- Email: [email protected]
Computer-assisted Hydrodynamic Gene Delivery
original article
© The American Society of Gene Therapy
Computer-assisted Hydrodynamic Gene Delivery Takeshi Suda1, Kieko Suda1 and Dexi Liu1 Department of Pharmaceutical Sciences, School of Pharmacy, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
1
The recently developed hydrodynamic delivery method makes it possible to deliver DNA and RNA into parenchyma cells by intravascular injection of nucleic acid– containing solution. While this procedure is effective in rodents, it is difficult to perform in large animals, because manual control while delivering the injection cannot be sufficiently reliable for achieving a just-right hydrodynamic pressure in targeted tissue. In order to overcome this problem, we have developed a computer-controlled injection device that uses real-time intravascular pressure as a regulator. Using the new injection device, and mouse liver as the model organ, we demonstrated continuous injection at a single pressure and different pressures, and also serial (repeated) injections at intervals of 250 ms, by programming the computer according to the need. When assessed by reporter plasmids, the computer-controlled injection device exhibits gene delivery efficiency similar to that of conventional hydrodynamic injection. The device is also effective in gene delivery to kidney and muscle cells in rats, with plasmids or adenoviral vectors as gene carriers. Successful gene delivery to liver and kidney was also demonstrated in pigs, with the computercontrolled injection being combined with image-guided catheterization. These results represent a significant advance in in vivo gene delivery research, with potential for use in gene therapy in humans. Received 2 January 2008; accepted 10 March 2008; published online 8 April 2008. doi:10.1038/mt.2008.66
INTRODUCTION Gene therapy has tremendous potential to provide highly specific, safe, and effective treatments for many diseases, ranging from single-gene defects to complex conditions involving genetic and environmental factors. In practice, gene therapy requires safe and effective methods for the delivery of therapeutic genes into the target cells. One such method is hydrodynamic delivery, which involves rapid intravascular injection of a relatively large volume of DNA solution in order to transiently permeabilize the endothelium and plasma membrane of parenchyma cells and facilitate efficient intracellular gene transfer. Earlier studies have shown that the hydrodynamics-based procedure is highly effective in gene delivery into the liver, muscle, and kidneys of rodents.1–3 In many cases, gene expression in hydrodynamically transfected animals is long-lasting and the gene products reach therapeutic levels.4–12
The high efficiency and simplicity of hydrodynamic gene delivery in rodents aroused the interest of the gene therapy community in this procedure as having potential for application in humans.13,14 Toward this end, significant efforts have been made in developing various strategies to reduce the volume required to be injected during hydrodynamic gene delivery. For example, Zhang et al. demonstrated in rats that the volume required for successful hydrodynamic gene delivery can be reduced from ~10% of body weight (BW) for conventional tail-vein injection to <2% of BW by performing localized hydrodynamic gene delivery through the hepatic vein into an isolated liver.15 Effective gene transfer to rat liver by localized injection of 125 μg of reporter plasmid in a volume of 2.5% BW has also been reported.16 Using an image-guided catheter insertion technique, Eastman et al. showed, in rabbits, that a volume of 15 ml/kg was sufficient for successful gene transfer into liver cells.17 Recently, Yoshino et al., Alino et al., and Fabre et al. reported hydrodynamic gene delivery into pig liver using image-guided catheterization.18–20 They showed that hydrodynamic gene delivery into pig liver is possible, and that tissue damage is minimal. These results confirm that hydrodynamic gene delivery is applicable to large animals comparable in size to humans. However, the problem revealed by these studies was that the level of transgene expression was not controllable in spite of the same surgeon using the same procedure.18 This is primarily because of the difficulty involved in placing a catheter at precisely the same site of a vasculature from time to time. Consequently, the tissue area covered by each hydrodynamic injection varies from one experiment to another, resulting in variations in hydrodynamic pressure in the vasculature, and consequent variations in gene transfer efficiency. In order to render hydrodynamic gene delivery reliable, we report here the design and development of a computer-controlled injection device capable of performing hydrodynamic gene delivery based on the intravascular pressure inside the targeted organ. Using mice, rats, and pigs as animal models, we demonstrated that our new design is effective in gene delivery to the liver, muscle, and kidney, and has great potential for clinical applications.
RESULTS Design of computer-controlled injection device Our design uses intravascular pressure inside the target tissue as a key regulator. The injection device (Figure 1) consists of four parts: a computer, a solution driver and delivery catheter, a pressure detection system, and an injection control system. The computer is used for data acquisition and analysis, and for operative control. The solution driver propels the DNA solution into the selected
Correspondence: Dexi Liu, Department of Pharmaceutical Sciences, University of Pittsburgh School of Pharmacy, 527 Salk Hall, Pittsburgh, Pennsylvania 15261, USA. E-mail: [email protected]
1098
www.moleculartherapy.org vol. 16 no. 6, 1098–1104 june 2008
© The American Society of Gene Therapy
Computer-controlled Gene Delivery
a
b IVC pressure (mm Hg)
Computer
L
A/D converter
Valve driver
d IVC pressure (mm Hg)
Amplifier
Pressure transducer
Three-way adaptor
Valve
Solution container
Solution driver
Figure 1 Schematic presentation of computer-controlled hydrodynamic injection device (A/D, analog-to-digital converter).
