Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Prostaglandin generation is enhanced in patients with atherosclerosis and after balloon angioplasty (Braden et al., 1991; Belton et al., 2000). The products include prostacyclin (PGI₂) from vascular endothelium and thromboxane A₂, which is largely derived from platelets. Prostaglandins are synthesized from arachidonic acid by the enzyme cyclooxygenase (COX), of which there are two isoforms, COX-1 and COX-2 (Hla and Neilson, 1992; Xi et al., 1994). COX-1 is constitutively expressed in most tissues, including platelets and is largely responsible for TXA₂ formation. COX-2 is normally undetectable or is expressed in low amounts (Masferrer et al., 1994). Nevertheless, COX-2 is largely responsible for the biosynthesis of PGI₂ in normal subjects (Cullen et al., 1998; Catella-Lawson et al., 1999; McAdam et al., 1999). COX-2 is induced in the vascular smooth muscle and inflammatory cells of human atherosclerotic plaque (Baker et al., 1999; Schonbeck et al., 1999; Belton et al., 2000) and we have shown that both isozymes are responsible for the increase in PGI₂ generation in patients with severe atherosclerosis (Belton et al., 2000). The expression of COX-2 in vascular injury is not unexpected, because COX-2 is readily induced by growth factors (Lin et al., 1989), cytokines (Jones et al., 1993), and free radicals (Adderley and Fitzgerald, 1999).

COX isozymes and prostaglandins may modify the response to vascular injury. Prostaglandins regulate the expression of genes involved in cell growth (Brown et al., 1999), apoptosis (Bornfeldt et al., 1997), and migration (Attiga et al., 2000), processes that have been implicated in the development of lesions in atherosclerosis (Ross, 1993) and after balloon angioplasty (Clowes and Reidy, 1991). Alternatively, vasodilator prostaglandins such as PGI₂ may limit the injury response because overexpression of prostacyclin synthase in the vessel wall decreases neointimal hyperplasia, potentially by accelerating regeneration of the endothelium (Harada et al., 1999). Other products such as PGJ₂ and its metabolites suppress the expression of inflammatory genes (Straus et al., 2000). COX-2 may also limit disease progression or the development of thrombotic complications through PGI₂-mediated inhibition of platelet function. Indeed, inhibition of COX-2-derived PGI₂ may underlie the reports of thrombosis in patients treated with selective COX-2 inhibitors (Bombardier et al., 2000; Crofford et al., 2000). To explore the functional role of COX isozymes in lesion development after vascular injury, we examined the expression of COX isozymes in the rat model of balloon injury to the carotid artery and explored their role in the subsequent development of restenosis.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animal Model of Intimal Hyperplasia. Eighty-five male Sprague-Dawley rats weighing 250 to 300 g were studied. The animals were maintained on normal diet and were allowed free access to chow and water throughout this study. Animal care and treatment were conducted in conformity with institutional guidelines in compliance with international laws and policies. A license and permission for the study were obtained from the Department of Health.

Immunohistochemistry Analysis. Left carotid artery samples were fixed in formal saline for 24 h. The tissues were paraffin embedded (Shandon Citadel 200; Shandon Lipshaw Inc., Pittsburgh, PA) and 5-µm sections were cut (Leitz 1512 microtome; Weltzar GmbH, Wetzlar, Germany). The sections were incubated in primary antibody against COX-1 or COX-2 (both polyclonal antibodies, Santa Cruz Biotechnology, Santa Cruz, CA) for 1 h at room temperature. After washing in phosphate-buffered saline the slides were incubated in the secondary biotinylated antibody and the immunocomplex visualized using the diaminobenzidine chromagen (ABC Complex, Vectastain Elite kit; Vector Laboratories, Burlingame, CA).

RT-PCR Analysis. COX-1, COX-2, and GAPDH mRNA were extracted and detected by RT-PCR as described previously (Belton et al., 2000). Each primer pair was designed to span at least one intron of the gene. The primers used were as follows: COX-2 sense: 5'-AACACGGACTTGCTCACTTT-3'; COX-2 antisense: 5'-TACTGTAGGGTTAATGTC-3; COX-1 sense: 5'-TCACAAGAGTACAGCTAT-3'; COX-1 antisense: 5'-TGGGCTGGCACTTCTCCA-3'; GAPDH sense: 5'-AACCCATCACCATATTCCAGGAGC-3'; and GAPDH antisense: 5'-CACAGTCTTCTGAGTGGCAGTGAT-3'.

