Prostacyclin synthase gene transfer inhibits neointimal formation in rat balloon-injured arteries without bleeding complications

Objective: This study was designed to compare the effects of prostacyclin synthase (PCS) gene transfer with those of a systemic infusion of beraprost sodium (BPS), a prostacyclin analogue, on vascular smooth muscle cell (VSMC) proliferation and neointimal formation after arterial injury. Methods: PCS gene (3 or 30 m g) was transfected into rat balloon-injured carotid arteries by a non-viral lipotransfection method. BPS (100 or 300 m g/kg/day) was subcutaneously infused with osmotic pumps after the injury. LacZ gene (30 m g) was used as a control.VSMC proliferation was estimated by the bromodeoxyuridine (BrdU) index (BrdU-positive nuclei/total nuclei) at day 7. Neointimal formation was evaluated at day 14. Each treatment group had six rats. Results: PCS gene transfer prevented the increase in intimal/medial area ratio (3 m g: 46.6%, 30 m g: 61.1% reduction; P , 0.05, P , 0.01, respectively), as did BPS 300 m g/kg/day (49.8% reduction; P , 0.05). BPS 100 m g/kg/day, however, had no effects on the ratio. PCS gene transfer and BPS 300 m g/kg/day signiﬁcantly suppressed the BrdU index. BPS 300 m g/kg/day group had more frequent hematoma and longer bleeding time. There were no signiﬁcant differences in blood pressure, heart rate, or urinary volume among all groups. Conclusion: Both PCS gene transfer and BPS 300 m g/kg/day reduced neointimal formation after arterial injury by inhibiting VSMC proliferation. PCS gene transfer may be a safer therapeutic modality against neointimal formation than a systemic infusion of BPS because the former method resulted in fewer bleeding complications. (cid:211) 1999 Elsevier Science B.V. rights reserved.


