Antiproliferative Strategies for the Treatment of Vascular Proliferative Disease
Vicente Andrés and Claudia Castro
Laboratory of Vascular Biology, Department of Molecular and Cellular Pathology and Therapy, Instituto de Biomedicina de Valencia, Spanish Council for Scientific Research (CSIC), Valencia, Spain.
Corresponding author:
Vicente Andrés
Tel: +34-96-3391752
FAX: +34-96-3690800
E-mail:
ABSTRACT
Excessive cellular proliferation contributes to the pathobiology of vascular obstructive disease (e. g., atherosclerosis, in-stent restenosis, transplant vasculopathy, and vessel bypass graft failure). Therefore, anti-proliferative therapies may be a suitable approach in the treatment of these disorders. Candidate targets for such strategies include the cyclin-dependent kinase/cyclin holoenzymes, members of the cyclin-dependent kinase family of inhibitory proteins, tumor suppressors, growth factors and transcription factors that control cell cycle progression. In this review, we will discuss the use of pharmacological agents and gene therapy approaches targeting cellular proliferation in animal models and clinical trials of cardiovascular disease.
KEY WORDS: atherosclerosis, restenosis, bypass graft failure, cell cycle, pharmacological therapy, gene therapy
LIST OF ABBREVIATIONS: apoE, apolipoprotein E; AP-1, activator protein-1; BMS, bare metal stent; CDK, cyclin-dependent kinase; CKI, CDK inhibitory protein; EC, endothelial cell; ERK, extracellular signal-regulated kinase; IVUS, intravascular ultrasound; JNK, c-jun NH2-terminal protein kinase; MAPK, mitogen-activated protein kinase; ODN, oligodeoxynucleotide; PCNA, proliferating cell nuclear antigen; PDGF, platelet-derived growth factor; pRb, retinoblastoma protein; PTCA, percutaneous transluminal angioplasty; SAPK, stress-activated protein kinase; TGF-transforming growth factor-; VSMC, vascular smooth muscle cell.
1. Introduction
2. Preclinical studies
2.1. Pharmacological therapies
2.1.1. CDK inhibitors
2.1.2. Rapamycin and everolimus
2.1.3. Microtubule polymerising agents (Paclitaxel, Docetaxel)
2.1.4. Tranilast
2.1.5. Troglitazone
2.2. Gene therapy
2.2.1. Antisense approach
2.2.1.1. CDKs and cyclins
2.2.1.2. Mitogen-responsive nuclear factors that promote cell growth
2.2.2. Ribozymes
2.2.3. Transcription factor ‘decoy’ strategies
2.2.3.1. E2F
2.2.3.2. AP-1
2.2.4. Overexpression of growth suppressors
2.2.4.1. CKIs
2.2.4.2. p53
2.2.4.3. pRb
2.2.4.4. GATA-6
2.2.4.5.GAX
2.2.5. Overexpression of transdominant negative mutants of positive cell cycle regulators.
2.2.5.1. Ras
2.2.5.2. MAPKs
3. Clinical studies
3.1. Rapamycin
3.2. E2F ‘decoy’
3.3. Tranilast
3.4. Placlitaxel
3.5. c-myc antisense ODN
4. Concluding remarks
5. Acknowledegments
6. References
1. Introduction
Animal studies and large-scale clinical trials conducted over the last decades have allowed the identification of independent risk factors that increase the prevalence and severity of atherosclerosis (e. g., hypercholesterolemia, hypertension, smoking). Cardiovascular risk factors initiate and perpetuate an inflammatory response within the injured arterial wall that contributes to neointimal lesion growth during atherogenesis [1,2]. Abundant proliferation of vascular cells is an important component of the chronic inflammatory response associated to atherosclerosis and related vascular occlusive diseases (e. g., in-stent restenosis, transplant vasculopathy, and vessel bypass graft failure) [3,4]. Thus, understanding the molecular mechanisms that control hyperplastic growth of vascular cells should help develop novel therapeutic strategies to reduce neointimal thickening.