blood vessel of the target organ. A pressure transducer connected to an amplifier is inserted into the blood vessel through a threeway adaptor to collect and send an intravascular pressure signal to the computer. Before the injection, the solution container is filled with DNA solution and connected to the solution driver (CO2 gas cylinder). The switch of the CO2 cylinder is set at the desired injection pressure. A catheter and a pressure transducer are inserted into the selected blood vessel. The injection starts when the valve-driver activates the valve (switching on). The DNA solution is pushed into the blood vessel of the animal through the inserted catheter. The pressure transducer senses the pressure near the tip of the inserted catheter and sends the pressure signal to the amplifier. This signal is converted and processed in the computer. When intravascular pressure reaches the preselected pressure value programmed into the computer, the injection valve is automatically inactivated by the valve-driver, and the injection stops. Similarly, if the intravascular pressure falls below the preselected value, the injection resumes. With this design, the real-time control of the injection is achieved by regulating the opening and closing of the valve in the injection route through a negative feedback circuit loop from the pressure transducer through the computer. As determined by a pressure profile programmed into the computer, the device injects the DNA solution into the animal through an inserted catheter when the pressure is below the preset pressure value, and shuts down when that value is reached. By varying the catheter size, injection pressure, and switching valve, this device can perform hydrodynamic injection at speeds ranging from 0.1 to 100 ml/s. We have named the computer-controlled injection device the “hydrojector” to reflect the hydrodynamic nature of injection.
Pressure profiles of hydrojector-assisted injection Programmed injection was demonstrated in anesthetized mice using the experimental set-up (Figure 2a). A catheter with a pressure transducer inserted in it was directly placed into the inferior vena cava (IVC). Two clamps were placed on the IVC, one at the segment below the catheter insertion site and the other at the section immediately below the diaphragm, in order to direct the injected solution into the liver in retrograde. Figure 2b shows examples of pressure profiles obtained at preset peak pressures of 10, 15, and 20 mm Hg. By using appropriate injection pressures, different durations of time could be taken to reach the same intravascular preset peak pressure (Figure 2c). Continuous Molecular Therapy vol. 16 no. 6 june 2008
20 10 0
e
30
f
20 10 0 0
Catheter
c
30
10
20
30 0
10 20 Time (s)
30 0
10
20
30
Figure 2 Pressure profiles of hydrojector-assisted injection. (a) Experimental setup for computer-assisted hydrodynamic gene delivery into liver [arrowheads: clamping sites on the inferior vena cava (IVC)]. (b) Pressure profiles of hydrodynamic injection with the peak pressure set at 10, 15, or 20 mm Hg. (c) Pressure profiles of hydrodynamic injection with fixed peak pressure and different injection durations for reaching peak pressure. (d) Pressure profile of a prolonged injection at the peak pressure. (e) Pressure profile of an injection at peak pressure followed by a sustained injection at lower pressure. (f) Pressure profile of repeated injections with the same peak pressure and time intervals of 250 ms.
injection at the same (Figure 2d) or different (Figure 2e) peak pressure, or repeated injections with time intervals of 250 ms (Figure 2f) or longer can also be achieved. In practice, our computer-controlled injection system is fully programmable for the various requirements of hydrodynamic delivery.
Comparison between computer-assisted and conventional hydrodynamic gene delivery Various parameters were considered in carrying out a comparison between conventional hydrodynamic injection through the tailvein and localized injection using the hydrojector. At a peak pressure of 30 mm Hg in mice, localized gene delivery by hydrojector required less than half the volume of DNA solution required in conventional hydrodynamic gene delivery (Figure 3a), for producing the same level of transgene expression (Figure 3b) and the same percentage of transfected liver cells (Figure 3c and d). Other than a small “shoulder” in the pressure profile because of the clamping on the IVC, the hydrojector generated a pressure profile in IVC identical to that of conventional hydrodynamic injection (Figure 3e), suggesting a better efficacy achieved by the hydrojector in establishing the hydrodynamic pressure. Whereas conventional hydrodynamic tail-vein injection requires large volumes of the injection to drive a retrograde flow of the injected solution into the liver, the hydrojector achieves a significant reduction in the volume of injection required, because it eliminates the volume needed to fill the entire IVC and the heart chambers before reaching the liver. Hydrojector-assisted gene delivery into rat liver (tenfold larger than that of a mouse) resulted in luciferase gene expression at the 2.9 × 108 relative light unit/mg level (single peak pressure: 50 mm Hg; injection time: 10 seconds; injection volume: 5.4 ± 0.2% of BW), thereby proving the effectiveness of the device in larger animals. Hydrojector-assisted gene delivery to muscle and kidney Hydrojector-assisted gene delivery into muscle and kidney was also performed. With the pressure profile shown in Figure 4a, 1099
© The American Society of Gene Therapy
c
15
Hydrodynamic
10
0
Local
Systemic
10
109
Kidney
107
Local
Systemic
Local
30 20
Immunohistochemistry
10 0 10 20 Time (s)
0 10
20
30
d
120
60
0 10 20 Time (s)
Luciferase activity (RLU/mg)
200
30
f
g
h
1010 108 6
10
2.5
4
10
102
8.0
10
10
108 106 4
10
8.0
10
Figure 5 Comparison of the levels of green fluorescent protein (GFP) expression in muscle and kidney. Fluorescence microscope images were taken (a,c) 4 days after computer-assisted hydrodynamic delivery, or (b,d) slow infusion of Ad-GFP vectors in (a,b) muscle, and (c,d) kidney. Immunohistochemical examination of the (e) nontargeted kidney sample, and targeted kidney sample (f) without or (g,h) with anti-GFP antibody staining. Clear positive signals are seen, both in the (g) medulla and (h) cortex of the surrounding glomerulus. Original magnifications are ×4 (a–d), ×10 (e,f), and ×40 (g,h).
