ARTICLE
doi:10.1006/mthe.2000.0053, available online at http://www.idealibrary.com on IDEAL
Optimization of Nonviral Gene Transfer of Vascular Smooth Muscle Cells in Vitro and in Vivo Sorin Armeanu,* Jaroslav Pelisek,* Eberhard Krausz,* Alexandra Fuchs,* Detlef Groth,† Rene Curth,† Oliver Keil,‡ Jacques Quilici,§ Pierre H. Rolland,§ Regina Reszka,† and Sigrid Nikol*,1 *Medical Clinic I, Klinikum Großhadern, Ludwig Maximilian University, Munich, Germany † Max-Delbru¨ck-Center for Molecular Medicine, Berlin, Germany ‡ Bergische Universita¨t, Wuppertal, Germany § Laboratoire Cardiovasculaire, Faculte´ de Me´de´cine, Universite´ de Marseille, France Received for publication January 28, 2000, and accepted in revised form March 8, 2000
Gene therapy strategies for the prevention of restenosis postangioplasty are promising. Nonviral gene transfer to the arterial wall in vivo has so far been limited by poor efficiency. This study aimed to optimize transfection of primary vascular smooth muscle cells using cationic nonviral formulations based on cholesterol derivates (DC-, DAC-, DCQ-, and Sp-Chol), double-chained amphiphils (LipofectAMINE, DOTMA, DOSGA, DOSPER, and DOCSPER), or heterogeneous reagents (Superfect, Effectene, and Tfx-50). Estimation of transfection efficiencies was performed using galactosidase assays at different ratios of transfection reagent to plasmid DNA with reporter gene. Toxicity was monitored by analyzing cell metabolism. Transfer efficiency and safety were determined in a porcine restenosis model for local gene therapy using morphometry, histology, galactosidase assays, and reverse-transcriptase polymerase chain reaction. The highest in vitro transfection efficiency was achieved using the recently developed DOCSPER liposomes, with transfer rates of at least 20% in vascular smooth muscle cells. Transfer efficiency was further enhanced up to 20% by complexing with poly-L-lysine. Transfection efficiency in vivo in a porcine restenosis model was up to 15% of adventitial cells using DOCSPER versus 0.1% using LipofectAMINE. Toxicity in vivo and in vitro was lowest using DOCSPER. Increased biological effects were demonstrated following optimization of transfer conditions. Key Words: liposomes; gene transfer; vascular smooth muscle cells; local drug delivery.
INTRODUCTION Positive long-term results of coronary angioplasty are limited by restenosis, the luminal renarrowing during the first 6 months after the procedure (1). Following angioplasty, vascular smooth muscle cells of the arterial wall leave the quiescent state and proliferate by progressing through the cell cycle (1). Smooth muscle cell proliferation and subsequent matrix formation make up the bulk of the restenotic narrowing, although a degree of remodeling may also contribute. Limiting smooth muscle proliferation has been shown to have a beneficial effect on restenosis (2, 3).
1 To whom correspondence should be addressed at Medical Department I, Klinikum Grohadern, Ludwig Maximilian University, D-81377 Munich, Germany. Fax: ⫹49 89 7095 6166. E-mail:
[email protected].
366
Restenosis develops in the arterial wall, which is relatively accessible at the time of angioplasty. This facilitates local deposition of active agents, which can be carried out through one of many local drug delivery catheters now available (4). One possibility is local perivascular delivery, providing a depot of active agent around the artery (5). Perivascular events have been shown to affect neointima: adventitial irritation results in focal intimal thickening and markers applied to the adventitia can be visualized in intima. Furthermore, it is hypothesized that cells originating from the adventitial layer migrate and contribute in large amounts to neointimal formation (6). Adventitial delivery may therefore be used to target agents that inhibit vascular smooth muscle cell proliferation and matrix formation, including gene therapy (1). Existing gene therapy protocols use several different vector systems: retroviruses (58%), adenoviruses (29%), adeno-associated viruses (5%), Herpes-derived viruses MOLECULAR THERAPY Vol. 1, No. 4, April 2000 Copyright © The American Society of Gene Therapy
ARTICLE TABLE 1 Transfection Reagents Tested Reagents
Name
DOPEa DOTMAb DOSGAb DOSPAc Lipofectinb LipofectAMINEc DCc DACc DCQc DOSPERc DOCSPERc DOCSPER–MPC Spc Tfx-50c Superfect Effectene
Dioleoylphosphatidylethanolamine N-[1-(2,3-Dioleoyloxy)-propyl]-N,N,N-trimethyl-ammonium chloride 1,3-Dioleoyloxy 2-[-alanyl-guanidyl]-propylamid 2,3-Dioleyloxy-(N-(2-sperminecarbxamido)ethyl)-N,N-dimethyl-propanaminium trifluoracetate DOTMA/DOPE DOSPA/DOPE 3:1 (w/w) Cholesteryl 3-[N-(N⬘,N⬘-dimethyl)-diaminoethyl]-carbamate Cholesteryl 3-[N-(N⬘,N⬘-dimethyl)-diaminoethyl]-carbamate Cholesteryl 3-[N-(N⬘,N⬘-dimethyl)-N-hydroxyethyl-diaminoethyl]-carbamate 1,3-Di-oleoyloxy-2(6-carboxy-spermyl)-propylamid Cholesteryl 3-[N-(N⬘,N⬘-dimethyl)-N-hydroxyethyl-diaminoethyl]-carbamate DOCSPER and mannosephosphate-cholesteryl (Dicarbobenzoxy-spermine-carbamoyl)-cholesterol N,N,N⬘,N⬘-Tetramethyl-N,N⬘-bis(2-hydroxyethyl)-2,3-dioleoyloxy-1,4-butanediammonium iodide/DOPE Polycationic dendrimer Two components nonliposomal
a
Neutral lipid. Monocationic lipid. c Polycationic lipid. b