Urinary Eicosanoid Excretion. Urine was collected from the animals over 24 h before surgery and on the third day after surgery. Urinary TXB₂ was determined by enzyme immunoassay (R & D Systems, Abingdon, UK). Five milliliters of urine was spiked with deuterated internal standards for 1 ng/ml PGF₂, 1 ng/ml PGE₂, and 1 ng/ml isoprostane 8-iso-PGF₂. The sample was purified and concentrated by solid phase extraction by using a C18 Sep-Pak cartridge.

COX-1 Activity in Whole Blood. Serum TXB₂ was assayed by allowing whole blood to clot in nonsiliconized glass tubes at 37°C for 1 h (Panara et al., 1995). Serum was separated by centrifugation at 1000g for 10 min. TXB₂ was measured by enzyme immunoassay (R & D Systems).

Quantification of Intimal Hyperplasia. The intimal area and medial area of each artery (n = 7 for each group) were obtained by computerized planimetry (Samba software; Alcatel, Grenoble, France). Neointimal thickening was quantified as the ratio of intima to media areas (I/M) from transverse sections of hematoxylin- and eosin-stained arterial sections. Two to four sections were examined for each section by an analyst blind to the treatment regimen and the average I/M calculated. The data are expressed as mean ± S.E.M.

Statistical Analysis. The data are expressed as mean ± S.E.M. For comparison between groups, the data were analyzed by analysis of variance followed by paired or unpaired tests as appropriate, or in instances where data distribution deviated from normality, by using the Kruskal-Wallis nonparametric analysis of variance and subsequent Mann-Whitney U test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Immunohistochemistry. Analysis of COX-1 and COX-2 protein expression was performed on sections of rat carotid artery 14 days after balloon injury. In the balloon injury group, a neointima encircling the lumen was clearly visible (Fig. 1). COX-1 was diffusely expressed throughout the intima and media of the normal vessel (Fig. 1A). COX-1 was also expressed in the cells of the neointima after carotid angioplasty (Fig. 1C). There was COX-2 expression in the adventitia of the carotid arteries from both the sham-operated animals and the animals undergoing balloon injury (Fig. 1, B and D). This probably reflects damage and inflammation triggered during isolation of the carotid artery because no COX-2 was detected in the absence of surgery (data not shown). In addition, there was a marked increase in COX-2 expression in the neointima in the animals that underwent balloon injury (Fig. 1D). The increase in COX-2 expression was confined to the neointima and was not evident in the media. The specificity of the COX-2 expression was confirmed by preabsorbing the COX-2 antibody with a peptide against which it was raised (Fig. 1E).

View larger version (122K): [in this window] [in a new window]   Fig. 1.   Immunohistochemistry analysis for COX-1 and COX-2 in the rat carotid artery after angioplasty. COX-1 (A) and COX-2 (B) expression in sham-operated animals. Only COX-1 is expressed in the media of the vessel. After balloon angioplasty, there is COX-1 (C) and COX-2 (D) expression in the neointima. The specificity of the COX-2 expression was confirmed by first preabsorbing the antibody with a COX-2 peptide against which the antibody was raised (E). A indicates adventitia. COX-1- and COX-2-positive staining is evident by the appearance of a brown color.

RT-PCR. In animals that were not subjected to anesthesia or any surgical procedure, there was no COX-2 mRNA expression in the carotid rat artery. In all samples from animals undergoing surgery with or without balloon injury (sham-operated animals), there was expression of COX-1 and COX-2. The results suggest that isolation of the vessel alone, even in the absence of balloon injury, induced COX-2 mRNA expression, consistent with the adventitial expression of COX-2 seen on immunohistochemistry in the sham-operated animals. COX-2 mRNA expression was unaffected by either of the COX inhibitors (data not shown).