.26
2
Systemic Local
Figure 4 Hydrojector-assisted hydrodynamic gene delivery into muscle and kidney in rats. (a) Intrafemoral vein pressure during a continued injection (10 seconds) with programmed peak pressure at 350 mm Hg. (b) Comparison between conventional (systemic) and computer-assisted (local) hydrodynamic gene delivery in respect of levels of luciferase gene expression in muscle. (c) Intrarenal vein pressure of computer-assisted hydrodynamic injection into the right kidney with intravascular pressure set at 100 mm Hg. (d) Comparison between conventional (systemic) and computer-assisted (local) hydrodynamic gene delivery in respect of the levels of luciferase gene expression in kidney. The values shown on the bars in b and d represent the injection volumes presented as a percentage of total body weight. The data are presented as mean values ± SD, n = 3. RLU, relative light unit.
the level of reporter gene expression in targeted muscle cells after hydrojector injection was 10,000-fold higher than that from hydrodynamic tail-vein injection (Figure 4b). More important, the volume of DNA solution required decreased from 8 to 2.5% of BW (values on the bars in Figure 4b). Similarly, hydrojectorassisted gene delivery through the renal vein at a peak pressure of 100 mm Hg (Figure 4c) resulted in 100 times more transgene product in rat kidney than could be achieved by conventional 1100
e 30
Luciferase activity (RLU/mg)
Femoral vein pressure (mm Hg) Renal vein pressure (mm Hg)
b
400
0
d
8
10
Figure 3 Comparison of conventional (systemic) and computerassisted (local) hydrodynamic gene delivery. (a) Comparison in respect of injection volume. (b) Comparison of levels of luciferase gene expression. (c,d) Comparison of densities of transfected cells in the liver. (e) Comparison of pressure profiles in the inferior vena cava (IVC) (blue line, computer-assisted injection; grey line, conventional hydrodynamic injection through the tail vein). RLU, relative light unit.
c
c
d
0
0
b
Systemic
10
e
a
Slow
a Muscle
5
IVC pressure (mm Hg)
b
Luciferase activity (RLU/mg)
a
Injection volume (% of BW)
Computer-controlled Gene Delivery
hydrodynamic tail-vein injection (Figure 4d). Significantly, the injection volume for muscle and kidney was only 2.5 and 0.26% of BW (values on the bars in Figure 4b and d), respectively, substantially lower than the injection volumes reported earlier.21–24
Hydrojector-assisted delivery of viral vectors As in the case of conventional nonviral approaches, the efficiency of gene delivery by viral vectors into kidney and muscle is relatively low primarily because the endothelial barrier hinders the direct access of viral vectors into the parenchyma cells. Taking advantage of the permeabilization effect that hydrodynamic injection has on the endothelium,25,26 we examined the usefulness of the hydrojector in enhancing viral vector–mediated gene delivery, using rat muscle and kidney as the targets. For gene delivery into muscle, 5 × 1011 adenoviral particles containing the green fluorescent protein (GFP) gene were injected through the femoral vein. GFP expression was seen in the targeted hindleg 4 days after injection (Figure 5a), whereas none was detected in the hindleg of the same animal receiving slow infusion of the same number of viral particles (Figure 5b). Again using the hydrojector, the same number of viral particles was injected into rat kidney through the right renal vein. www.moleculartherapy.org vol. 16 no. 6 june 2008
© The American Society of Gene Therapy
Computer-controlled Gene Delivery
Figure 5c shows clear GFP expression in the targeted right kidney but not in the left kidney that received a slow infusion of the virus (Figure 5d). Immunohistochemical staining of kidney sections with anti-GFP antibodies revealed enriched GFP expression in both the cortex and the medulla of the hydrodynamically treated kidney (Figure 5g and h) (Figure 5e and f, negative control).