(1%), and nonviral systems, including liposomal gene transfer (7%) (7). The most efficient methods of gene transfer today involve the use of engineered viruses, with therapeutic genes inserted at the site of viral genetic information. Such recombinant viruses cannot replicate and harm patients once they have been delivered. Investigations in animal models for gene therapy to prevent restenosis have so far demonstrated that the most powerful vector systems use recombinant adenoviruses, with transfer rates shown to be as high as 20 –30% into intimal smooth muscle cells in a porcine restenosis model (8, 9). However, adenoviral gene transfer in patients can be complicated by pronounced immune responses against the adenoviral vector, reducing transfection efficiency to as low as 5% (10, 11). Complexing recombinant DNA with cationic liposomes to introduce foreign genes into cells (termed transfection or lipofection) avoids some of the disadvantages of viral gene transfer and may be safer and easier to handle in clinical practice The mechanism of lipofection is not completely understood, but it is recognized that the degree of successful gene transfer is highly dependent on the cationic lipid type, liposomal formulation, and cell type (12). Cationic vectors bind to negatively charged DNA, resulting in a condensation reaction and the formation of stable complexes. Fusion with the cell membrane or endocytosis allows incorporation of DNA into the cell. Recent reports suggest that new liposomal formulations, individually optimized for the target tissue, with better protocols and/or continuously administered (poly-)cationic liposomes may substantially increase transfection efficiencies (13, 14). We screened nonviral formulations by transfecting primary vascular smooth muscle cells in vitro using different concentrations and combinaMOLECULAR THERAPY Vol. 1, No. 4, April 2000 Copyright © The American Society of Gene Therapy
tions of nonviral vectors as well as varying DNA/vector ratios and adjuvants. Thus, 16 vectors and five combinations of liposomes with cholesterol were investigated in vitro to optimize gene transfer. We then used the three most effective vector formulation, adjuvant, and DNA/ vector ratios to transfect vascular cells in a porcine restenosis model in vivo. In addition, the needle injection catheter was used to ensure local adventitial delivery without any endoluminal loss.
MATERIALS
AND
METHODS
Plasmids. The -galactosidase reporter gene vectors pCMV (7.2 kb, CMVdriven -galactosidase gene, Clontech, Heidelberg, Germany) and pBAG [12 kb, retroviral vector (15)] were used for transfection. For in vivo experiments, pre-pro-cecropin A cDNA from Hyalophora cecropia (GenBank Accession No. X06672) was cloned into the pTRACER-SV40 vector (Invitrogen, Groningen, The Netherlands) (2). Plasmids were isolated with an endotoxin-free kit (Qiagen, Hilden Germany). Transfection reagents. Several commercially available transfection reagents were tested (Table 1): DOSPER and FuGENE 6 (Roche, Mannheim, Germany), Lipofectin and LipofectAMINE (GIBCO BRL, Eggenstein, Germany), Tfx-50 (Promega, Mannheim, Germany), and Superfect and Effectene (QIAGEN, Hilden Germany). Additionally, different formulations were prepared using cationic lipids listed in Table 1, each mixed with different amounts of neutral lipid DOPE (Roche). DOCSPER was also formulated with mannose covalently linked to cholesterol (DOCSPER– MPC) which may mediate an uptake of DNA–vector complexes via the cell-surface mannose receptor (14). Transfection in vitro. Primary vascular smooth muscle cells derived from porcine aortic vessels and human vein grafts were used in the first four passages (2). Prior to transfection, 20,000 cells were seeded for 24 h into 96-well cell culture plates in 100 l DMEM/20% FCS. Transfection was performed with pCMV unless otherwise noted. For transfection, 0.1, 0.33, or 1 g plasmid DNA or 1 g DNA complexed with 0.5 g poly-Llysine (MW 25,000 Da, Sigma, Munich, Germany) was mixed with 10, 5, 2.5, 1.3, 0.6, or 0.3 g transfection reagent in 100 l serum-free OptiMEM1
367
ARTICLE TABLE 2 Primers Used for RT-PCR and PCR Specificity
Forward primer
Reverse primer
PCR product (bp)
pCMV pBAG -Gal gene GAPDH gene
TCTAGAGGATCCGGTACTCGAG GGTGGAACTGACGAGTTCGGAACAC GCATCATCCTCTGCATGGTCAG CGTCTTCACCACCATGGAGAAGGC
GACCGTAATGGGATAGGTTAC GACCGTAATGGGATAGGTTAC AAACCGCCAAGACTGTTACCCATCG AAGGCCATGCCAGTGAGCTTCCC
647 619 572 398
medium (GIBCO BRL, Eggersheim, Germany). After 4 h, 100 l DMEM/ 40% FCS was added to the transfection mix and cells were incubated for an additional 48 h. Cells were either stained with X-gal to estimate transfection efficiencies or lysed for a quantitative ONPG -galactosidase assay, as described previously (2). To evaluate the toxicity of reagents used, the CellTiter method for determining the number of viable cells with tetrazolium compound was performed (Promega, Heidelberg, Germany). Cultures were transfected with 0.33 g DNA complexed with 1 to 30 g transfection reagents. After 48 h, medium was changed and the tetrazolium compound was added. Optical density at 540 nm with a reference at 690 nm was measured every 30 min. The relative cell viability was calculated as relative optical density to a culture grown without transfection. Transfection in vivo. All animal procedures were approved by the Animal Care and Use Review Committee of the University of Marseille according to French guidelines. Study animals were Pietrain pigs (average weight 35 kg at initial procedure; n ⫽ 27). Anesthesia and balloon angioplasty in iliac arteries were performed as previously described (2, 16). Balloon catheters were introduced via six French intraarterial sheaths into femoral arteries and balloons of 4-cm length were inflated three times at 8⫻ atmosphere for 1 min. After deflation and withdrawal of the balloon from each femoral artery of study animals, 30 g plasmid DNA with DOCSPER, LipofectAMINE, or Effectene in 2 ml serum-free DMEM was deposited into two experimental segments using a needle injection catheter (BMT, Germany) as described (2, 16). Sacrifice took place 7 days after transfection, using the same anesthetic protocol, including orotracheal