TGF-β1 levels increase after vascular injury and promote vascular smooth muscle cell (VSMC) proliferation. We define a nonviral gene delivery system that targets αvβ3 and α5β1 integrins that are expressed on proliferating VSMCs and strongly induced by TGF-β1. A 15-amino acid RGDNP-containing peptide from American Pit Viper venom was linked to a Lys(16) peptide as vector (molossin vector) and complexed with Lipofectamine or fusogenic peptide for delivery of luciferase or β-galactosidase reporter genes to primary cultures of human, rabbit, and rat VSMCs. Preincubation of VSMCs with TGF-β1 for 24 h, but not with PDGF-BB, interferon-γ, TNF-α, nor PMA, increased αvβ3 and α5β1 expressions on VSMCs and enhanced gene delivery of molossin vector. Thus β-galactosidase activity increased from 35 ± 5% (controls) to 75 ± 5% after TGF-β1 treatment, and luciferase activity increased fourfold over control values. Potential use of this system in vessel bypass surgery was examined in an ex vivo rat aortic organ culture model after endothelial damage. Molossin vector system delivered β-galactosidase to VSMCs in the vessel wall that remained for up to 12 days posttransfection. The molossin vector system, when combined with TGF-β1, enhances gene delivery to proliferating VSMCs and might have clinical applications for certain vasculoproliferative diseases.
- peptide vector
- molossin rector
proliferation and migration of vascular smooth muscle cells (VSMCs) play major roles in the development of intimal hyperplasia and are crucial events that lead to chronic allograft rejection, postangioplasty restenosis and in-stent stenosis (1, 15). The major rationale for early genetic intervention in the treatment of these disorders is to deliver functional copies of genes that control proliferation and/or migration specifically into VSMCs. However, those nonviral DNA vectors that currently are available have low levels of efficiency and lack specificity, whereas viral vectors, such as retroviral or adenoviral vectors, display high gene transfer efficiencies but suffer from immunogenicity, potential oncogenic risks when used in the clinical setting, and they are not specific for proliferating and migrating VSMCs (20). Therefore, a nonviral DNA vector system, with no toxicity but high efficiency and selectivity for proliferating and migrating VSMCs, would greatly enhance the prospect of gene therapy for treating vasculoproliferative diseases.
Integrins are heterodimeric transmembrane receptors consisting of noncovalently linked α- and β-subunits that mediate many cellular responses in VSMCs and other cell types, including differentiation, proliferation and migration (11, 23, 27). Furthermore, α5β1 and αvβ3 have been implicated in pathological conditions such as atherosclerosis, restenosis after angioplasty and constrictive vascular remodeling after injury (25). Vascular injury induces α5β1 integrin expression exclusively in proliferating VSMCs at the luminal surface of the neointima (17), and VSMC invasion from the tunica media to the intima has been shown to be dependent on αvβ3 integrin expression (13). Interestingly, integrin-mediated cell entry is exploited by a number of intracellular pathogens including adenovirus and encovirus (5). The widespread use of this means of cell entry in nature suggests that integrin-binding ligands might be suitable moieties for the development of receptor-mediated gene delivery vectors (2, 10).
TGF-β1 has profound autocrine and paracrine biological effects on VSMCs and on neointima formation after allograft transplantation and angioplasty (4, 28). TGF-β1 also contributes to arterial remodeling after injury by stimulating VSMC adhesion molecule synthesis at sites of vascular damage (8) and induces β3 and α5 integrin expression in VSMCs (3, 12). Given the important implications of TGF-β in the pathogenesis of local occlusive vascular diseases, a selective gene delivery system, which is enhanced by a local effect from elevated levels of TGF-β1 in proliferating and migrating VSMCs, would be advantageous.
We have used a nonviral gene delivery system consisting of a 15-amino acid peptide derived from the venom of the American pit viper (Crotalus molossus) linked to 16 consecutive lysines, to permit electrostatic binding of DNA, as a vector (molossin vector) (6, 14, 24). Molossin vector contains an RGDNP motif that displays high affinities toward αvβ3 and α5β1 integrins (as opposed to the RGDW motif present in other pit viper venoms) (22, 31). In the present study, we have used molossin vector with Lipofectamine or a fusogenic peptide to introduce genes into proliferating VSMCs pretreated with TGF-β1. We also tested the potential application of the molossin vector system to deliver genes into the vessel wall in an aortic organ culture model.
