-Jun DNAzymes Inhibit Vein Graft Stenosis after Autologous End-to-Side Transplantation in Rabbits-
The preceding findings using adenovirus to increase SMC levels of

. Dz13 inhibits vein graft stenosis after autologous end-to-side transplantation in rabbits.
A, jugular veins were extracted and transfected with FITC-labeled Dz13. Tissue was frozen in OCT, cross-sectioned, DAPI-stained, and viewed under fluorescence microscopy. B, veins were treated intraoperatively with 500 g of DNAzyme or vehicle, and sutured end-to-side into the carotid arteries of the same rabbits. After 28 days, vein grafts were harvested, fixed, and embedded in paraffin. Sections were stained by modified Miller's elastic and van Gieson staining and photographed under 200ϫ magnification. Intimal thickening is expressed as the neointimal area as a percentage of total vessel area. C, representative cross-sections of grafts at 28 days or of freshly isolated veins. D, immunohistochemical analysis for c-Jun, PCNA, and MMP-2 in the respective groups (magnification, 400ϫ). E, immunohistochemical analysis for JunB and c-Fos in Dz13-and Dz13scr-treated veins (magnification, 400ϫ). F, jugular vein segments were transduced ex vivo with adenovirus-c-Jun (10 10 pfu/ml) over 30 min, prior to end-to-side transplantation into carotid arteries, and assessment of intimal thickening after 28 days. Immunohistochemical analysis was performed for c-Jun, PCNA, and MMP-2 (magnification, 400ϫ). c-Jun, and Dz13 to knockdown c-Jun, demonstrate that SMC growth and wound repair in vitro are tightly controlled by this transcription factor. Dz13 has previously been used in animal models of arterial injury (7,8), but it has not been deployed in models of autologous vein transplantation. Using the E2F decoy approach of intra-operative, ex vivo oligonucleotide delivery, we incubated rabbit jugular veins in a solution containing Dz13 or Dz13scr then grafting the vein into the carotid artery of the same animal end to side, as is performed in standard CABG. Intra-operative, ex vivo delivery of Dz13 was indicated by green fluorescence visualized in the graft following delivery of FITC-labeled Dz13 (Fig. 5A) in vehicle containing the commercial transfection agent FuGENE 6. This model results in a thickened neointima within the graft 4 weeks after anastomosis (Fig. 5, B and C). Dz13 attenuated intimal thickening (Fig. 5, B and  C), whereas neointima formation after treatment of the veins with Dz13scr did not differ from the vehicle control (Fig. 5, B  and C). Immunohistochemical analysis of cross-sections revealed that c-Jun, PCNA, and MMP-2 were poorly expressed in freshly isolated jugular veins, but increased in the grafts after 4 weeks (Fig. 5D), coincident with increased intimal thickening (Fig. 5, B and C). c-Jun, PCNA, and MMP-2 expression were attenuated by Dz13, but not Dz13scr (Fig. 5D). In efforts to demonstrate that the effects of the DNAzyme are specific for c-Jun, we performed immunostaining of DNAzyme-treated veins with antibodies to other key members of the AP-1 family, JunB and c-Fos (17). Although Dz13 blocked c-Jun expression (Fig.  5D), it had no effect on levels of JunB, nor did it influence levels of c-Fos (Fig. 5E).