intubation and mechanical ventilation. Control vessels were removed before arteries were treated with gene therapy using RNase-free instruments. Vessels were cross-sectioned starting at the proximal site in 0.5-cm segments and each second segment was embedded in OCT compound (Miles Scientific, Elkhart, IN); other samples were snap-frozen. Blood samples and noninjured carotid arteries, skeletal muscle adjacent to the treated vessel, and organs (liver, lung, spleen, heart, small intestine, and kidney) were obtained for safety analysis and demonstration of plasmid retention at 7 days. Detection of transferred plasmid DNA. Snap-frozen tissue was homogenized in liquid nitrogen and DNA prepared with the QIAamp Tissue DNA extraction kit (Qiagen). Plasmid DNA was detected by polymerase chain reaction (PCR) amplifying a partial sequence of the plasmid construct with the specific primer (Table 2, synthesized by MWG Biothech, Eggersheim, Germany) for pCMV or pBAG in 35 cycles with an annealing temperature of 60°C. Each PCR assay contained 0.1 g total DNA, 5 pmol of each primer, and 0.5 units Taq DNA polymerase. Transfer of plasmid DNA was analyzed in treated and untreated arteries and potential spread of the plasmid into other organs was evaluated by similar analyses. The molecular weight marker VIII was a mixture of pUC21 cleaved with HpaII and a HindIII/DraI digest of pUC21 for the lower molecular weight range. Detection of expressed reporter gene. Reverse-transcription PCR analyses were performed to detect specific mRNA from the -galactosidase genes. Snap-frozen tissue was homogenized in liquid nitrogen and RNA was prepared with RNAzol B reagent (CINNA/MRC, Cincinnati, OH). Contaminating DNA was digested with DNase (Roche) and RNA was precipitated with ethanol. One to five micrograms of total RNA was reverse transcribed with 200 units of M-MLV reverse transcriptase, 10 mM each dNTP, and 25 units ribonuclease inhibitor rRNasin (Promega, Heidelberg, Germany). PCR was performed with 35 cycles with 5 pmol of each specific primer pair for the -galactosidase genes (Table 2) using 500 ng reverse-transcribed
368
total RNA as template. The quality of transcribed RNA was shown by amplification of the 5⬘ end of the GAPDH gene transcript (Table 2). Specificity of amplified fragments was verified using restriction pattern analysis. Protein extract was prepared from snap-frozen tissue in lysis buffer (-galactosidase assay, Promega, Munich, Germany). -Galactosidase activity was detected in samples of 100 g protein extract with the CPRG substrate (17). Toxicity in vivo. Sections were stained with either hematoxylin and eosin or elastic trichrome and examined by light microscopy at various magnifications (4 to 500⫻). The toxic effect of DNA/transfection reagent complexes was evaluated by differential counting of leukocytes in the arterial wall using hematoxylin- and eosin-stained tissue sections. Fibrosis was visualized by elastic trichrome staining and quantified by morphometric analysis. Morphometric analysis. Morphometry data were obtained by analyzing five sections of each injury site in triplicate, at least 1 mm apart. Thus, 360 fields were examined from 12 therapeutic gene-transfected and 12 control gene-transfected arteries from a total of six experimental segments. Computer-assisted morphometry by measuring areas was performed at 4⫻ magnification with a microscope linked to a charge-coupled device camera connected to computer (CCD, Nikon 104, Du¨ sseldorf, Germany). Digitized pictures were analyzed with the computer program Adobe Photoshop (Adobe Systems Inc., Seattle, WA). Fibrotic areas were related to the corresponding area of medium and compared in arteries treated with nonviral vectors at different concentrations. Neointimal areas were compared in arteries treated with therapeutic and control genes. Statistical analysis. Results are expressed as means ⫾ SEM for each artery. Comparisons use a one-way analysis of variance with Scheffe’s F test. Statistical significance was set at P ⬍ 0.05.
RESULTS Lipofection efficiencies in vitro. The optimal transfection efficiency of 16 nonviral vector formulations into vascular smooth muscle cells was determined by titration, using different amounts of plasmid DNA at different vector/ DNA ratios (Fig. 1). Transfection efficiencies ranged from 0.1 to 20% of transfected vascular smooth muscle cells. The recently developed polycationic spermin derivate DOCSPER (14) demonstrated the best efficiency, obtained without adding neutral lipid under defined conditions (DOCSPER100, Fig. 2A). A 50:50 formulation of DOCSPER and DOPE revealed a much lower efficiency (data not shown). The polycationic spermine derivative Sp-Chol 20 formulated with 80% DOPE (Fig. 2B) and the cholesterol derivative containing a quaternary ammonium group, DCQ-Chol 25, formulated with 75% DOPE (w/w) showed significantly lower transfer efficiencies. The widely used 50:50 (w/w) DOTMA/DOPE (Fig. 2C) formulation was as effective as Sp-Chol 20 and DCQ-Chol 25. Formulation of liposomes with higher cationic lipid amount did not reMOLECULAR THERAPY Vol. 1, No. 4, April 2000 Copyright © The American Society of Gene Therapy
ARTICLE
FIG. 1. Profile of transfection efficiencies over increasing transfection reagent/DNA ratios. Transfection reagent/DNA ratios from 1 to 30 using pCMV plasmid shown for DOCSPER100, DOTMA50, Superfect, and DOSGA25. For transfection, 0.1, 0.33, or 1 g plasmid DNA or 1 g DNA complexed with 0.5 g poly-L-lysine was complexed with 10, 5, 2.5, 1.3, 0.6, or 0.3 l (1 g/l) vector. The transfection efficiency was proportional to the activity of -galactosidase in milliunits (mU) per well. Error bars represent SEM of triplicate experiments.