MATERIALS AND METHODS
Materials. PMA, human recombinant TGF-β1, human recombinant TNF-α, human recombinant PDGF-BB, and human recombinant interferon-γ (INF-γ) were from Sigma (St. Louis, MO). Mouse IgG1 monoclonal antibody to molossin peptide (LC2–64) was raised as described previously (14). Mouse IgG1 antibodies to human αvβ3 integrin (CD49e) and α5β1 integrin (CD51/CD61) were from Serotec (Bicester; Oxon, UK). Mouse IgG2a antibody to α-smooth muscle actin and mouse monoclonal antibody to β-galactosidase (β-gal) were from Sigma.
Peptides. Molossin peptide vector, consisting of 16 lysines at the amino terminus NH2-(Lys)16-Ile-Cys-Arg-Arg-Ala-Arg-Gly-Asp-Asn-Pro-Asp-Asp-Arg-Cys-Thr-COOH, was synthesized, purified and cyclized by Cambridge Research Biochemicals (Zeneca-CRB; Northwich, Cheshire, UK). HPLC confirmed purity as 96.7%. The 20-amino acid fusogenic peptide of influenza virus (NH2-Gly-Leu-Phe-Glu-Ala-Leu-Leu-Glu-Leu-Leu-Glu-Ser-Leu-Trp-Glu-Leu-Leu-Leu-Glu-Ala-COOH) was synthesized and purified by Zinsser Analytic (Maidenhead, Berkshire, UK). Purity was >90%. Peptides were dissolved in 0.15 M NaCl at stock concentrations of 0.1 mg/ml and then stored in small aliquots at –35°C.
Primary culture of VSMCs. Normal primary cultures of human aortic VSMCs (batch F-14579) were obtained from American Type Culture Collection (Manassas, VA) and maintained in Ham's F-12K medium supplemented with 10% fetal calf serum (FCS), 2 mM l-glutamine, 10 μg/ml insulin (Life Technologies, Paisley, UK), 1 μg/ml transferrin (Sigma), 10 ng/ml sodium selenite (Sigma), and 20 μg/ml endothelial cell growth supplement (Sigma). Rat and rabbit VSMCs were isolated from thoracic aortas as described previously (14). Cells from all three species were used between passages 2 and 6 and were in the synthetic (proliferative) stage of growth. In all cases, >99% of cells stained positive for α-smooth muscle actin.
Formation of transfection complexes and transfection of VSMCs. The methods and concentrations of each component required to form transfection complexes have been examined and optimized previously (14, 24, 30, 31). A molossin (fuso-molossin)-vector/DNA/Lipofectamine ratio of 5:1:2 was used to achieve maximum transfection efficiency (14). VSMCs (1 × 105/well in 24-well plates) were seeded in appropriate culture medium the day before transfection. Cells then were cultured for appropriate periods of time in the presence or absence of TGF-β1 (1–20 μg/ml). Cells were transfected with 0.4 ml of DNA complex for 3–4 h after which an equal volume of culture medium containing 20% FCS was added followed by overnight culture. Medium then was replaced with fresh medium. Three days after initial transfection, cells were analyzed for expression of reporter genes.
Reporter gene assays. β-Gal and luciferase reporter gene assays were performed as described previously (14). For organ transfection, vessels were cut into 5-mm segments and stained for β-gal as described (14). Vessels were fixed in 10% formalin, wax-embedded, cut into 5-μm sections, and counterstained with neutral red (to stain nuclei). Luciferase enzyme activity was measured with a luciferase assay kit (Promega), and results were expressed as relative light units per milligram of total protein.
Ex vivo transfection of aortic vessel with β-gal gene and immunocytochemistry. Adult male Wistar rats were anesthetized terminally by injection of pentobarbitone (0.6 mg/g body wt ip). The thoracic aorta was removed, and the endothelium was injured by three passes with a 3-mm diameter balloon. After being washed, one end of the vessel was clamped with a small hemostat and the other end of the aorta was mounted on a small plastic tube connected to a 1-ml syringe. The liquid in the lumen was decanted, and the aorta was filled with transfection (β-gal) complex solution. The aorta was placed into DMEM and incubated at room temperature for 10 min after which both ends of the aorta were trimmed off and the remaining vessel cut into 1-cm pieces. Three pieces per well were placed into 6-well plates and incubated in 3 ml of transfection complex solution for 4 h after which an equal volume of medium containing 20% FCS was added overnight. The medium was then replaced with fresh culture medium. Vessels were cultured for a further 3–14 days, and β-gal expression in the vessel wall was examined by histochemical staining or immunofluorescence on frozen tissue sections as previously described (14). All animals were housed and cared for in accordance with current UK Home Office Guidelines. All protocols adhered to APS's “Guiding Principles in the Care and Use of Animals.”