Although arterial cell proliferation occurs in animal models during all phases of atherogenesis [1,5-7], studies with hyperlipidemic rabbits have shown an inverse correlation between atheroma size and cellular proliferation within the atheromatous plaque [8-10]. Consistent with the response-to-injury hypothesis [1], medial cell proliferation at early stages of atherogenesis in fat-fed rabbits increased as a function of intimal lesion size [5]. Experimental angioplasty is also characterized by abundant proliferation of vascular smooth muscle cells (VSMCs), followed by the reestablishment of the quiescent phenotype typically within 2-4 weeks [11-13]. These animal studies suggest that vascular cell proliferation prevails at the onset of atherogenesis and restenosis.
Expression of proliferation markers in human primary atheromatous plaques and restenotic lesions has been well documented [14-26]. However, controversy exists regarding the magnitude of the proliferative response, ranging from a very low index of cell proliferation [15,16,18,20,22,26] to abundance of dividing cells [17,24,25,27]. Aside from methodological issues (e. g., differences in the fixatives used for tissue preservation, antigen accessibility, diversity of proliferation markers analyzed in these studies), some of the reported variance with regard to the issue of cell proliferation might relate to differences in the arteries being analyzed (i. e., peripheral, coronary and carotid arteries) and variance in the stage of atherogenesis at the time of tissue harvesting [28].
Proliferating cells within human atherosclerotic tissue include VSMCs, leukocytes and endothelial cells (ECs) [14-16,18-20,22]. Histological examination in 20 patients undergoing antemorten coronary angioplasty revealed that the extent of intimal proliferation was significantly greater in lesions with evidence of medial or adventitial tears than in lesions with no or only intimal tears [25]. Regarding the relative magnitude of intimal and medial cell proliferation, analysis of human carotid plaques revealed more proliferative activity in the intimal lesion versus the underlying media [20]. This study also disclosed differential distribution of proliferating cells in the intima versus the media; while the prevailing proliferative cell type in the intima was the monocyte/macrophage (46% versus 9.7% -actin immunoreactive VSMCs, 14.3% ECs, 13.1% T lymphocytes), VSMCs were the preponderant proliferating cell type in the media (44.4% versus 20% ECs, 13.0% monocyte/macrophages, and 14.3% T lymphocytes). It is noteworthy that studies in human peripheral and coronary lesions have suggested more prominent proliferation in restenotic compared to primary lesions [18,26,27]. Moreover, cultured VSMCs from human advanced primary stenosing disclosed lower proliferative capacity than cells from fresh restenosing lesions [29]. Thus, similar to the situation in animal models, proliferation during human atherosclerosis and restenosis might peak at the onset of these pathologies and then progressively decline.
Cell cycle progression is controlled by several cyclin-dependent kinases (CDKs) that associate with regulatory cyclins [30]. Mitogenic stimuli activate CDK/cyclin holoenzymes, thus causing hyperphosphorylation of the retinoblastoma protein (pRb) and the related pocket proteins p107 and p130 from mid G1 to mitosis. The complex interaction among E2F transcription factors and individual pocket proteins determines whether E2F proteins function as transcriptional activators or repressors [31]. Interaction of CDK/cyclins with CDK inhibitory proteins (CKIs) attenuates CDK activity and promotes growth arrest [32]. CKIs of the Cip/Kip family (p21Cip1, p27Kip1 and p57Kip2) bind to and inhibit a wide spectrum of CDK/cyclin holoenzymes, while members of the Ink4 family (p16Ink4a, p15Ink4b, p18Ink4c, p19Ink4d) are specific for cyclin D-associated CDKs. Mitogenic and antimitogenic stimuli affect the rates of synthesis and degradation of CKIs, as well as their redistribution among different CDK/cyclin pairs [32]. For example, p27Kip1 promotes the assembly of CDK4/cyclin D complexes by binding to them, thus facilitating CDK2/cyclin E activation through G1/S phase. Moreover, the protooncogen c-myc plays a key role in p27 sequestration through modulation of the level of cyclin D and E proteins.