sult in higher efficiencies for Sp-Chol, DC-Chol, DCQChol, and DOSGA. In contrast, doubling the amount of the cationic lipid DAC-Chol (50%) increased the transfer efficiency (Table 3). Control experiments using plasmid DNA without the addition of vectors resulted in transfer rates below detection limits (Fig. 2D). In an additional experiment, 1 g DNA was mixed with 0.5 g poly-L-lysine before addition of vectors for transfection. Enhanced gene transfer efficiency was observed for DOCSPER and with the formulations containing a lower percentage of cationic lipids, such as DC-Chol 25, DCQ-Chol 25, or DAC-Chol 25, which contain 75% (w/w) neutral lipid DOPE (Fig. 1). A growing number of transfection reagents, mostly polycationic vectors, have been developed for optimal efficiency in commonly used tumor cell lines. On primary porcine vascular smooth muscle cells transfer efficiencies of DOSPER (Boehringer Mannheim), Lipofectin and LipofectAMINE (GIBCO BRL), Tfx-50 (Promega), and the dendrimer Superfect and Effectene (QIAGEN) were compared with the newly developed DOCSPER100 reagent (Table 3). MOLECULAR THERAPY Vol. 1, No. 4, April 2000 Copyright © The American Society of Gene Therapy
Although the monocationic Lipofectin demonstrated very low efficiency with vascular smooth muscle cells, the polycationic reagents LipofectAMINE and DOSPER as well as the Superfect reagent provided higher efficiencies (Table 3). None reached the efficiency of the DOCSPER reagent previously tested on porcine vascular smooth muscle cells. For comparison, these reagents were also tested in human vascular smooth muscle cells. Overall, the transfection efficiency was reduced to 1% of the transfection observed in porcine vascular smooth muscle cells (Table 4). Toxicity in vitro. Toxicity was estimated by assessing the grade of cell confluence. Optimal concentrations of DC, DAC, and DOSGA which led to the highest transfection efficiencies demonstrated high toxicity on vascular smooth muscle cells, with a cell confluence below 50%. In contrast, Sp-Chol and DOCSPER were well tolerated by cells and cell densities of up to 90% were measured (data not shown). For DOCSPER, LipofectAMINE, and Effectene, relative viability of cells compared with untreated cells was determined using the CellTiter test. DOCSPER
369
ARTICLE
FIG. 2. Transfection efficiencies using different liposomes under optimal conditions. Visualization by X-gal staining of transfected primary vascular smooth muscle cells: (A) DOCSPER100, (B) Sp-Chol 20, (C) DOTMA50, (D) control. Bar represents 100 m.
MOLECULAR THERAPY Vol. 1, No. 4, April 2000 Copyright © The American Society of Gene Therapy
370
ARTICLE TABLE 3 Highest Transfection Efficiencies in Vascular Smooth Muscle Cells Evaluated by -Galactosidase Activity Assays in Comparison with the Maximum Value Measured for DOCSPER 100 (Set at 1.00)
Nonviral vector
Relative to DOCSPER100
DC25 DC50 DAC25 DAC50 DCQ25 DCQ50 Sp20 Sp30 DOTMA50 Superfect100 DOCSPER50 DOCSPER100 DOCSPER50–MPC DOSGA25 DOSGA75
0.16 0.13 0.13 0.23 0.36 0.15 0.46 0.24 0.35 0.34 0.23 1.00 0.69 0.20 0.17
Tfx-50 Lipofectin LipofectAMINE Superfect Effectene DOSPER Control/no vector
0.35 0.16 0.73 0.76 0.62 0.83 0.04
Conditions of transfection (g vector/g pCMV/ g poly-L-lysine) 2.5/1.0/0.5 0.6/0.33/0 1.3/1.0/0.5 1.3/0.33/0 2.5/1.0/0.5 10/1.0/0 1.3/0.1/0 5/1.0/0 1.3/0.33/0 5/0.33/0 10/1.0/0 5/1.0/0.5 10/1.0/0 2.5/0.1/0 2.5/0.33/0 5/1.0/0 5/0.33/0 10/1.0/0 5/1.0/0 5/0.33/0 10/1.0/0 5/0.33/0 0/1.0/0
Note. Numbers following the vector abbreviations indicate the percentages of cationic lipid in neutral lipid DOPE.