Adeno-c-Jun Potentiates Vein Graft Stenosis after Autologous Endto-Side Transplantation in Rabbits-Dz13
was used throughout this study as an interventional agent targeting endogenous c-Jun. Accordingly, Dz13's inhibition of c-Jun expression (Figs. 1D and  5D), c-Jun-dependent gene expression (Figs. 3A, 3B, 3C, and 5D), cell growth (Fig. 3, D and E), cell migration (Figs. 4A and B), and intimal thickening (Figs. 1B,  1C, 4A, 4B, and 4C) demonstrate that endogenous c-Jun is causally involved in all these processes. Throughout this study, Dz13scr, the scrambled arm counterpart of Dz13, provided an important sizeand overall charge-matched, sequence-specific negative control. To provide further evidence for the causal role of c-Jun in vein graft neointima formation, we performed further experiments in rabbit vessels, independently of Dz13. We transduced jugular vein segments ex vivo with adenovirus-c-Jun (10 10 pfu/ml) over 30 min, prior to end-to-side transplantation into carotid arteries (as used for Dz13/ Dz13scr in Fig. 5, B and C) and assessed the extent of intimal thickening after 28 days. Fig. 5F demonstrates that Adenoc-Jun significantly potentiates neointima formation in the vessels compared with its control (Adeno-LacZ). Immunohistochemical staining of the rabbit tissue provides supportive data. Adeno-c-Jun increased c-Jun immunoreactivity in the veins (Fig. 5F, middle upper), together with the expression of the c-Jun-dependent gene, MMP-2 (Fig. 5F, middle lower), and the mitogenic marker, PCNA (Fig. 5F, middle  center). These in vivo data with Adeno-c-Jun complement our findings in the same model with Dz13 and provide compelling evidence for c-Jun playing a causal role in vein graft thickening.

JOURNAL OF BIOLOGICAL CHEMISTRY 4043
Induction of c-Jun in SMCs Exposed to Fluid Shear Stress Is ERK1/2-and JNK-dependent and p38-independent-The process of intimal thickening as a prelude to bypass graft ste-nosis may be initiated by changes in fluid biomechanical forces sensed by the cells in the graft (18). For example, conduits undergo an acute static-to-shear transition after the reperfu-  (magnification, 200ϫ). B, growth-quiescent SMCs, transfected with Dz13, Dz13scr, or incubated with PD98059, SB202190, or JNKII, were exposed to 10 dynes/cm 2 fluid shear stress for 1 h and fixed, and immunocytochemical analysis was performed. c-Jun expression (FITC ϩ ) was visualized under fluorescence microscopy. Blue, DAPI-Blue staining indicates nuclei. C, SMCs were treated as in B, except that the cells remained under static conditions for the 1 h. D, SMCs were exposed to shear for the indicated times, and immunocytochemical analysis was performed for c-Jun or Egr-1 on fixed cells. FITC ϩ staining was expressed as a proportion of FITC ϩ -and DAPI-Blue-stained nuclei in five random fields of view. E, effect of Dz13, Dz13scr, ED5, ED5 SCR, PD98059, SB202190, or JNKII on shear-inducible c-Jun and Egr-1 expression after 90-min exposure to fluid shear stress. FITC ϩ staining was expressed as a proportion of FITC ϩ -and DAPI-Bluestained nuclei in five random fields of view. F, Western blot analysis for Egr-1 or c-Jun using total cell extracts of SMCs (pretreated with Dz13, Dz13scr, ED5, or ED5SCR), injured by scraping and left for 1 h prior to harvest. In G: Left/upper panels, immunohistochemical analysis for Egr-1 in rabbit venoarterial autologous bypass transplants 28 days after the veins were treated ex vivo with Dz13, Dz13scr, and anastomosis. Right/lower panels, Egr-1 staining on vein transplants 28 days after the veins were transduced with Adeno-c-Jun or Adeno-LacZ.

c-Jun, Egr-1, and Vein Graft Stenosis
sion. Moreover, because arterial shear forces are considerably higher than those in veins, venous conduits experience biomechanical forces not previously sensed by veins. SMCs represent the most prominent cellular element of the intimal lesion (19), and immunohistochemical analysis by our group (Fig. 6A) and others (20,21) indicates that SMCs, rather than endothelium, can make contact with flowing blood in human vein grafts. It is possible therefore that the SMC response to flow may influence shear-dependent thickening in the de-endothelialized vein graft. Immunocytofluorescence analysis revealed that c-Jun is poorly expressed in cultured growth-quiescent SMCs but is induced by laminar shear stress (10 dynes/cm 2 ) within 1 h (Fig. 6, B and C). The induction of c-Jun by shear stress was completely abrogated by Dz13 (Fig. 6B), whereas Dz13scr had no effect (Fig. 6B). To determine kinase(s) mediating shearinducible c-Jun expression, we exposed SMCs to pharmacologic inhibitors of MEK-ERK1/2 (PD98059), JNK (JNKII inhibitor), or p38 (SB202190) prior to exposure to shear. Fig. 6B clearly demonstrates that the induction of c-Jun by shear is ERK1/2-and JNK-dependent and is p38-independent.