Hydrojector-assisted hydrodynamic gene delivery in pigs In order to evaluate the feasibility of using our computer-controlled injection device for gene delivery into liver and kidney in large (comparable in size to humans) animals we used swine as an animal model. A standard image-guided insertion procedure employed in the clinic was used for placing a balloon catheter into the right lateral hepatic vein or renal vein to target the liver or
b
d
150 100 50 0 0
e
10 20 30 Time (s)
40
7.4 × 106 3.7 × 106 6.1 × 106
300 200 100 0 0
10 20 30 Time (s)
Tissue damage assessment Tissue damage caused by computer-assisted hydrodynamic gene delivery into the liver, kidney, and muscle was evaluated in rats, using serum biochemistry (Figure 7) and histological examination. As found after hydrodynamic tail-vein injection, tissue-specific
40
Caudate lobe 2.9 × 103
Right lateral lobe 4.8 × 106 3.0 × 106 103
Right middle lobe
f
1.8 × 10 1.2 × 105 1.4 ×
3.1 × 103
Left middle lobe
5
105
ALT (U/l)
6.7 ×
a
Left lateral lobe
7.3 × 101
g
4.7 × 101
5.7 × 105
Figure 6 Image-guided hydrodynamic gene delivery to pig liver and kidney. (a,b) Fluoroscopic images of the pig liver and kidney showing the relative locations of the inserted balloon catheters [the injection balloon catheter (*), the occlusion balloon in portal vein (**), and the occlusion balloon in the IVC (***)]. The distribution of phase contrast medium injected through the injection catheter to verify obstruction of blood vessels by the balloon catheters is also indicated (****). (c) Lateral hepatic vein pressure profile during and soon after hydrodynamic injection with peak pressure set at 100 mm Hg. (d) Right renal vein pressure profile during and soon after hydrodynamic injection with peak pressure set at 240 mm Hg. (e) Photograph of hydrodynamically transfected pig liver in cross-section, showing the levels of luciferase gene expression at the approximate sites from which liver samples were collected. (f) Photograph of hydrodynamically transfected right kidney cut in half in the middle, showing the levels of luciferase gene expression at the sites from which tissue samples were collected. (g) Photograph of untreated left kidney cut in half in the middle, showing the background level of luciferase activity. The levels of luciferase gene expression were determined 6 hours after gene delivery. The data presented are from two pigs, one each for gene delivery to the liver and the kidney. The values in e–g represent luciferase activity in relative light unit/mg of extracted proteins.
Molecular Therapy vol. 16 no. 6 june 2008
104 103 102 101 100
b
2
Crt (mg/dl)
Hepatic vein pressure (mm Hg)
c
* ***
**
Right renal vein pressure (mm Hg)
*
*** *
** *
1
c
300
CPK (U/l)
a
kidney, respectively. Figure 6a and b demonstrates the relative positions of the inserted balloon catheters. In order to minimize leakage of the injected DNA solution into the portal vein and/or IVC, occlusion balloons were inserted from the superior mesenteric vein to the portal vein for the liver, and/or from the right femoral vein to the IVC for liver (located between diaphragm and hepatic vein) and kidney (covering the renal–IVC junction). Complete occlusion was confirmed by injection of a small amount of phase-contrast medium (gray area in Figure 6a beyond the balloon of the injection catheter). Figure 6c and d shows the pressure profiles of injection into the right hepatic vein and right renal vein, respectively. The total injection volume was 800 ml delivered in 20 seconds (40 ml/s) for the liver, and 55 ml in 11.7 seconds (4.7 ml/s) for the kidney. Figure 6e shows reporter gene expression in the liver in both targeted (right lateral and caudate) and nontargeted (right medial, left medial, and left lateral) lobes. Figure 6f and g reveals luciferase activity in the treated and untreated kidneys, respectively. Luciferase gene expression was seen in both targeted and nontargeted liver lobes because of the inner-lobular vascular connection. The level of luciferase gene expression seen in the targeted right lateral lobe, however, was 1,000-fold higher than that in the nontargeted lobe. A significant level of luciferase gene expression was seen in the hydrodynamically treated kidney (Figure 6f), but not in the untreated one (Figure 6g).
200
0
100 0
0
100 Time (h)
200
Figure 7 Assessment of tissue damage after hydrodynamic gene delivery using the computer-controlled injection device. Serum concentrations of tissue-specific marker enzymes in rats [(a) alanine aminotransferase (ALT) for liver, (b) creatinine (Crt) for kidney, and (c) creatinine phosphokinase (CPK) for muscle] at 6, 24, 48, 72, 120, and 168 hours after (closed circles) computer-assisted hydrodynamic gene delivery, or after (open circles) a slow (60 seconds) infusion of the same volume of saline. Open bars in each plot represent the serum concentration of marker enzyme in untreated animals. The data are presented as mean values ± SD, n = 3.
1101
Computer-controlled Gene Delivery
enzyme concentration was initially high after hydrojector-assisted injection, and decreased rapidly with time. At 24 hours, the serum concentration of alanine aminotransferase, a liver-specific marker, was three- to fourfold higher in hydrodynamically treated animals than in untreated animals (225 ± 131 U/l versus 62 ± 16 U/l) (Figure 7a). Similarly, the concentrations of the kidney- and muscle-specific markers, creatinine and creatine phosphokinase, 24 hours after hydrodynamic treatments, were 0.3 ± 0.04 mg/dl and 82 ± 27 U/l, respectively (Figure 7b and c), which was similar to the levels in control animals that had received no treatment (open bar insert). All the elevated values of enzyme concentrations in treated animals returned to the normal ranges within 3 days, which is in full agreement with results obtained with conventional hydrodynamic delivery.1,27,28 Histological evaluation of the rat livers revealed no obvious tissue damage (data not shown). Examination of the targeted kidney revealed no obvious bleeding, tubular cell detachment, apoptosis, necrosis, or inflammatory cell infiltration in either the cortical or medullary regions. The interlobular vein also showed no obvious abnormality. Muscles in the targeted hindleg revealed no obvious tissue damage such as edema, torn bundles, or inflammatory cell infiltration. Muscle fibers with multiple nuclei and regular cross-striation patterns were clearly observed, indicating mature muscles without obvious sarcomere unit destruction. All these results confirm previous findings in mice, that the structural impact of hydrodynamic gene delivery is transient, and the procedure is safe.1,26