and Effectene were well tolerated up to 20 g transfection reagent/ml in contrast to LipofectAMINE, which demonstrated a reduced relative viability at this concentration (Fig. 3). DOCSPER in combination with 50% of the neutral lipid DOPE was well tolerated at concentrations higher than 40 g/ml. The vector concentration at 50% cell viability (LD50) was calculated as a parameter for toxicity using Fig. 3a. Significantly less toxicity was observed for DOCSPER (LD50 ⫽ 33 g/ml) compared to Effectene (LD50 ⫽ 25.8 g/ml) or LipofectAMINE (LD50 ⫽ 17.5 g/ml) (P ⬍ 0.05). Transfection efficiencies in vivo. DOCSPER, LipofectAMINE, and Effectene were selected from the in vitro studies for comparison of transfection efficiencies in a porcine restenosis model. Seven days after local application of transfection complexes into femoral arteries, plasmid DNA was detected using PCR in distinct arterial segments (Fig. 4A). Reporter gene expression was detected in those arterial segments by reverse transcription and PCR (Fig. 4B). Transfection efficiencies of pCMV complexed with DOCSPER or Effectene were five times higher than with LipofectAMINE as assayed by -galactosidase activity, which measured 1 mU/100 g protein (Fig. 4C). X-gal staining on cryostat slides revealed that up to 15% of cells MOLECULAR THERAPY Vol. 1, No. 4, April 2000 Copyright © The American Society of Gene Therapy
in the adventitia were successfully transfected using pCMV/DOCSPER complexes (Fig. 5). Toxicity in vivo. Toxicity was assessed by measuring the fibrotic areas after arterial injury by balloon angioplasty and gene transfer. After transfection with LipofectAMINE, DOCSPER, and Effectene in the optimal ratios with 3, 30, or 150 g plasmid, the size of the fibrotic area was not significantly different from that in control balloon-injured arteries where there was no transfection. Seven days following balloon dilatation, the mean neointima/medium ratio was 0.35 (⫾0.1). The number of infiltrating eosinophilic cells was increased in arteries treated with LipofectAMINE (32 ⫾ 5/tissue section) and Effectene (21 ⫾ 8), but not with DOCSPER (4 ⫾ 2), which was used at up to 665 g per application site without increasing the number of local inflammatory cells. Transfection in vivo with a therapeutic gene. LipofectAMINE, Effectene, and DOCSPER were compared in the pig balloon angioplasty model using plasmids coding for the antibacterial peptide cecropin A. The therapeutic effect was quantified by morphometric analysis by measuring the extent of neointima 7 days after angioplasty. Transfer of 30-g plasmids carrying the gene encoding for the antimicrobial peptide pre-pro-cecropin A reduced neointima formation in vivo. The amount of measured neointima depended on the transfection reagent used. Neointimal areas were smaller following gene transfer of pTRACER-CA combined with DOCSPER (0.20 ⫾ 0.01 mm2) than following the use of Effectene (0.25 ⫾ 0.02 mm2) or LipofectAMINE (0.28 ⫾ 0.03 mm2) as nonviral vector. Neointimal areas were also smaller when compared to control transfection using pCMV/DOCSPER as nontherapy control (0.44 ⫾ 0.06 mm2). There was statistical significance (P ⬍ 0.05) between neointimal areas of therapy and nontherapy arteries. In addition, there was statistical significance (P ⬍ 0.05) between neointimal
TABLE 4 Transfection Efficiencies in Porcine and Human Vascular Smooth Muscle Cells Using Different Vectors Evaluated by -Galactosidase Activity Assays pVSMCs
hVSMCs
Transfection reagent
mU BG/well
⫾SEM
mU BG/well
⫾SEM
Sp-Chol 20 DOCSPER100 DOSPER Tfx-50 LipofectAMINE Lipofectin Superfect
11.0 37.0 32.3 13.7 24.3 11.7 34.0
1.00 1.16 0.88 0.33 3.48 1.67 3.79
0.57 1.8 1.2 0.26 0.33 0.28 0.33
0.20 0.1 0.06 0.04 0.06 0.05 0.04
Note. pVSMCs, porcine vascular smooth muscle cells; hVSMCs, human vascular smooth muscle cells; mU BG, activity of -galactosidase; SEM, standard error of the mean of triplicate experiments.
371
ARTICLE
FIG. 3. Toxicity of transfection reagents. Toxicity of DOCSPER (D100 or D50), LipofectAMINE (LiA), and Effectene (Eff) was evaluated by incubation of cells with transfection reagents at increasing concentrations. Cell numbers (A) or metabolic derivation of tetrazolium compound (B) was evaluated after 48 h in triplicate experiments for vectors concentrations between 0 and 40 g/ml. The reagent concentrations at 50% cell viability was calculated (LD50). The high LD50 for DOCSPER revealed less toxic effects on cells than the other vectors tested (P ⬍ 0.05).
areas following transfection with DOCSPER compared to the use of the other two vectors as transfection reagents.