VSMC proliferation in the balloon-injured rat carotid artery is associated with a temporally and spatially coordinated expression of CDKs and cyclins [23,33]. Moreover, induction of these factors correlated with increased CDK2 and CDC2 activity [23,34], demonstrating the assembly of functional CDK/cyclin holoenzymes in the injured arterial wall. Expression of CDK2 and cyclin E was also detected in human VSMCs within atherosclerotic and restenotic tissue [17,23,35], suggesting that induction of positive cell-cycle control genes is a hallmark of vascular proliferative disease in human patients.
In the following sections, we will discuss the use of pharmacological agents (Table 1) and gene therapy strategies targeting cellular proliferation in animal models (Table 2) and clinical trials (Table 3) of cardiovascular disease.
2. Preclinical studies
2.1. Pharmacological therapies
2.1.1. CDK inhibitors
Over fifty low molecular weight pharmacological CDK inhibitors that target the ATP-binding pocket of the catalytic site of CDKs have been identified and grouped in different families (purines, alkaloids, indirubins, flavonoids, paullones, butyrolactone I, hymenialdisine and pyrazolo[3,4-b]quinoxalines). Structural information on these agents and their application in cancer therapy was recently reviewed [36,37].The rat carotid model of balloon angioplasty has been used to assess the therapeutical efficacy of some pharmacological CDK inhibitors. The purine CVT-313 reduced neointimal lesion formation by 80% when delivered at a dose of 1.25 mg/kg for 15 minutes under pressure at the site of balloon angioplasty [38]. Likewise, flavopiridol at 5 mg/kg administered orally for 5 days beginning at the day of balloon angioplasty reduced neointima formation by 35% and by 39% at day 7 and 14 after intervention, respectively [39].
2.1.2. Rapamycin and everolimus
Rapamycin (sirolimus) is a macrolide antibiotic with potent immunosuppressant, antiproliferative and antimigartory properties [40]. The efficacy of rapamycin in attenuating neointimal thickening by both alloimmune and mechanical injury has been demonstrated in several animal models, including cardiac transplantation and restenosis post-angioplasty [41-48]. Although stabilization of the growth suppressor p27Kip1 appears to be a critical mediator of rapamycin-dependent growth arrest in vitro [49,50], the in vivo therapeutic efficacy of rapamycin in inhibiting neointima formation after mechanical injury was similar in wild-type and p27Kip1-null mice [51]. Oral administration of everolimus (40-O-(2-hydroxyethyl)-rapamycin, also dubbed RAD or certican) was well tolerated and suppressed in-stent neointimal growth in the rabbit iliac artery [52]. Moreover, everolimus attenuated the development of macrotaline-induced pulmonary vascular neointimal formation in pneumonectomized rats [53], and fluvastatin in combination with everolimus significantly reduced graft vascular disease in rat cardiac allografts [54].
2.1.3. Microtubule polymerising agents (Paclitaxel, Docetaxel)
Paclitaxel (Taxol) induces tubulin polymerization, thus stabilizing microtubules and causing G2/M arrest [55]. Both continuous exposure to paclitaxel and applications for 24 hours or even 20 minutes caused a complete and prolonged inhibition of the growth of human VSMCs stimulated in culture with several mitogens, with an IC50 of 2.0 nmol/L [56]. Moreover, locally delivered paclitaxel prevented neointimal thickening in animal models of balloon angioplasty and arterial stent implantation [56-61], thus making this agent a promising candidate for local antiproliferative therapy of restenosis. Of note, paclitaxel protected against vascular endothelial growth factor-mediated increase in neointimal formation after balloon angioplasty in femoral artery of cholesterol-fed rabbits [62]. Local delivery of docetaxel, a novel microtubule polymerising agent, also reduced neointimal hyperplasia in a balloon-injured rabbit iliac artery model [63].
2.1.4. Tranilast
The anti-inflammatory agent tranilast (SB 252218)attenuates mitogen-dependent proliferation and migration of VSMCs [64,65]. These inhibitory effects of tranilast are associated with increased expression of p53 and p21Cip1 and elevated complexing of p21Cip1 with CDK2 and CDK4. Tranilast inhibited neointimal formation in the rat carotid model of balloon angioplasty and in murine models of cardiac allograft vasculopathy, and these in vivo activities also correlated with p21Cip1 upregulation [64,66-68]. Importantly, genetic inactivation of p21Cip1 abolished tranilast-dependent reduction of neointimal thickening in a murine model of mechanical vascular injury [69], further implicating this CKI in the antiproliferative activity of tranilast. Tranilast also attenuated the proinflammatory activity of human monocytes, adding to its potential efficacy as a therapeutic agent in restenosis [70].