DISCUSSION We have demonstrated in this study the successful development of a computer-controlled injection device for hydrodynamic gene delivery. The system uses negative feedback from the vascular pressure at the injection site to precisely control the injection, based on a preselected pressure profile for maximal gene transfer efficiency and minimal tissue damage. The most significant feature of our design is to use intravascular pressure as the key regulator for the injection. The injection shuts down automatically when the desired intravascular pressure is reached. Similarly, when the pressure falls below a preselected level, the injection automatically resumes. Intravascular pressure at the injection site can thereby be maintained within any desired range over the duration of the injection. Because the injection is controlled by the intravascular pressure, another important feature of this injection device is that the volume needed in order to develop a sufficiently elevated pressure, and thereby achieve a successful gene transfer, is self-adjusting based on the size and anatomical structure of the selected organ of an individual. This system can be used for both viral and nonviral gene delivery, and provides a potential method for safe, reliable, and efficient in vivo gene delivery. One important finding made during this study was that the pressure profiles that are effective for gene delivery vary from one target organ to another, and also from one animal species to another. Larger animals tend to require higher pressure for successful gene transfer to a given target organ. Although such interspecies differences could be caused by differences in the structure of the endothelium and parenchyma cell membrane of the target organs, it is likely that higher pressure required for larger organs is because of the low elasticity of the vasculature and parenchyma 1102
© The American Society of Gene Therapy
cells. The significantly high pressure and relatively low luciferase activity associated with kidney and muscle appears to bear out such a notion. Compared to the liver, both kidney and muscle have much lower elasticity. More research is needed to determine whether the turbulence generated by the injection inside the vasculature influences the efficiency of gene delivery. As far as the applications of the hydrojector system are concerned, future studies must establish optimal pressure profiles for each targeted organ so as to enable the development of computer programs that will guide safe and efficient gene delivery. With continuing efforts, and with new data being generated, it is foreseeable that an effective and clinically applicable hydrodynamic procedure will be established.
MATERIALS AND METHODS All experiments performed on mice, rats, and pigs were approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh (Pittsburgh, Pennsylvania). Conventional hydrodynamic gene delivery was performed in accordance with the procedure published earlier.2 Pressure changes in the IVC during conventional hydrodynamic injection through the tail vein were measured according to the method described by Zhang et al.27 Materials. Plasmids of pCMV-Luc and pCMV-LacZ were purified using
CsCl-ethidium bromide density gradient ultracentrifugation and stored in Tris–EDTA buffer. The purity of the plasmids was verified by absorbency at 260 and 280 nm and 1% agarose gel electrophoresis. Ad-GFP vectors were generously provided by Xiao Xiao (University of North Carolina at Chapel Hill). A single batch of high titer adenovirus stock (1012 viral particles/ml) was used throughout the experiments. The luciferase assay kit was from Promega (Madison, WI). CD-1 mice (18–20 g, female) and Wister rats (180–200 g, female) were from Charles River (Wilmington, MA), and pigs (20–30 kg) were from Wally Whippo (Enon Valley, PA). The 12 F sheath for image-guided catheter insertion was from COOK (Bloomington, IN), and guidewire was from Terumo Medical (Somerset, NJ). The 9 F balloon catheter for injection in pigs was custom-made by Clinical Supply (Gifu, Japan). The occlusion balloon catheters were purchased from Medtronic (Minneapolis, MN). Mikro Tip catheter transducer was from Millar Instruments (Houston, TX). Construction of the computer-controlled injection device. The injection
device was built according to the illustration shown in Figure 1. The computer program running the entire system was developed in-house to monitor pressure every 100 ms and operate the opening and closing of the valve at intervals of 250 ms. A regular CO2 tank, commonly used for cell culture, was used as the solution driver. The miniature pressure transducer (SPR671 or SPR-407) was from Millar Instruments. A solenoid or pneumatic valve was placed between the solution container and the catheter, and the opening and closing of the valve were controlled through activation and inactivation of a power supply (RK-80; Matsusada Precision, Shiga, Japan), regulated by transistor-transistor-logic signals generated from a signal generator (ADA16-32/2(CB) F; Contec, Osaka, Japan). Computer-assisted gene delivery to the liver. A midline incision was made
in the abdomen of anesthetized mice and rats to expose the internal organs. A 22G peripheral catheter (Terumo Medical, Somerset, NJ) was inserted into the IVC near the bifurcation of the iliac veins, and advanced further, above the confluence with the left renal vein, at which point a vessel clamp (AROSurgical, Newport Beach, CA) was placed to prevent backflow of the injected solution. A pressure transducer was inserted by means of a threeway adaptor into the IVC through the inserted catheter. Saline with or without plasmid DNA (pCMV-Luc, 10 μg/ml) was injected, and the pressure profile and injection volume were recorded. During injection, the IVC www.moleculartherapy.org vol. 16 no. 6 june 2008
© The American Society of Gene Therapy
was manually compressed using a cotton swab at the section immediately below the diaphragm. The total occlusion time was 40 seconds including the injection time. At the end of the procedure, hemostasis was established by applying gelatin sponge and n-butyl cyanoacrylate adhesive (3M Health Care, St Paul, MN) at the insertion site, and the midline incision was sutured layer-by-layer using 4-0 silk. Computer-assisted gene delivery to kidney. Similar to the procedure
undertaken for the liver, a midline incision was made in the abdomen of anesthetized rats. A peripheral catheter (22G) was inserted through the sidewall of the IVC into the right renal vein. After insertion of the pressure transducer through the inserted catheter, the right renal vein was taped using 4-0 silk and temporarily tied off. The peak pressure was set at 100 mm Hg, and the injection was administered, consisting of saline with plasmid DNA (100 μg/ml) or Ad-GFP viral vectors (5 × 1011 viral particle in volume of 0.25% BW). The same procedure as that used for liver was employed for hemostasis and incision closure. Computer-assisted gene delivery to muscle. On an anesthetized rat, a
midline skin incision was made in the lower abdomen. A 22G peripheral catheter was inserted into the exposed femoral vein in a hindleg near the bifurcation from the external iliac vein, and advanced into the medial saphenous vein. After inserting a pressure transducer through the inserted catheter, both the femoral vein and artery were taped and temporarily tied off using a 4-0 silk. Hydrodynamic injection of saline containing pCMVLuc (100 μg/ml) or Ad-GFP viral particles (5 × 1011 viral particles in volume of 3% BW) in phosphate-buffered saline was administered. At the end of the procedure, hemostasis was achieved by simple compression with sterile gauze, and the incision was closed. Image-guided hydrodynamic gene delivery to liver and kidney in swine.