Statistical analysis. Quantitative data are shown as means ± SD unless otherwise stated. Differences in mean measurements among experimental groups were tested by ANOVA, followed by a Student's t-test with a Bonferroni correction, and differences were considered significant at the 95% level.
Effect of TGF-β1 on α5β1 and αvβ3 integrin expression in human VSMCs. Human aortic VSMCs were used to assess the effect of TGF-β1 on α5β1 and αvβ3 integrin expressions. Cells were pretreated with or without TGF-β1 for 24 h and examined for the expression of α5β1 and αvβ3 integrins by flow cytometric analysis. In the absence of TGF-β1, VSMCs expressed basal levels of α5β1 and αvβ3 integrins (Fig. 1). However, after treatment with TGF-β1, VSMCs increased α5β1 and αvβ3 integrin expressions resulting in a right shift of the flow cytometry profile. These results are in accordance with previous reports that TGF-β1 induces α5 and β3 integrin expression in vascular cells (3, 12) and suggest that use of a molossin vector, which is known to target α5β1 and αvβ3 integrins, might be more effective as a gene delivery system in VSMCs when under the influence of TGF-β1.
Effect of TGF-β1 on lipo-molossin- or fuso-molossin-mediated gene transfection. The molossin vector previously has been used in cultured VSMCs (14) and in other cell types and tissues (24, 31), and the system needs Lipofectamine or fusogenic peptide to assist endocytic exit (14, 30). To determine the transfection efficiency before and after TGF-β1 pretreatment of VSMCs, we used the β-gal and GFP reporter genes. Figure 2 shows that transfection of rat VSMCs with the β-gal gene in combination with the Lys(16) alone failed to yield any β-gal expression (Fig. 2A), whereas molossin vector (Fig. 2B) or Lipofectamine (Fig. 2C), when used alone, enabled ∼2–3% transfection efficiency. Figure 2D shows that cells transfected with β-gal in combination with Lipofectamine-Lys(16) led to a small increase in transfection efficiency (5 ± 2%) compared with Lipofectamine (Fig. 2C). There was no difference in the level of transfection between cells treated with or without TGF-β1. Interestingly, when molossin vector was used in combination with Lipofectamine (lipo-molossin) (Fig. 2E), transfection efficiency was increased to 75 ± 5% in rat VSMCs pretreated with TGF-β1 compared with 35 ± 5% in cells without TGF-β1 pretreatment (Fig. 2F). Thus the effect was dependent on the presence of the integrin targeting moiety of the vector, because TGF-β1 pretreatment did not improve gene delivery when a Lys(16) control peptide was used with Lipofectamine (Fig. 2D). Similarly, fusogenic peptide when used alone (Fig. 2G) displayed no evidence of transfection as determined by measuring GFP expression regardless of whether cells were pretreated with or without TGF-β1. However, when molossin vector formed complexes with fusogenic peptide (fuso-molossin) it successfully delivered the GFP reporter gene to human VSMCs. The transfection efficiency of fuso-molossin was 40 ± 5% in TGF-β1 pretreated human VSMCs (Fig. 2H) compared with 25 ± 5% in the same cells without TGF-β1 pretreatment (data not shown). Thus pretreatment of cells with 10 ng/ml TGF-β1 for 24 h before transfection increases transfection efficiency of both lipo-molossin and fuso-molossin vectors twofold when compared with cells not pretreated with TGF-β1.
Using neutralizing antibodies to α5β1 and αvβ3 integrins, we demonstrated that lipo-molossin targets VSMCs. Thus Fig. 3 shows that incubation of cells with antibodies to α5β1 or αvβ3 for 1 h before transfection significantly (P < 0.03) inhibited molossin vector delivery of the luciferase gene (as measured by luciferase activity) compared with cells incubated with medium only. The combination of the two antibodies inhibited luciferase activity by 85% without TGF-β1 and 89% with TGF-β1. These results suggest that molossin vector targets α5β1 and αvβ3 integrins and that it might be more effective as a gene delivery system in VSMCs when under the influence of TGF-β1.