Introduction
demonstrated that coronary restenosis resulted primarily from vascular remodeling and contraction rather than from Percutaneous transluminal coronary angioplasty (PTCA) intimal growth [7,8]. Although several randomized and is a well-established strategy against severe and symptom-placebo-controlled studies have revealed that coronary atic coronary artery diseases. However, its efficacy is stents which generally prevented vascular remodeling limited by a high incidence of chronic restenosis which has reduced restenosis rates after angioplasty [9,10], the issue been reported to be as high as 20-50% within 3-6 months of restenosis mainly caused by intimal thickening after [1][2][3]. Several lines of evidence indicate that this re-coronary intervention has not been resolved. Combined stenosis might be multifactorial and caused mainly by therapy against vascular remodeling and neointimal formavascular remodeling, the migration of vascular smooth tion may be required to eliminate restenosis after coronary muscle cells (VSMCs) from the media into the intima, and intervention. In animal models, several pharmacological their unregulated proliferation [4][5][6]. Intravascular ultra- [11][12][13] and genetic [14][15][16]  protects vessels by dilating the vessels [17], suppressing a 1.5% agarose gel, stained with ethidium bromide and platelet aggregation [18] and inhibiting VSMC prolifer-extracted. The 1.6 kb rat PCS cDNA was cloned in the ation [19]. Although several animal studies have shown cytomegalovirus (CMV) enhancer / promoter-directed vecthat a PGI analogue suppressed experimental neointimal tor (pTargeT, Promega), which contains a neomycin 2 hyperplasia [20, 21], restenosis occurred after coronary resistance gene under the control of an SV 40 enhancer / angioplasty with a PGI analogue in a clinical trial early promoter with a polyadenylation signal sequence 2 conducted by Gershlick et al. [22]. However, they em-(pCMV-PCS). The plasmid carrying the LacZ gene substiployed a relatively small dose (4 ng / kg / min) of the tuted for the PCS gene was constructed as a control analogue, which was infused for a relatively short period (pCMV-LacZ). The plasmid DNAs (pCMV-PCS and (36 h). This may have led to the failure of the analogue to pCMV-LacZ) were purified with a plasmid purification kit prevent restenosis. Thus, the effects of PGI on VSMC (Qiagen) according to the manufacturer's protocol. The 2 migration and proliferation after balloon injury have not sequence of plasmids was confirmed by a DNA sequence yet been clarified. We speculated that one of the reasons for reader (4000L, LI-COR Tween-20), and incubated with anti-rat PCS antibody expression, shown above as the underlined region. The (1:500 in TBST, Cayman) for 1 h at room temperature. non-underlined sequences are based on the original se- The membranes were then washed three times with TBST, quence of rat PCS). Total RNA (5 mg) from rat aorta was incubated with anti-rabbit immunoglobulin antibody reverse-transcribed using a first-strand cDNA synthesis kit (1:5000), washed again three times with TBST, developed (Pharmacia Biotech), containing Moloney Murine with enhanced chemiluminescence reagent (ECL Western Leukemia Virus reverse transcriptase. In each tube, 0.5 mg Blotting Analysis System, Amersham) and exposed to of oligo(dT)12-18 primer (Invitrogen) was used as a X-ray film. primer for the first-strand cDNA synthesis. A polymerase chain reaction (PCR) was performed for 40 cycles of 2.4. Transgene efficiency in vitro denaturation at 958C for 30 s, annealing at 508C for 30 s, and extension at 728C for 1 min with a final 10-min Transfected VSMCs were detected by LacZ gene exextension period. The PCR product was electrophoresed on pression using b-galactosidase staining [5 mmol / l K Fe(CN) , 5 mmol / l K Fe(CN) , 1 mg / ml 5-bromo-4-cedures among all groups. Rats were divided into five fection (BPS-100), and (5) BPS 300 mg / kg / day without Therefore, cells with the blue cytosol but without dark blue transfection (BPS-300) (n56 each, Table 1). nuclei were not counted as transducted cells. pCMV-LacZ, the carotid arteries were perfusion-fixed with The solution was washed with water-saturated diethylether 2.5% phosphate-buffered glutaraldehyde and harvested three times and lyophilized. The cAMP levels in an under anesthesia with sodium pentobarbital. The LacZ appropriate dilution were measured by an enzyme imgene expression was determined by b-galactosidase stainmunoassay kit (Amersham) according to the manufacturing for 24 h, and counterstaining with eosin. The ber's protocol.
galactosidase positive cells were defined as described above. The transduction efficiency was calculated by the 2.6. Gene delivery in vivo and regimen of BPS infusion following formula: (positive cells stained by b-galactosidase staining) /(total nuclei stained by eosin) (n53). This investigation conformed with the Guide for the Care and Use of Laboratory Animals published by the US 2.8. PGI and TXA production in rat carotid artery 2 2 National Institutes of Health (NIH Publication No. 85-23, revised 1996). The in vivo gene transfer into rat carotid The injured rat carotid arteries treated with pCMV-PCS arteries was performed as previously described [24]. Male (30 mg) or pCMV-LacZ (30 mg) were resected 14 days Sprague-Dawley rats (Japan SLC, Inc.) weighing 360-410 after the balloon injury. The arteries were cut into 5-mm g were maintained under a 12-h light / dark cycle at 238C lengths, washed with 9.57 mmol / l PBS (pH 7.4), and and were fed standard laboratory chow (RC4, Oriental incubated in 1 ml of 50 mmol / l Tris-HCl at 378C for 45 Yeast Co.) and water ad libitum. Rats were anesthetized min. Six-keto-PGF and TXB in the medium were 1a 2 with sodium pentobarbital (50 mg / kg intraperitoneally, measured to evaluate local PGI and TXA production, 2 2 i.p.). After a midcervical incision, the right common respectively, with the enzyme immunoassay kits described carotid artery and its bifurcation were exposed. To prevent above. acute thrombosis during the procedure, heparin sodium (200 IU / kg) was intravenously injected 5 min before the balloon injury according to the method of Zoldhelyi et al.
2.9. Measurement of hemodynamic parameters and [25]. The right common carotid artery was balloon-injured urinary excretion of prostanoids three times with a 2 F balloon catheter (2 F Fogarty, Baxter Healthcare) inserted through the external carotid Systolic blood pressure, heart rate and urinary volume artery. Two vascular clips were placed at the distal end and were measured before and 14 days after balloon injury. in the middle of the injured arterial segment. A 24-gauge Systolic blood pressure and heart rate were measured by cannula was introduced into the common carotid artery via the tail-cuff method (BP98A, Softron). Urine was collected the external carotid artery. The lumen of the injured every 2 h to avoid evaporation and centrifuged at 10003g segment between the two clips (10 mm long) was washed for 5 min to eliminate particles. The urinary level of with 9.57 mmol / l phosphate-buffered saline (PBS).
6-keto-PGF was measured with the enzyme immuno-1a pCMV-PCS or pCMV-LacZ with 100 ml Lipofectamine assay kit described above. Plus (GIBCO BRL) or vehicle (9.57 mmol / l PBS) with 100 ml Lipofectamine Plus was instilled into the lumen for by the Dunn procedure was performed to analyze non-The abdomen was shaved carefully and an incision (10 parametric data. Differences in noncontinuous variables mm in length and 3 mm in depth) was made with a scalpel.
were tested with the chi-square test. A value of P,0.05 Bleeding from the incision site was wiped away with filter was considered significant. paper every 20 s, until it stopped. The template bleeding time was defined as the time elapsed from the initiation of the incision to hemostasis. The frequency of subcutaneous hematoma at the site of surgery was also determined on 3. Results day 14. Hematoma was defined as a mass which was more than 10 mm in diameter, and was confirmed macro-3.1. Prostacyclin production induced by pCMV-PCS scopically.
transfer in cultured rat VSMCs