Urinary Eicosanoid Production. Urinary TXB₂, PGE₂, PGF₂, and 8-iso-PGF₂ were measured before surgery and at 3 days after angioplasty (Fig. 2). TXB₂, PGE₂, and PGF_2, (ng/mg creatinine) were unaltered in sham-operated controls after surgery compared with levels seen preoperatively (TXB₂ 7.2 ± 0.5 versus 5.5 ± 0.7 (n = 8), p = 0.19; PGF₂ 10.6 ± 1.4 versus 8.9 ± 1.2 (n = 8), p = 0.39; and PGE₂ 21.8 ± 3.6 versus 23.5 ± 2.6 (n = 8), p = 0.21). However, after balloon angioplasty there was a marked increase in the their excretion compared with sham-operated controls [TXB₂: 19.8 ± 1.7 (n = 8), p < 0.001; PGF₂: 24.8 ± 2.7 (n = 8), p <0.001; and PGE₂: 66.4 ± 4.9 (n = 8), p < 0.001].

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Urinary prostaglandin generation in sham animals and after balloon angioplasty. Data are mean ± S.E.M.; $, p < 0.001 versus before surgery; **, p < 0.01; and ***, p < 0.001 versus balloon injury on no treatment.

2

2

2

2

2

Serum TXB₂. Serum TXB₂, an assay of COX-1 activity, was reduced by SC-560 alone and when combined with SC-236. SC-236 alone had no effect on serum TXB₂ formation (Fig. 3).

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Measurement of COX-1 activity. Serum TXB2 from whole blood after balloon angioplasty. SC-560 and combination of SC-560 with SC-236 reduced serum TXB2 levels. *, p < 0.05 versus no treatment.

Quantification of Intimal Hyperplasia. Intimal hyperplasia developed over the 14 days after balloon injury in the group that received no treatment (I/M ratio 1.76 ± 0.2; n = 7). SC-236, the selective COX-2 inhibitor, had no effect on the development of intimal hyperplasia (I/M ratio 1.61 ± 0.2; n = 7, p = 0.92). Administration of SC-560 significantly reduced intimal hyperplasia (I/M ratio 1.03 ± 0.2; n = 7, p < 0.05) as did the combination of SC-560 and SC-236 (I/M ratio 1.06 ± 0.1; n = 7, p < 0.05) (Fig. 4).

View larger version (11K): [in this window] [in a new window]   Fig. 4.   Intima/media ratio of the carotid artery 14 days after balloon angioplasty. Measurement of intimal hyperplasia in animals that did not receive any treatment and in animals that received SC-236, SC-560, or a combination of both treatments. Data are mean ± S.E.M., *, p < 0.05 versus balloon injury on no treatment.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Our study demonstrates expression of both COX-1 and COX-2 in the neointima after balloon injury of the carotid artery. The expression of the two COX isoforms in the neointima of vascular lesions is not surprising. COX-2 is induced by several factors implicated in the development of vascular proliferation and restenosis, including growth factors, free radicals, and cytokines (Lin et al., 1989; Jones et al., 1993; Adderley and Fitzgerald, 1999). COX-1 is also inducible (McAdam et al., 2000) and in any event is constitutively expressed in virtually all cells. There was also a marked increase in prostaglandin formation and thromboxane. Both PGE₂ and thromboxane are relevant, because they are major products of vascular smooth muscle and inflammatory cells. Moreover, COX-2 colocalizes with PGE synthase in human atherosclerotic plaque (Cipollone et al., 2001). Although the source of the increase was not defined, because there was no increase in sham-operated animals, it is reasonable to infer that the rise in urinary product excretion derived from the site of vascular injury. However, it should be noted that these are primary prostaglandins and we cannot exclude a renal origin.

To explore the functional roles of COX-1 and COX-2 in this model, we examined the effect of SC-236 and SC-560. SC-236 is highly selective for COX-2 (Kishi et al., 2000). Consistent with this, SC-236 had no effect on serum TXB₂, an assay of COX-1 activity, despite marked suppression of urinary prostaglandin metabolites. SC-560 is selective for COX-1 (Smith et al., 1998) and as expected inhibited serum TXB₂, although only by 60%. Other studies have described an 85% reduction in TXB₂ production in calcium ionophore-stimulated whole blood with 10 mg/kg SC-560 (Smith et al., 1998). The incomplete suppression of both serum TXB₂ and whole blood-stimulated TXB₂ is probably due to the fact that SC-560, although selective for COX-1, is a competitive inhibitor.