We also used lipo-molossin to deliver the luciferase reporter gene into human aortic VSMCs pretreated with different concentrations of TGF-β for 24 h or at a fixed concentration (10 ng/ml) but during variable periods (Fig. 4A). Pretreatment of cells with TGF-β1 significantly increased luciferase reporter gene expression. Thus the luciferase activity in TGF-β1 (10 ng/ml) pretreated VSMCs was approximately sixfold of that present in cells without TGF-β1 pretreatment (Fig. 4A, top). Therefore, 10 ng/ml of TGF-β1 was used in all subsequent experiments. The effect of TGF-β1 on transfection efficiency into VSMCs was time dependent because the maximal effect, which produced approximately sevenfold greater luciferase activity than the activities detected in cells without TGF-β1 pretreatment, was observed 24 h after pretreatment (Fig. 4B, bottom). The effect of enhancing molossin vector gene delivery to VSMCs is TGF-β1 specific because pretreatment of VSMCs with PDGF-BB (10 ng/ml), TNF-α (100 U/ml), INF-γ (100 U/ml), or the PKC activator PMA (100 ng/ml) for 24 h had no effect on lipo-molossin gene delivery (Fig. 4B). The effect of TGF-β1 was further examined in primary VSMC cultures derived from human, rabbit, and rat. Compared with cells without TGF-β1 treatment, TGF-β1 pretreatment significantly increased the efficiency of gene delivery of both fuso-molossin and lipo-molossin vectors to all three species of VSMCs (P < 0.03) (Fig. 4C). Interestingly, lipo-molossin gene delivery was greater than that achieved with fuso-molossin in these studies. One reason for this might involve an increased binding of transfection complex to the cell surface and better endocytic escape of molossin vector/DNA with Lipofectamine. Finally, Fig. 4D shows that TGF-β1 enhances molossin vector delivery of the luciferase gene in human VSMCs ∼20-fold more selectively than in some other cell types, including hepatocytes and fibroblasts.
Ex vivo gene delivery to VSMC in the vessel wall using lipo-molossin. Figure 5 illustrates that molossin vector bound effectively to medial VSMCs (Fig. 5B), whereas control tissue, which had not been incubated with molossin vector (Fig. 5A), showed limited background staining.
We then examined the potential application of this gene transfer technology to deliver genes into VSMCs in the vessel wall in an ex vivo model of rat aortic organ culture. The β-gal reporter gene was employed to assess the efficiency of lipo-molossin and was visualized by immunohistochemical staining (Fig. 5, C–E). Thus control vessels transfected with lipo-molossin alone (Fig. 5C) or with Lipofectamine plus β-gal DNA alone (Fig. 5D) showed no evidence of β-gal expression. In contrast, vessels transfected with β-gal DNA plus lipo-molossin showed β-gal expression (by X-gal staining) mainly in medial VSMCs (Fig. 5E, blue color). Some β-gal expression also was detected in the junction area between the media and adventitia (Fig. 5E). Expression of β-gal in the vessel wall also was confirmed by immunofluorescence staining by using a monoclonal antibody against Escherichia coli β-gal. Expression of β-gal was detectable 3 days after initial transfection, whereas 12 days after transfection, VSMCs had migrated out from the media layer to form neointima. Interestingly, β-gal expression was detectable in neointima cells (Fig. 6). To quantitatively measure the gene delivery efficiency in rat aortic vessels, we performed a series of luciferase activity studies (summarized in Table 1). Lipo-molossin vector enhanced luciferase activity seven- and fourfold when compared with enzyme levels measured in the presence of Lipofectamine alone or Lipofectamine and Lys(16) peptide alone, respectively (Table 1).
In this study, we have demonstrated use of a nonviral and highly efficient gene delivery system utilizing a nontoxic peptide vector that targets α5β1 and αvβ3 integrins expressed highly on proliferating VSMCs. The targeting moiety of the vector consists of a 15-amino acid peptide exhibiting a RGDNP motif that binds preferentially to α5β1 and αvβ3 integrins (22) and a DNA binding moiety, comprising 16 consecutive lysine residues, which promotes electrostatic binding of DNA. RGD is an integrin recognition sequence that is particularly effective in those adhesion proteins that bind to the β1, -3, and -5 family of integrins (11, 21). However, internalized DNA is targeted to lysosomes for destruction (7) and one important limiting step for receptor-mediated gene delivery is the exit of DNA from endosomes (19). In our study, this process was achieved by inclusion of the cationic lipid, Lipofectamine (14, 30), or a fusogenic peptide derived from the hemagglutinin of influenza virus (9, 19, 30).