Quantification of neointimal formation
The baseline levels of both 6-keto-PGF and TXB 1a 2 produced by the pCMV-PCS-and pCMV-LacZ-transfected Fourteen days after the vascular injury, rats were VSMCs did not differ significantly from those of the anesthetized with sodium pentobarbital (50 mg / kg i.p.). untransfected control VSMCs (Table 2). Treatment with After a midsternal incision was made, a blood sample was 10 mmol / l sodium arachidonate for 60 min increased the collected by heart puncture, and the common carotid production of 6-keto-PGF by 2.5-fold in the untransfect-1a arteries were perfusion-fixed with 2.5% phosphate-buf-ed VSMCs, by 3-fold in the pCMV-LacZ-transfected fered glutaraldehyde and harvested. The arteries were VSMCs, and by 6-fold in the pCMV-PCS-transfected stained with hematoxylin-eosin or elastica van Gieson VSMCs (Table 2). Thus, the pCMV-PCS-transfected stain. Cross-sectional areas of the lumen, intima, media VSMCs produced twice as much PGI in the presence of 2 and area surrounded by external elastic lamina were exogenous arachidonic acid as in the pCMV-LacZ-transdetermined with computed planimetry on four sections fected VSMCs and the untransfected VSMCs, which was a from each artery that spanned the 1.0-cm region of arterial significant difference (P,0.05, each). In contrast, the injury and were averaged.
TXB production did not change even in the presence of 2 exogenous arachidonic acid in each VSMC group.
The baseline levels of urinary 6-keto-PGF were not 1a significantly different among the five groups. Fourteen days after balloon injury, the urinary levels of 6-keto-PGF increased by 1.7-fold in BPS-100 group (P,0.05), nificantly change (Fig. 5). Thus, pCMV-PCS locally transgene efficiency in rat VSMCs confirmed by b-galactodelivered to the carotid artery did not affect the level of sidase staining was 21.561.7% (Fig. 2). plasma 6-keto-PGF . 1a The platelet counts measured at both day 0 and day 14 were not significantly different among the five groups 3.

Effects of gene transfer on the injured arterial wall
( Table 4). The template bleeding times at day 0 were not different among the five groups. The bleeding time in the The cells stained by b-galactosidase were scattered BPS-300 group was significantly prolonged at day 14 (day mainly in the media, partly in the adventitia (Fig. 3). The 14: 18569 vs. day 0: 14567 s, P,0.05), while those of transduction rate was 8.560.7%. the other groups were not (Table 4). In addition, the The pCMV-PCS-transfected segments of balloon-injured BPS-300 group had significantly more frequent hematoma carotid arteries produced significantly more prostacyclin at the operated neck or the back where the osmotic pumps estimated by 6-keto-PGF than their pCMV-LacZ-trans-1a had been implanted (control: 0 / 6 vs. BPS-300: 3 / 6, P, fected counterparts (6.960.7 vs. 2.160.3 mg / ml, P, 0.05). In contrast, the pCMV-PCS groups showed neither 0.0001) (Fig. 4). The TXB production was not sig-2 bleeding tendency nor hemorrhagic complications. nificantly different between the two segments (Fig. 4). Thus, balloon-injured carotid arteries transfected with PCS gene were able to produce a larger amount of PGI without 2