Shear-and Injury-inducible Egr-1 Expression Is Controlled by c-Jun-To provide additional insights on the molecular pathways regulated by c-Jun in stressed SMCs. Accordingly, we performed microarray analysis to identify differently expressed c-Jun-dependent genes in SMCs 24 h after transduction with Adeno-c-Jun and Adeno-LacZ. Using 44K Agilent GeneChip Arrays, we found that c-Jun increased the expression of 607 genes, and reduced that of 180 genes by at least 2-fold after this time. Among the genes induced by c-Jun was the zinc finger transcription factor, Egr-1, which increased 6-fold. Egr-1 is an immediate-early gene, which, like c-Jun, plays a master regulator role by switching on the expression of many pathophysiologically important genes in numerous models of cardiovascular pathology (22). We demonstrated over a decade ago that Egr-1 is induced by shear in vascular endothelial cells (23). However, the transcriptional pathway responsible for its induction by shear is unknown, as is whether Egr-1 is activated in other cell types by shear stress. The time course of c-Jun and Egr-1 induction by shear stress revealed that the peak of c-Jun expression precedes that of Egr-1 (Fig. 6D). Intriguingly, shear-induced Egr-1 expression was inhibited by Dz13 (Fig. 6E). In contrast, shear-induced c-Jun expression was not inhibited by ED5, a DNAzyme that targets Egr-1 (Fig. 6E).
These findings were not confined to shear stress. We found that injury-induced Egr-1 expression was inhibited by Dz13 (Fig. 6F). In contrast, injury-induced c-Jun expression was not inhibited by ED5 (Fig. 6F). We further determined that Dz13 suppressed Egr-1 expression in rabbit venoarterial autologous bypass transplants (Fig. 6G), in which intimal thickening was inhibited by Dz13 (Fig. 5, B and C). Moreover, Egr-1 levels were increased in veins transduced with Adeno-c-Jun (Fig. 6G). These studies thus demonstrate that inducible c-Jun expression is ERK1/2-and JNK-dependent and p38-independent and that c-Jun controls Egr-1.

DISCUSSION
CABG disease represents an unsolved problem in interventional cardiology and heart surgery. Late occlusion of autologous saphenous vein bypass grafts is a consequence of neointima formation underpinned by SMC migration and proliferation. The beneficial effects of 3-hydroxy-3-methylglutaryl-CoA reductase inhibitors on saphenous vein bypass grafts independent of their cholesterol-lowering properties have been attributed to their ability to reduce intimal hyperplasia by suppressing SMC invasion and proliferation (24). The present study offers several novel advances over prior work. First, the present study provides the first ever demonstration that Dz13 inhibits neointima formation in venous tissue. All previous models involving Dz13 have only used arteries, and findings from arteries cannot necessarily be extrapolated to veins (25). Second, the end-to-side venoarterial autologous bypass transplantation model used here in rabbits is analogous to the anastomosis commonly performed in human CABG. Agents that inhibit neointima formation in rat vessels do not necessarily perturb intimal thickening in higher species or humans. Table 1 illustrates that drugs such as heparin, carve- This study demonstrates that Dz13 can inhibit restenosis in human (Fig. 1, A and B) and rabbit (Fig. 5, B and C) venous tissue. c-Jun, is an attractive target for interventional strategies aimed at improving the long term survival of venous grafts, for several reasons. First, this study demonstrates that c-Jun is expressed with PCNA in SMCs of failed human saphenous veins used for CABG. These findings are supported by independent studies that have shown that c-Jun mRNA is up-regulated in stenotic saphenous aortocoronary bypass grafts by microarray analysis (26). In contrast, c-Jun is poorly expressed in normal arteries and veins. Second, as shown here, c-Jun overexpression, by adenovirus, accelerates SMC proliferation. It also stimulates SMC wound repair after scraping injury. Moreover, Dz13 inhibits SMC proliferation and migration potentiated by Adeno-c-Jun. Third, Dz13 inhibits intimal thickening in a rabbit end-to-side vein graft model and builds on our earlier demonstration of c-Jun DNAzyme suppression of neointima formation in rat (7) and rabbit (8) arteries. Finally, we demonstrate here that shearand injury-inducible Egr-1 expression is suppressed by the c-Jun DNAzyme. DNAzymes to Egr-1, however, had no inhibitory effect on c-Jun levels.