For gene delivery to the liver, the animal was put under general anesthesia, and laid out on a fluoroscopy table. Sterile drapes were placed appropriately, and a midline skin incision was made near the mandibular arch to expose the jugular vein. An 18G peripheral catheter was inserted over a stylet into the exposed jugular vein. The stylet was removed and a 0.035-in hydrophilic guide wire was inserted and advanced through the catheter into the IVC under fluoroscopic guidance. Retaining the guide wire in place, the peripheral catheter was replaced with a 12 F-size short sheath. After removing the guide wire and washing out blood from the sheath with saline/heparin, a 9 F-size balloon catheter was inserted through the sheath into the right lateral lobe of the liver, followed by insertion of the pressure transducer. Two occlusion balloons (8 F, 46-mm diameter when fully inflated) were inserted, one to occlude passage from the femoral vein to the IVC and the other to occlude passage from the superior mesenteric vein into the main portal trunk, in order to block leakage of injected solution into the IVC and the portal vein, respectively. The occlusion balloons were inflated by infusion of phase-contrast medium. A small volume of contrast medium was injected to verify a complete occlusion. Saline containing plasmid DNA (100 μg/ml) was injected through the hydrojector with peak pressure set at 100 mm Hg. For gene delivery to the kidney, the catheter insertion was made into the right renal vein following the same procedure, with the exception that only the IVC occlusion balloon was utilized, and the peak pressure was set at 240 mm Hg. Intravascular pressure was continuously monitored and recorded during and soon after the injection. The volume actually injected was calculated on the basis of the amount of solution in the solution container before and after injection. Reporter gene expression assay. The level of reporter gene expression was
determined using procedures established earlier.1,27 In the case of mouse liver, mouse kidney, or rat kidney for the purpose of luciferase assay, the whole organ was collected whereas, for rat liver, pig liver, and pig kidney, tissue cubes (~200 mg) were obtained from the respective organs. In animals that had received hydrodynamic treatment in the legs, tissue samples were obtained from the quadriceps femoris, biceps femoris, tibialis anterior, and Molecular Therapy vol. 16 no. 6 june 2008
Computer-controlled Gene Delivery
gastrocnemius muscles. The tissues were immediately frozen in powdered dry ice and stored at −80 °C until use. Histochemical β-galactosidase assay on mouse liver was performed in accordance with the method described earlier.1 Immediately after tissue sectioning, the expression of GFP was evaluated under a fluorescence microscope.25 Immunohistochemistry for GFP expression in kidney was performed on 4-μm thick formalin-fixed paraffin-embedded kidney sections using biotin–avidin enzyme conjugates and rabbit anti-GFP antibody of ab290 (Abcam, Cambridge, MA) at a 1,000-fold dilution overnight at 4 °C. The distribution of the primary antibody was visualized through avidin–biotin complex amplification using a Vectastain Elite ABC Kit (Vector, Burlingame, CA) with 3, 3ʹ-diaminobenzidine as chromaogen (Vector). Assessment of tissue damage. Saline was injected using the hydrojector to
target the muscle, kidney, and liver in separate rats. At 6, 24, 48, 72, 120, and 168 hours after the injection, blood samples were collected from the lateral tail vein for determination of creatinine and creatine phosphokinase concentrations, and from the retro-orbital plexus for quantification of alanine aminotransferase. Serum concentrations of the marker enzymes were determined using commercial kits (Stanbio Laboratory, Boerne, TX) in accordance with the manufacturer’s instructions. Histological evaluation was performed using formalin-fixed paraffin-embedded sections of the target organ obtained 24 hours after the injection. Hematoxylin and eosin staining was performed for the liver and muscles, whereas periodic acid-Schiff stain was employed for the kidney.
ACKNOWLEDGMENTS We thank Joseph E. Knapp for his critical reading of the manuscript. This work was supported in part by grants from National Institutes of Health (EB2002946 and HL075542).