2.1.5. Troglitazone
The thiazolidinediones troglitazone and pioglitazone are novel insulin sensitizing agents that have been shown to inhibit the mitogenic effect of several growth factors in cultured VSMCs [71,72]. Inhibition of c-fos induction by troglitazone appeared to occur via a blockade of the mitogen-activated protein kinase (MAPK) pathway. When examined in vivo, troglitazone attenuated neointimal formation 14 days after balloon injury of the aorta compared with injured rats that received no troglitazone [72].
2.2. Gene therapy
Antiproliferative gene therapy strategies designed for the treatment of experimental cardiovascular disease can be grouped into two main categories: 1) Antisense approaches, ribozymes, and transcription factor ‘decoy’ strategies to inactivate positive cell cycle regulators (e. g., CDK/cyclins, protooncogenes, E2F, growth factors), and 2) Forced overexpression of negative regulators of cell growth (e. g., CKIs, p53, pRb, GAX, and GATA-6).
2.2.1. Antisense approach
This strategy typically utilizes a synthetic antisense oligodeoxynucleotide (ODN) that hybridizes in a complementary fashion and stoicheometrically with the target mRNA, thereby inactivating the gene of interest.
2.2.1.1. CDKs and cyclins
Several ODN strategies targeting CDKs and cyclins have proven effective in reducing neointimal lesion formation in animal models of balloon angioplasty, including ODNagainst CDK2 [34,73], CDC2 [34,74], and cyclin B1 [74]. Interestingly, cotransfection of antisense ODN against CDC2 kinase and cyclin B1 resulted in further inhibition of neointima formation, as compared to blockade of either gene target alone [74]. Of note, Morishita et al. reported sustained inhibition of neointima formation in the rat carotid balloon-injury model after a single intraluminal molecular delivery of combined CDC2 and proliferating cell nuclear antigen (PCNA) antisense ODNs [75], whereas this approach had no effect in the coronary arteries of pigs after balloon angioplasty [76]. Downregulation of cyclin G1 expression by retrovirus-mediated antisense gene transfer inhibited VSMC proliferation and neointima formation after balloon angioplasty [77]. ODN against CDK2 [78], and a combination of antisense ODN against PCNA and CDC2 [79], also attenuated experimental graft atherosclerosis.
2.2.1.2. Mitogen-responsive nuclear factors that promote cell growth
Several “immediate-early” genes (e. g., c-fos, c-jun, c-myc, c-myb, egr-1) are activated in serum-stimulated VSMCs, and their overexpression can induce VSMC proliferation in vitro[80-87]. Higher levels of c-myc mRNA are present in VSMCs cultured from atheromatous plaques than in VSMCs from normal arteries [88], and arterial injury induced the expression of several “immediate-early” genes [89-92]. Antisense ODNs against c-myc and c-myb reportedly inhibited in a sequence-specific manner both VSMC proliferation in vitro[83,93-100], and neointima formation after angioplasty [98,100-104] and vein grafting [105] in vivo. However, other studies have suggested that these inhibitory effects might be caused by a nonantisense mechanism [106-110].
2.2.2. Ribozymes
Targeted gene inactivation can be achieved by the use of ribozymes, a unique class of RNA molecules that catalytically cleave the specific target RNA. Su et al. designed a DNA-RNA chimeric hammerhead ribozyme targeted to human transforming growth factor-1 (TGF-1) that significantly inhibited angiotensin II-stimulated TGF-1 mRNA and protein expression in human VSMCs, and efficiently inhibited the growth of these cells [111]. Likewise, cleavage of the platelet-derived growth factor (PDGF) A-chain mRNA by hammerhead ribozyme attenuated human and rat VSMC growth in vitro [112,113].