The mechanism with which c-Jun controls intimal hyperplasia and SMC growth and migration is likely to involve Egr-1. Egr-1 is another key transcription factor strongly implicated in vascular disorders, that like c-Jun, is poorly expressed in normal arteries but dramatically induced after acute vascular injury (22,27). c-Jun overexpression stimulates Egr-1 expression in the neointima (Fig. 6G), whereas Dz13-targeting c-Jun prevents Egr-1 induction (Fig. 6, E-G).
Although we have not demonstrated here whether Egr-1 inhibition and overexpression affects intimal hyperplasia and SMC growth in the same way as c-Jun, we have previously shown that Egr-1 suppression in carotid and coronary arteries reduces neointima formation in models of permanent ligation (28), balloon injury (29), and in-stent restenosis (30). We showed that Dz13 inhibits MMP-2 expression (Figs. 3A, 3B, 3C, and 5D). Interestingly, MMP-2 is under the transcriptional control of both c-Jun (14,31) and Egr-1 (32), suggesting that these and other nuclear factors control the same genes in a coordinated manner. Control of one important injury-and shear-responsive transcription factor by another suggests the existence of transcriptional amplification waves that magnify the signaling burst after acute cellular stress.
MMPs and SMC hyperplasia have long been implicated in the pathogenesis of intimal hyperplasia and CABG failure. George and colleagues reported that TIMP-2 overexpression using an adenovirus reduces neointimal thickening, primarily by inhibiting MMP activity and blocking SMC migration (33). This group later showed that adenovirus-mediated TIMP-3 overexpression inhibits late vein graft neointima formation in human and porcine models (34). Others have suggested that surgical preparative injury initiates phenotypic changes in a subpopulation of medial SMCs to a synthetic phenotype increasing MMP activity thereby causing matrix degradation, SMC migration, intimal thickening, and graft failure (35). Small interference RNA-based silencing of either MMP-2 or MMP-9 inhibits invasion of human saphenous vein SMC (36). Moreover, Berceli et al. found that MMP-2 is significantly up-regulated in grafts with enhanced intimal thickening and is the predominant gelatinase regulating early vein graft remodeling (37).
Localized "anti-gene" therapy, rather than classic gene therapy, in the surgically isolated vessel at the time of harvest ex vivo provides an alternative and potentially safer means to treat transplantable grafts. This approach has distinct strategic advantages for the inhibition of graft failure for several reasons. First, conduits are generally free of disease at the time of bypass and are, therefore, amenable to genetic manipulation. Second, manipulation and irrigation of the vessel is already a routine part of the surgical procedure. Third, ex vivo treatment of the graft would reduce the propensity of systemic side effects. Finally, SMC hyperplasia in graft failure, like restenosis after percutaneous transluminal coronary angioplasty or stenting, occurs early after injury, which suits acute single administration (38). The safety and feasibility of treating human vein grafts ex vivo with oligo- nucleotides was demonstrated in the PREVENT trials using edifoligide, a double-stranded synthetic phosphorothioated oligonucleotide bearing the E2F binding site first described by Dzau et al. in 1995 (39). PREVENT provides important proof-of-principle clinical evidence of the practicability of gene-manipulative treatment of grafts ex vivo prior to implantation. It also suggests the likelihood of even greater efficacy using more potent inhibitors and approaches to established or new targets. Recent studies have shown that topical pretreatment of vein grafts with the ERK1/2 inhibitor UO126 effectively reduces neointimal formation after vein graft arterialization (40). Local delivery of c-Jun-targeting agents, alone, or in combination with other strategies, provides a new opportunity to improve vein graft patency following coronary artery or peripheral vascular bypass surgery.