REFERENCES
1. Liu, F, Song, Y and Liu, D (1999). Hydrodynamics-based transfection in animals by systemic administration of plasmid DNA. Gene Ther 6: 1258–1266. 2. Zhang, G, Budker, V and Wolff, JA (1999). High levels of foreign gene expression in hepatocytes after tail vein injections of naked plasmid DNA. Hum Gene Ther 10: 1735–1737. 3. Maruyama, H, Higuchi, N, Nishikawa, Y, Hirahara, H, Iino, N, Kameda, S et al. (2002). Kidney-targeted naked DNA transfer by retrograde renal vein injection in rats. Hum Gene Ther 13: 455–468. 4. Zhang, G, Song, YK and Liu, D (2000). Long-term expression of human α1-antitrypsin gene in mouse liver achieved by intravenous administration of plasmid DNA using a hydrodynamics-based procedure. Gene Ther 7: 1344–1349. 5. Miao, CH, Thompson, AR, Loeb, K and Ye, X (2001). Long-term and therapeutic-level hepatic gene expression of human factor IX after naked plasmid transfer in vivo. Mol Ther 3: 947–957. 6. Stoll, SM, Sclimenti, CR, Baba, EJ, Meuse, L, Kay, MA and Calos, MP (2001). EpsteinBarr virus/human vector provides high-level, long-term expression of α1-antitrypsin in mice. Mol Ther 4: 122–129. 7. Yew, NS, Przybylska, M, Ziegler, RJ, Liu, D and Cheng, SH (2001). High and sustained transgene expression in vivo from plasmid vectors containing a hybrid ubiquitin promoter. Mol Ther 4: 75–82. 8. Alino, SF, Crespo, A and Dasi, F (2003). Long-term therapeutic levels of human α-1 antitrypsin in plasma after hydrodynamic injection of nonviral DNA. Gene Ther 10: 1672–1679. 9. Chen, ZY, He, CY, Ehrhardt, A and Kay, MA (2003). Minicircle DNA vectors devoid of bacterial DNA result in persistent and high-level transgene expression in vivo. Mol Ther 8: 495–500. 10. Score, PR, Belur, LR, Frandsen, JL, Guerts, JL, Yamaguchi, T, Somia, NV et al. (2006). Sleeping Beauty-mediated transposition and long-term expression in vivo: use of the LoxP/Cre recombinase system to distinguish transposition-specific expression. Mol Ther 13: 617–624. 11. Dai, C, Yang, J and Liu, Y (2002). Single injection of naked plasmid encoding hepatocyte growth factor prevents cell death and ameliorates acute renal failure in mice. J Am Soc Nephrol 13: 411–422. 12. Chen, L and Woo, SL (2005). Complete and persistent phenotypic correction of phenylketonuria in mice by site-specific genome integration of murine phenylalanine hydroxylase cDNA. Proc Natl Acad Sci USA 102: 15581–15586. 13. Suda, T and Liu, D (2007). Hydrodynamic gene delivery: its principles and applications. Mol Ther 15: 2063–2069. 14. Herweijer, H and Wolff, JA (2007). Gene therapy progress and prospects: hydrodynamic gene delivery. Gene Ther 14: 99–107. 15. Zhang, X, Dong, X, Sawyer, GJ, Collins, L and Fabre, JW (2004). Regional hydrodynamic gene delivery to the rat liver with physiological volumes of DNA solution. J Gene Med 6: 693–703. 16. Inoue, S, Yakamata, Y, Kaneko, M and Kobayashi, E (2004). Gene therapy for organ grafts using rapid injection of naked DNA: application of the rat liver. Transplantation 77: 997–1003.
1103
Computer-controlled Gene Delivery
17. Eastman, SJ, Baskin, KM, Hodges, BL, Chu, Q, Gates, A, Dreusicke, R et al. (2002). Development of catheter-based procedures for transducing the isolated rabbit liver with plasmid DNA. Hum Gene Ther 13: 2065–2077. 18. Yoshino, H, Hashizume, K and Kobayashi, E (2006). Naked plasmid DNA transfer to the porcine liver using rapid injection with large volume. Gene Ther 13: 1696–1702. 19. Alino, SF, Herrero, MJ, Noguera, I, Dasi, F and Sanchez, M (2007). Pig liver gene therapy by noninvasive interventionist catheterism. Gene Ther 14: 334–343. 20. Fabre, JW, Grehan, A, Whitehorne, M, Sawyer, GJ, Dong, X, Salehi, S et al. (2007). Hydrodynamic gene delivery to the pig liver via an isolated segment of the inferior vena cava. Gene Ther 15: 452–462. 21. Budker, V, Zhang, G, Danko, I, Williams, P and Wolff, J (1998). The efficient expression of intravascularly delivered DNA in rat muscle. Gene Ther 5: 272–276. 22. Su, LT, Gopal, K, Wang, Z, Yin, X, Nelson, A, Kozyak, BW et al. (2005). Uniform scale-independent gene transfer to striated muscle after transvenular extravasation of vector. Circulation 112: 1780–1788.