Both compounds reduced PGE₂ and TXB₂ excretion. This is not surprising, because the two COX isoforms are capable of generating all prostaglandins, and although there may be quantitative differences in product formation, no one product discriminates between the two isoforms. The findings are similar to human studies of atherosclerosis, where both isoforms are responsible for the increase in prostaglandin generation (Belton et al., 2000). SC-560 markedly inhibited TXA₂ generation based on a marked reduction in TXB₂ excretion in animals. This is to be expected, because platelet COX-1 is a major source of thromboxane generation in vivo (Clarke et al., 1991). The selective COX-2 inhibitor SC-236 also inhibited TXA₂ although to a smaller degree. Tissue COX-2 may contribute to TXA₂ in some settings, for example, in patients with unstable angina where there is persistent formation of TXA₂ despite treatment with aspirin (Cipollone et al., 1997). Indeed, several cell types found in vascular lesions express COX-2 and generate TXA₂, including vascular smooth muscle cells and monocytes (Penglis et al., 2000; Young et al., 2000).

Smooth muscle cell proliferation is a key event in the development of atherosclerosis, restenosis, and intimal hyperplasia (Clowes and Reidy, 1991; Ross, 1993). Prostaglandins may influence the process in several ways. TXA₂ and PGF₂, and the isoprostane 8-iso-PGF₂ (a free radical-derived product of arachidonic acid) induce proliferation of vascular smooth muscle cells through activation of G protein-coupled transmembrane receptors (Craven et al., 1996; Pakala et al., 1997; Zucker et al., 1998). Platelets that adhere at the site of injury may also contribute to lesion formation through the release of growth factors or by triggering thrombosis (Clowes and Reidy, 1991; Ross, 1993). Therefore, in addition to a direct effect on vascular tissue, TXA₂ may contribute to lesion formation by amplifying platelet activation at the site of vascular injury.

Alternatively, prostaglandins may limit lesion development. Overexpression of COX-2 or prostacyclin synthase suppresses the development of vascular lesions and the growth of vascular smooth muscle cells (Hara et al., 1995; Harada et al., 1999). These effects may be mediated through peroxisome proliferator-activated receptors (Kilewer, 1997; Staels et al., 1998), a series of nuclear membrane receptors for prostaglandins that heterodimerize with other transcription factors and regulate the expression of genes in vascular smooth muscle cells, including those involved in cell growth and/or apoptosis.

In the model of balloon injury to the carotid artery, COX-1 inhibition reduced lesion development and no additional effect was seen when both isozymes were inhibited by combining the two drugs. The results suggest a role for COX-1 alone in the development of the neointima. Although this could reflect suppression of tissue prostaglandin generation, overexpression of COX-1 in vascular tissue by using an adenoviral vector protects against lesion development by preventing the local formation of thrombus (Zoldhelyi et al., 1996). Alternatively, the effects of SC-560 may be due to the suppression of platelet and/or vascular TXA₂, a potent platelet activator and mitogen for vascular smooth muscle cells. Similar inhibition of the vascular response to injury has been seen with selective TXA₂/prostaglandin endoperoxide receptor antagonists, supporting the hypothesis (Cayatte et al., 2000).

No inhibition of neointimal hyperplasia was detected using the selective COX-2 inhibitor despite a marked reduction in prostaglandin formation. Because COX-2 is a major source of endogenous PGI₂, concern has been raised that COX-2-selective inhibitors may allow unopposed TXA₂-mediated effects on platelets and place patients at risk of thrombosis (Cullen et al., 1998; Catella-Lawson et al., 1999; McAdam et al., 1999). Reports of thrombosis in patients receiving selective COX-2 inhibitors are consistent with this hypothesis (Bombardier et al., 2000; Crofford et al., 2000). We saw no increase in lesion development and no thrombosis, although it should be emphasized that we did not measure PGI₂ biosynthesis in vivo in this model, a limitation given that PGI₂ may influence lesion development by inhibiting platelet activity at the site of vascular damage (Zoldhelyi et al., 1996). Moreover, thrombosis is not a major feature of animal models. Thus, further studies are required to explore the effect of COX-2 on PGI₂ formation and the development of thrombus in vivo.

In conclusion, selective inhibition of COX-1 reduced the development of the neointimal hyperplasia that follows balloon injury of the carotid artery, whereas inhibition of COX-2 had no effect. Our findings are consistent with findings showing evidence of a role for COX-1 but not COX-2 in a murine model of atherosclerosis (Pratico et al., 2001).