DISCUSSION Gene transfer using cationic lipid/plasmid complexes has become a widely used strategy for gene transfer and is being tested preclinically as well as in several clinical trials (7, 12). However, the benefits and limitations of each transfer system depend on reliable, safe, and efficient gene transfer. For the commercially available transfection reagents and newer nonviral vectors that we tested, the transfection efficiencies in vitro in primary vascular smooth muscle cells ranged from 0.1 to 20%. Published results have been different in other cell types and cell
372
lines; thus, each reagent apparently has cell-type-specific and additional species-specific characteristics (18). In addition, conditions of use may need to be optimized for each laboratory. Commercially available reagents are optimized for commonly used tumor cell lines and show highest efficiencies with these lines. Clearly, these may not be reproducible in primary cultures. Vascular smooth muscle cells seem to be more difficult to transfect compared to other cell types. For example, Pickering et al. (19) demonstrated a 30-fold lower efficiency using DOTAP on vascular smooth muscle cells in comparison with NIH 3T3 cells, which can probably be attributed to a lower proliferation rate in the vascular smooth muscle cells. Takeshita et al. demonstrated that Lipofectin-mediated transfection yields in vascular smooth muscle cells could be improved MOLECULAR THERAPY Vol. 1, No. 4, April 2000 Copyright © The American Society of Gene Therapy
ARTICLE
FIG. 4. Detection of plasmid DNA and gene expression in homogenates of transfected arterial tissues. Total DNA and RNA were isolated from consecutive frozen arterial segments 7 days following gene transfer. For PCR and RT-PCR 0.5 g of total DNA and RNA prepared from arterial samples was taken as template. PCR was performed using specific primers for plasmid pCMV (A). RT-PCR was performed using primers for the transfected reporter gene for -galactosidase (B). Persistent DNA plasmids, transcripts, and -galactosidase activity (C) were found at different degrees in all analyzed arteries following transfection. S1–S9, analyzed segments of transfected arteries; K1, nontransfected artery (negative controls); K2, pCMV (10⫺3 ng) (positive control); E1–E3, endogenous -galactosidase activity in nontransfected artery; error bars represent SEM of triplicate measurements.
7- to 13-fold if performed 3 to 7 days postangioplasty in both organ culture and in vivo (20), when proliferation is likely to be maximal (approximately 30% of vascular smooth muscle cells are in a proliferative state in this period). However, transfection efficiency at this point was still less than 1%. Furthermore, a second catheter-based procedure for lipofection several days following balloon angioplasty will not be acceptable in clinical practice, although strategies with depot gene therapy may be able to target this time point (2). Even relatively poor efficiencies, however, may be sufficient to result in successful secretion of proteins with measurable effects in vivo, including the transfer of platelet-derived growth factor (PDGF) (21), fibroblast growth factor (FGF-1) (22), transforming growth factor-1 (TGF1) (23), nitric oxide synthetase (NOS) (24), cecropin (2), or prodrug-activating enzymes, such as HSV–thymidine kinase (25). However, for gene transfer of proteins where the therapy aims to influence the cell cycle by specific inhibitors or by other antiproliferative strategies and where individual transduced cells need to be targeted, a much higher efficiency is needed for an effect to be observed (16). Transfer conditions need to be evaluated carefully, especially within the narrow window for efficient transfection using DOCSPER. DOCSPER demonstrated optimal transfection efficiencies at 5 g liposomes/1 g plasmid. In combination with the neutral lipid DOPE, transfer efficiency was only 60% compared with DOCSPER alone, although this was under a wider range of transfer conditions, e.g., lipid and DNA ratios. In contrast, DOTMA and DCQ-Chol demonstrated a broader range of possible transfer conditions but had a lower level of transfection efficiency. Reagents such as LipofectAMINE and Effectene, which consist of two components, still allow at least 80% of reporter gene activity if optimal DNA/transfection reagent ratios are varied by 50% or more. MOLECULAR THERAPY Vol. 1, No. 4, April 2000 Copyright © The American Society of Gene Therapy
Controlled cellular access and uptake, intracellular trafficking, and nuclear retention of plasmids must also be achieved to reach optimal gene transfer. To that end, peptide may be incorporated into multicomponent lipoplexes for DNA condensation, cell-specific targeting, endosomolysis, or nuclear transport (26). Poly-L-lysine condensates DNA and thus enhances transfection efficiency of some nonviral vector formulations. In our hands, the poly-L-lysine fraction of 400,000 kDa increased transfection efficiencies, but it also had a toxic effect on vascular smooth muscle cells. Regardless of the reagent used, human vascular smooth muscle cells showed a much lower transfer efficiency than porcine vascular smooth muscle cells (about 20-fold less for DOCSPER). Similar differences were reported previously for rabbit and rat smooth muscle cells (18, 19). The human vascular smooth muscle cells used in our experiments were derived from an atherectomy specimen of a 70-year-old patient. The cells appeared to proliferate much more slowly than the cells in primary culture from younger animals, which are usually used for in vitro studies and which are similar to the age of animals in in vivo studies. Experiments demonstrated that the conditions evaluated for porcine vascular smooth muscle cells did not differ from those evaluated for human vascular smooth muscle cells for at least seven formulations. Thus, conditions of use may be transferred from one species to another for the same cell type. In contrast, conditions evaluated for a reference cell line of human hepatocellular carcinoma (HepG2) were completely different regarding both lipid and conditions used (data not shown). Results obtained from cell lines cannot be generalized to primary cell cultures. The mechanism of enhancement of gene transfection rates using DOCSPER is not well understood. In general, nonviral vectors facilitate gene transfer in different ways even in the same cell type, due to complex extracellular
373
ARTICLE
FIG. 5. Detection of gene expression of transfected arterial tissues in situ. In situ -galactosidase staining was performed on tissue sections of arteries prepared from pigs 7 days following gene transfer using (A) pCMV/DOCSPER, (B) pCMV/LipofectAMINE, (C) pCMV/Effectene, and (D) pTRACER-CA/DOCSPER as control under optimal transfection conditions (Table 3). Sections were counterstained with hematoxylin. Blue cells were detected mostly in adventitial arterial wall layers. Arrows delineates the media layer. Bar represents 100 m.