A particular attribute of the molossin vector system is that it takes advantage of the TGF-β1 effect on VSMC integrin expression and proliferation such that a much higher transfection efficiency is achieved in TGF-β1 treated cells. The clinical relevance and importance of this finding is highlighted by the fact that TGF-β1 can be produced by almost every vascular cell in the vessel wall, and its expression is increased significantly after vessel injury and/or during other pathological conditions in which VSMC proliferation forms a part of the pathology (4, 18). Indeed, in a rat balloon injury model, high TGF-β1 expression levels were found in the injured vessel wall, which promoted restenosis after angioplasty (26), whereas local blockage of TGF-β1 using ribozyme oligonucleotides inhibited neointima formation after vascular injury (29). One important mechanism responsible for the effects of TGF-β1 on VSMC function and neointima formation is that it increases α5 and β3 integrin expressions (3, 12). Our study confirms that pretreatment of human VSMCs with TGF-β1 significantly increased cellular expression of α5β1 and αvβ3 integrins. Furthermore, we have shown that molossin vector targets α5β1 and αvβ3 integrins on proliferating human VSMCs. This has formed the foundation of our studies using the molossin vector for gene transfer into proliferating VSMCs after vessel injury. The concentration of TGF-β1 (1–20 ng/ml) necessary to increase molossin vector gene delivery is physiologically relevant. Thus circulating levels of TGF-β1 range from 2–10 ng/ml in healthy individuals (8). Moreover, at sites of vascular injury, in which vessel wall TGF-β1 synthesis is induced and activated platelets release TGF-β1, the local concentration of TGF-β1 may commonly reach the levels (>10 ng/ml) found to be optimal in this study.
In many restrictive vessel cases, a reactive cellular proliferative response leads to the regrowth of VSMCs locally that reduces lumen diameter and thereby compromises blood flow. This can occur at sites of anastomoses in allograft transplants and in bypass grafts and is a major component of restenosis and in-stent stenosis after angioplasty. A suitable nonviral and nontoxic vector that mimics the functions of viral proteins for DNA delivery into proliferating and migrating VSMCs would greatly improve the outcome for patients after such interventional therapy. With increases in the number of allograft transplantations and vessel bypass procedures, it would be advantageous to transfer the gene ex vivo into the isolated grafts before transferring to the recipient. Indeed, Mann and colleagues (16) recently transferred decoy oligonucleotides, which target the E2F transcription factor family, to veins ex vivo before grafting into patients. This approach led to a reduction in failure rates of this procedure and improved the patency of grafts and demonstrates the effectiveness of ex vivo gene therapy for certain vascular diseases. In the present study, we have shown that successful ex vivo gene transfer can be achieved with the molossin vector in rat aortic vessels after injury of the endothelium. Of particular interest was the observation that molossin vector bound efficiently to VSMCs in aortic vessels denuded of endothelium (Fig. 5B). Thus gene delivery occurs readily in proliferating VSMCs in areas of vessel injury, whereas delivery would likely be minimal or would not occur at all in uninjured vessels.
In summary, we have demonstrated the use of a nonviral gene delivery system for proliferating VSMCs both in vitro and ex vivo that is enhanced significantly in the presence of TGF-β1. It is possible that the combination of molossin vector with TGF-β1 might not be suitable for all in vitro transfection procedures (due to unwanted effects from the cytokine). In such cases, the use of integrin-activating agents, e.g., activating antibodies in place of TGF-β1, might prove more preferable. However, the current approach certainly is advantageous for ex vivo/in vivo gene delivery in areas in which TGF-β1 levels are raised naturally, e.g., in injured vessels. The system provides advantages over viral vectors including safety and a lack of any limit on the size of DNA insert that can be packaged. The whole system contains only three components: molossin vector, DNA, and fusogenic peptide or cationic lipid. All components are molecularly defined, easily undergo self-assembly into an active DNA delivery system and the whole procedure can be achieved within a short time frame, thereby facilitating the application of this technology within the operating theater during bypass graft or organ transplantation operations.
This study was supported, in part, by the British Heart Foundation, the National Heart Research Fund, and the Royal Society.
We thank John Fabre for helpful discussions.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2003 by the American Physiological Society