The first evidence that ribozymes might represent useful tools in cardiovascular therapy came from studies using experimental models of angioplasty. Frimerman et al. reported the efficacy of chimeric hammerhead ribozyme to PCNA in reducing stent-induced stenosis in a porcine coronary model [114]. Moreover, ribozyme strategy against TGF-1 inhibited neointimal formation after balloon injury in the rat carotid artery model [115].
2.2.3. Transcription factor ‘decoy’ strategies
This approach consists of delivering a double-stranded ODN corresponding to the optimum DNA target sequence of the transcription factor of interest, thus leading to the sequestration of the specific trans-acting factor and attenuation of its interaction with the authentic cis-elements in cellular target genes.
2.2.3.1. E2F
E2F participates in the transcriptional activation of several growth and cell-cycle regulators (e. g., c-myc, pRb, cdc2, cyclin E, cyclin A), and genes encoding proteins that are required for nucleotide and DNA biosynthesis (e. g., DNA polymerase , histone H2A, pcna, thymidine kinase) [116,117]. E2F inactivation using synthetic ‘decoy’ ODN containing an E2F consensus binding site prevented experimental neointimal thickening in balloon-injured arteries [118], vein grafts [119,120], and cardiac allografts [121]. Ahn et al. developed a novel E2F ‘decoy’ ODN with a circular dumbbell structure (CD-E2F) and compared its properties with those of conventional phosphorothioated E2F ‘decoy’ ODN (PS-E2F) [122]. CD-E2F displayed more stability and stronger antiproliferative activity than PS-E2F when assayed in cultured VSMCs, and was more effective in inhibiting neointimal formation in vivo.
2.2.3.2. Activator protein-1 (AP-1)
Cell proliferation in the rat carotid arterymodel of angioplastycorrelated with elevated expression and high DNA-binding activity of transcription factors of the AP-1 family [89-91,123,124]. AP-1 ‘decoy’ ODN delivery into cultured human VSMCs significantly reduced cell number and TGF-1 production under conditions of PDGF stimulation [125], and attenuated neointimal thickening when applied at the site of balloon angioplasty in rabbit carotid [125] and minipig coronary arteries [126]. Compared to conventional phosphorothioated AP-1 decoy ODN, circular dumbbell AP-1 ‘decoy’ ODN was more effective in inhibiting the proliferation of VSMCs in vitro and neointimal hyperplasia in vivo [127].
2.2.4. Overexpression of growth suppressors
2.2.4.1. CKIs
The first evidence that p21Cip1 and p27Kip1 may function as negative regulators of neointimal hyperplasia was suggested in animal studies showing the upregulation of these CKIs at late time points following balloon angioplasty, coinciding with the restoration of the quiescent phenotype after the initial proliferative wave [21,128]. The protective role of p27Kip1 against neointimal thickeninghas been rigorously demonstrated in hypercholesterolemic apolipoprotein E (apoE)-deficient mice, in which genetic inactivation of one or two p27Kip1 alleles progressively accelerated atherogenesis [6]. However, neointimal hyperplasia after mechanical damage of the arterial wall was similar in wild-type and p27Kip1-null mice [51]. Redundant roles between p21Cip1 and p27Kip1, or compensatory increase in p21Cip1 expression (or other CKIs) might account for the lack of phenotype of p27Kip1-null mice in the setting of mechanical arterial injury.
Several studies have suggested a role of CKIsin establishing regional phenotypic variance in VSMCs from different vascular beds. Using human VSMCs isolated from internal mammary artery and saphenous vein, Yang et al. suggested that sustained p27Kip1 expression in spite of growth stimuli may contribute to the resistance to growth of VSMCs from internal mammary artery and to the longer patency of arterial versus venous grafts [129]. Likewise, dissimilar proliferative response of intimal and medial VSMCs towards basic fibroblast growth factor (bFGF or FGF2) correlated with distinct expression of p15Ink4b and p27Kip1 in these cells [130]. Intrinsic differences in the regulation of p27Kip1 might also play an important role in creating variance in the proliferative and migratory capacity of VSMCs isolated from different vascular beds, which might in turn contribute to establishing regional variability in atherogenicity [131].