and intracellular events involved during the transfection process. Membrane binding, internalization, endosomal release, uncoating, and nuclear translocation constitute basic cellular events during nonviral gene transfer and may be involved at different degrees depending on the vector used (12). Attempts to perform in vivo arterial gene transfer using nonviral vectors have been compromised by a low transfection efficiency of vascular smooth muscle cells in the case of the earlier generation of liposomes (monocationic), i.e., DOTAP/Lipofectin (8, 20, 27, 28). Less than 1% of cells were successfully transfected in porcine iliofemoral arteries (2, 16) and less than 0.1% in a rabbit model using Lipofectin (27). With polycationic liposomes such as DOCSPER, up to 15% of adventitial cells were transfected following perivascular delivery. The transfection efficiencies in cultures of proliferating vascular smooth muscle cells were higher than those observed in vivo. The in vivo transfection efficiency may be reduced by the low accessibility of the cell membrane for the nonviral
374
vector complex or by the low permeability of the nucleus for the plasmid DNA. Interstitial proteins and the highly organized extracellular matrix in the artery wall may bind vector/DNA complexes by electrostatic forces. The low proliferation rate of vascular cells compared to those in cell cultures reduces the probability for uptake of foreign DNA into the nucleus. Two ways to improve transfer efficiency and expression of the transfected gene are nuclear targeting of plasmid DNA to nondividing cells and the use of local gene delivery for higher concentrations of DNA/vector complexes compared to systemic delivery. Proliferation of cultured cells can be negatively affected by cationic liposomes through activation of second-messenger pathways, and cellular metabolism may be affected by incorporation of non-biodegradable lipids into biological membranes (29). DOCSPER is well tolerated because of its biodegradability (14) and provides high transfection efficiency in vitro. In contrast to LipofectAMINE, a mixture of the cationic lipid DOSPA and the neutral lipid DOPE, DOCSPER has an enzyme-accessible glycerol moiMOLECULAR THERAPY Vol. 1, No. 4, April 2000 Copyright © The American Society of Gene Therapy
ARTICLE ety. In combination with DOPE, DOCSPER had no impact on cell proliferation at concentrations over 40 g/ml. Toxic effects described for systemic applications in vivo, including hemolysis, complement activation, and platelet aggregation, may be a result of reduced complexing of DNA. Angioplasty initiates a number of responses in the vessel wall including cellular migration, proliferation, and matrix accumulation, as well as apoptosis of vascular smooth muscle cells, inflammatory cells, and adventitial fibroblasts (30). Use of DOCSPER demonstrated a minimal increase in eosinophilic infiltration, but no resultant increase in fibrosis. Neointima may be reduced by the transfer of genes encoding proteins with antiproliferative effects, such as cecropin A, an antimicrobial peptide. These have antiproliferative properties in mammalian tumor cells (31) and on vascular smooth muscle cells (2). High transfer efficiency using the optimized nonviral gene transfer strategy with DOCSPER increased the therapeutic effect of cecropin A in a porcine restenosis model compared to that with LipofectAMINE. In conclusion, up to 20-fold higher transfer efficiencies can be achieved in vascular smooth muscle cells in vivo when transfer conditions for nonviral gene transfer are optimized. This includes the use of newer generation liposomes with high efficiency and low toxicity. Thus, it may be possible to circumvent viral gene transfer in the cardiovascular field to achieve valid therapeutic effects in vascular gene therapy. ACKNOWLEDGMENTS This work was in part funded by the Deutsche Forschungsgemeinschaft (NI 331/3-2). The authors thank Tanya Huehns, M.D., for help in preparing the manuscript and gratefully acknowledge the technical assistance of Brigitte Leitermeier and Tanja Pamp.
REFERENCES 1 Nikol, S., Huehns, T. Y., and Hofling, B. (1996). Molecular biology and post-angioplasty restenosis. Atherosclerosis 123: 17–31. 2 Nikol, S., Huehns, T. Y., Kraausz, E., Armeanu, S., Engelmann, M. G., Winder, D., Salmons, B., and Ho¨fling, B. (1999). Needle injection catheter delivery of the gene for an antibacterial agent inhibits neointimal formation. Gene Ther. 6: 737–748. 3 Bauters, C., and Isner, J. M. (1997). The biology of restenosis. Prog. Cardiovasc. Dis. 40: 107–116. 4 Hofling, B., and Huehns, T. Y. (1995). Intravascular local drug delivery after angioplasty. Eur. Heart J. 16: 437– 440. 5 Huehns, T. Y., Gonschior, P., and Ho¨fling, B. (1996). Adventitia as a target for intravascular local drug delivery. Heart 75: 437– 438. 6 Shi, Y., Piniek, M., Fard, A., et al. (1996). Adventitial remodeling after coronary arterial injury. Circulation 93: 340 –348. 7 Wiley, J. (1999). The Journal of Gene Medicine. Available: http://www.wiley.co.uk/ genetherapy [online], September. 8 Laitinen, M., Pakkanen, T., Donetti, E., Baetta, R., Luoma, J., Lehtolainen, P., Viita, H., Agrawal, R., Miyanohara, A., Friedmann, T., et al. (1997). Gene transfer into the carotid artery using an adventitial collar: Comparison of the effectiveness of the plasmid–liposome complexes, retroviruses, pseudotyped retroviruses, and adenoviruses. Hum. Gene Ther. 8: 1645–1650.