The effects of PCS gene transfer and BPS infusion
any changes in their ability to produce TXA . 2 Fig. 6 shows representative histological photomicrography of common carotid arteries 14 days after the injury. Systolic blood pressure, heart rate and urinary volume Table 5 shows the morphometric analyses of these arteries.      Both the local expression of PCS and the high dose of BPS surrounded by the external elastic lamina. Larger lumen (300 mg / kg / day) markedly suppressed neointimal forma-areas were obtained in the pCMV-PCS-30 and the BPStion and the intimal / medial area ratios, but did not affect 300 groups. A lower dose of BPS (100 mg / kg / day), the medial area and the vessel size expressed by the area however, had no inhibitory effect on neointimal formation.   The BrdU index values are shown in Table 6. The application than a systemic administration of BPS in light pCMV-PCS-30 transfer and systemic infusion of BPS 300 of attenuating the systemic side effects of PGI . 2 mg / kg / day reduced the BrdU index not only in the In the present study we used cationic liposomes to neointima but also in the media to a greater extent than the enhance the entry of plasmid DNA into cells. It has been pCMV-LacZ transfer (control).

Systemic effects of gene transfer and BPS infusion
reported that the transduction efficiency of liposomal vectors into vascular cells (4-5%) is lower than that of the other vectors [32]. Adenoviral vectors have been most frequently used for vascular gene transfer [14,25,33,34].

Discussion
Although the adenovirus-mediated gene transfer method achieves high transgene expression in vivo (10-30%), it In the present study we demonstrated that a local has several limitations. For example, adenovirus vectors overexpression of PGI derived from PCS gene transfer retain the majority of the parent virus genome, which is 2 and a systemic infusion of BPS prevent VSMC migration associated with undesired gene expression that results in and proliferation (neointimal formation) in balloon-injured both immune and vascular inflammatory responses [35,36]. rat carotid arteries. We also found that a high dose of BPS Gene expression from adenoviral vectors in vivo peaks causes more frequent hemorrhagic complications.
within 1 week and is limited to 2-4 weeks due to these Many early responses to balloon injury, including immune responses. In contrast, the liposomal vector methplatelet adhesion and aggregation, thrombus formation, and od has several advantages. It is not antigenic, is feasible to adhesion and invasion of macrophages and lymphocytes to prepare, and can be used repeatedly. Our liposomal vector arterial walls may take place, leading to the synthesis had an acceptable transgene efficiency of 8.5% in vivo. and / or release of various kinds of growth factors such as Our present findings showed that PCS gene transfer with platelet-derived growth factor-A, transforming growth a liposomal vector markedly suppressed neointimal formafactor-b and basic fibroblast growth factor [26]. These tion after balloon injury. This desirable result may have growth factors in turn stimulate VSMCs to migrate from been due to the following possible mechanisms. First, PGI 2 the media into the intima, where they start to proliferate was overexpressed at the balloon-injured arterial segments and secrete extracellular matrix components [27,28]. PGI on which it should act in order to prevent intimal thicken-2 has been reported to inhibit the chemotaxis and adhesion ing. The PGI thus produced may be advantageous because 2 of monocytes [29], cytokines derived from activated it prevents the proliferation of untransfected cells as well macrophages, and VSMC migration and proliferation [30].
as transfected cells through an autocrine / paracrine loop Moreover, it is likely that PGI suppresses some of these [24]. Second, PGI has many protective effects on vessels 2 2 reactions to balloon injury. Although each protective effect as well as a direct inhibitory effect on cellular growth [30], of PGI may not be strong enough to inhibit VSMC similar to the other vasodilative molecules such as C-type 2 proliferation, these effects may be combined and result in a natriuretic peptide [24] [31]. Our present study also demonstrated that the high of PGH may cause vascular contraction and platelet 2 dose (300 mg / kg / day) of BPS caused more frequent aggregation [38][39][40], and may also result in an overhemorrhagic complications associated with increased production of TXA . Pritchard et al. [41] reported that 2 levels of plasma and urinary 6-keto-PGF , although it did PGH production increased after balloon injury via the 1a 2 prevent neointimal formation after balloon injury. In induction of COX-2 in the vessel wall, although the injury contrast, local PCS gene transfer to the injured arterial disturbed constitutive COX-1 activity. We speculated that a segment prevented neointimal formation without causing sufficient supply of PGH led to an increased PGI 2 2 any hemorrhagic complications, and without increasing the production catalyzed by overexpressed PCS in the injured plasma and urinary levels of 6-keto-PGF . Hence, local arteries. In this study, the balloon-injured arteries transfect-1a PCS gene transfer may be more suitable for clinical ed with PCS gene produced higher levels of PGI without aurintricarboxylic acid prevents restenosis after vascular injury in affecting TXA production.