MOLECULAR THERAPY Vol. 1, No. 4, April 2000 Copyright © The American Society of Gene Therapy
9 Yang, Z. Y., Simari, R. D., Perkins, N. D., San, H., Gordon, D., Nabel, G. J., and Nabel, E. G. R. A. (1996). Role of the p21 cyclin dependent kinase inhibitor in limiting intimal cell proliferation in response to arterial injury. Proc. Natl. Acad. Sci. USA 93: 7905–7910. 10 Kass-Eisler, A., Leinwand, L., Gall, J., Bloom, B., and Falck Pedersen, E. (1996). Circumventing the immune response to adenovirus mediated gene therapy. Gene Ther. 3: 154 – 162. 11 Laitinen, M., and YlaHerttuala, S. (1998). Adventitial gene transfer to arterial wall. Pharmacol. Res. 37: 251–254. 12 Clark, P. R., and Hersh, E. M. (1999). Cationic lipid-mediated gene transfer: Current concepts. Curr. Opin. Mol. Ther. 1: 158 –176. 13 Kichler, A., Zauner, W., Ogris, M., and Wagner, E. (1998). Influence of the DNA complexation medium on the transfection efficiency of lipospermine/DNA particles. GeneTher. 5: 855– 860. 14 Groth, D., Keilb, O., Lehmana, C., Schneider, M., and Rudo, M. (1998). Preparation and characterisation of a new lipospermine for gene delivery into various cells. Int. J. Pharm. 162: 143–157. 15 Price, J., Turner, D., and Cepko, C. (1987). Lineage analysis in the vertebrate nervous system by retrovirus-mediated gene transfer. Proc. Natl. Acad. Sci. USA 84: 156 –160. 16 Huehns, T. Y., Krausz, E., Mrochen, S., Schmid, M., Engelmann, M. G., Esin, S., Schrittenloher, P.-K., Ho¨fling, B., Gu¨nzburg, W. H., and Nikol, S. (1999). Neointimal growth can be influenced by local adventitial gene delivery via a needle injection catheter. Athersclerosis 144: 135–150. 17 Groth, D., Keil, O., Schneider, M., and Reszka, R. (1998). Transfection assay for dual determination of toxicity and gene expression. Anal. Biochem. 258: 141–143. 18 Keogh, M. C., Chen, D., Lupu, F., Shaper, N., Schmitt, J. F., Kakkar, V. V., and Lemoine, N. (1997). High efficiency reporter gene transfection of vascular tissue in vitro and in vivo using a cationic lipid DNA complex. Gene Ther. 4: 162–171. 19 Pickering, J. G., Jekanowski, J., Weir, L., Takeshita, S., Losordo, D. W., and Isner, J. M. (1994). Liposome-mediated gene transfer into human vascular smooth muscle cells. Circulation 89: 13–21. 20 Takeshita, S., Gal, D., Leclerc, G., Pickering, J. G., Riessen, R., Weir, L., and Isner, J. M. (1994). Increased gene expression after liposome-mediated arterial gene transfer associated with intimal smooth muscle cell proliferation. In vitro and in vivo findings in a rabbit model of vascular injury. J. Clin. Invest. 93: 652– 661. 21 Nabel, E. G., Yang, Z., Liptay, S., San, H., Gordon, D., Haudenschild, C. C., and Nabel, G. J. (1993). Recombinant platelet-derived growth factor B gene expression in porcine arteries induces intimal hyperplasia in vivo. J. Clin. Invest. 91: 1822–1829. 22 Nabel, E. G., Yang, Z. Y., Plautz, G., Forough, R., Zhan, X., Haudenschild, C. C., Maciag, T., and Nabel, G. J. (1993). Recombinant fibroblast growth factor-1 promotes intimal hyperplasia and angiogenesis in arteries in vivo. Nature 362: 844 – 846. 23 Nabel, E. G., Shum, L., Pompili, V. J., Yang, Z. Y., San, H., Shu, H. B., Liptay, S., Gold, L., Gordon, D., Derynck, R., et al. (1993). Direct transfer of transforming growth factor beta 1 gene into arteries stimulates fibrocellular hyperplasia. Proc. Natl. Acad. Sci. USA 90: 10759 –10763. 24 von der Leyen, H. E., Gibbons, G. H., Morishita, R., Lewis, N. P., Zhang, L., Nakajima, M., Kaneda, Y., Cooke, J. P., and Dzau, V. J. (1995). Gene therapy inhibiting neointimal vascular lesion: In vivo transfer of endothelial cell nitric oxide synthase gene. Proc. Natl. Acad. Sci. USA 92: 1137–1141. 25 Ohno, T., Gordon, D., San, H., Pompili, V. J., Imperiale, M. J., Nabel, G. J., and Nabel, E. G. (1994). Gene therapy for vascular smooth muscle cell proliferation after arterial injury. Science 265: 781–784. 26 Mahato, R. I., Monera, O. D., Smith, L. C., and Rolland, A. (1999). Peptide-based gene delivery. Curr. Opin. Mol. Ther. 1: 226 –244. 27 Leclerc, G., Gal, D., Takeshita, S., Nikol, S., Weir, L., and Isner, J. M. (1992). Percutaneous arterial gene transfer in a rabbit model. Efficiency in normal and balloon-dilated atherosclerotic arteries. J. Clin. Invest. 90: 936 –944. 28 Nabel, E. G., Plautz, G., and Nabel, G. J. (1990). Site-specific gene expression in vivo by direct gene transfer into the arterial wall. Science 249: 1285–1288. 29 Leventis, R., and Silvius, J. R. (1990). Interactions of mammalian cells with lipid dispersions containing novel metabolizable cationic amphiphiles. Biochim. Biophys. Acta 1023: 124 –132. 30 Malik, N., Francis, S. E., Holt, C. M., Gunn, J., Thomas, G. L., Shepherd, L., Chamberlain, J., Newman, C. M., Cumberland, D. C., and Crossman, D. C. (1998). Apoptosis and cell proliferation after porcine coronary angioplasty. Circulation 98: 1657–1665. 31 Winder, D., Guenzburg, W. H., Erfle, V., and Salmons, B. (1998). Expression of antimicrobial peptides has an antitumour effect in human cells. Biochem. Biophys. Res. Commun. 242: 608 – 612.
375