1104
© The American Society of Gene Therapy
23. Hagstrom, JE, Hegge, J, Zhang, G, Noble, M, Budker, V, Lewis, DL et al. (2004). A facile nonviral method for delivering genes and siRNAs to skeletal muscle of mammalian limbs. Mol Ther 10: 386–398. 24. Danialou, G, Comtois, AS, Matecki, S, Nalbantoglu, J, Karpati, G, Gilbert, R et al. (2005). Optimization of regional intraarterial naked DNA-mediated transgene delivery to skeletal muscles in a large animal model. Mol Ther 11: 257–266. 25. Brunetti-Pierri, N, Palmer, DJ, Mane, V, Finegold, M, Beaudet, AL and Ng, P (2005). Increased hepatic transduction with reduced systemic dissemination and proinflammatory cytokines following hydrodynamic injection of helper-dependent adenoviral vectors. Mol Ther 12: 99–106. 26. Suda, T, Gao, X, Stolz, DB and Liu, D (2007). Structural impact of hydrodynamic injection on mouse liver. Gene Ther 14: 129–137. 27. Zhang, G, Gao, X, Song, YK, Vollmer, R, Stolz, DB, Gasiorowski, JZ et al. (2004). Hydroporation as the mechanism of hydrodynamic delivery. Gene Ther 11: 675–682. 28. Kobayashi, N, Nishikawa, M, Hirata, K and Takakura, Y (2004). Hydrodynamics-based procedure involves transient hyperpermeability in the hepatic cellular membrane: implication of a nonspecific process in efficient intracellular gene delivery. J Gene Med 6: 584–592.
www.moleculartherapy.org vol. 16 no. 6 june 2008
© The American Society of Gene Therapy
Computer-assisted Hydrodynamic Gene Delivery Takeshi Suda1, Kieko Suda1 and Dexi Liu1 Department of Pharmaceutical Sciences, School of Pharmacy, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
1
The recently developed hydrodynamic delivery method makes it possible to deliver DNA and RNA into parenchyma cells by intravascular injection of nucleic acid– containing solution. While this procedure is effective in rodents, it is difficult to perform in large animals, because manual control while delivering the injection cannot be sufficiently reliable for achieving a just-right hydrodynamic pressure in targeted tissue. In order to overcome this problem, we have developed a computer-controlled injection device that uses real-time intravascular pressure as a regulator. Using the new injection device, and mouse liver as the model organ, we demonstrated continuous injection at a single pressure and different pressures, and also serial (repeated) injections at intervals of 250 ms, by programming the computer according to the need. When assessed by reporter plasmids, the computer-controlled injection device exhibits gene delivery efficiency similar to that of conventional hydrodynamic injection. The device is also effective in gene delivery to kidney and muscle cells in rats, with plasmids or adenoviral vectors as gene carriers. Successful gene delivery to liver and kidney was also demonstrated in pigs, with the computercontrolled injection being combined with image-guided catheterization. These results represent a significant advance in in vivo gene delivery research, with potential for use in gene therapy in humans. Received 2 January 2008; accepted 10 March 2008; published online 8 April 2008. doi:10.1038/mt.2008.66
INTRODUCTION Gene therapy has tremendous potential to provide highly specific, safe, and effective treatments for many diseases, ranging from single-gene defects to complex conditions involving genetic and environmental factors. In practice, gene therapy requires safe and effective methods for the delivery of therapeutic genes into the target cells. One such method is hydrodynamic delivery, which involves rapid intravascular injection of a relatively large volume of DNA solution in order to transiently permeabilize the endothelium and plasma membrane of parenchyma cells and facilitate efficient intracellular gene transfer. Earlier studies have shown that the hydrodynamics-based procedure is highly effective in gene delivery into the liver, muscle, and kidneys of rodents.1–3 In many cases, gene expression in hydrodynamically transfected animals is long-lasting and the gene products reach therapeutic levels.4–12
The high efficiency and simplicity of hydrodynamic gene delivery in rodents aroused the interest of the gene therapy community in this procedure as having potential for application in humans.13,14 Toward this end, significant efforts have been made in developing various strategies to reduce the volume required to be injected during hydrodynamic gene delivery. For example, Zhang et al. demonstrated in rats that the volume required for successful hydrodynamic gene delivery can be reduced from ~10% of body weight (BW) for conventional tail-vein injection to <2% of BW by performing localized hydrodynamic gene delivery through the hepatic vein into an isolated liver.15 Effective gene transfer to rat liver by localized injection of 125 μg of reporter plasmid in a volume of 2.5% BW has also been reported.16 Using an image-guided catheter insertion technique, Eastman et al. showed, in rabbits, that a volume of 15 ml/kg was sufficient for successful gene transfer into liver cells.17 Recently, Yoshino et al., Alino et al., and Fabre et al. reported hydrodynamic gene delivery into pig liver using image-guided catheterization.18–20 They showed that hydrodynamic gene delivery into pig liver is possible, and that tissue damage is minimal. These results confirm that hydrodynamic gene delivery is applicable to large animals comparable in size to humans. However, the problem revealed by these studies was that the level of transgene expression was not controllable in spite of the same surgeon using the same procedure.18 This is primarily because of the difficulty involved in placing a catheter at precisely the same site of a vasculature from time to time. Consequently, the tissue area covered by each hydrodynamic injection varies from one experiment to another, resulting in variations in hydrodynamic pressure in the vasculature, and consequent variations in gene transfer efficiency. In order to render hydrodynamic gene delivery reliable, we report here the design and development of a computer-controlled injection device capable of performing hydrodynamic gene delivery based on the intravascular pressure inside the targeted organ. Using mice, rats, and pigs as animal models, we demonstrated that our new design is effective in gene delivery to the liver, muscle, and kidney, and has great potential for clinical applications.
RESULTS Design of computer-controlled injection device Our design uses intravascular pressure inside the target tissue as a key regulator. The injection device (Figure 1) consists of four parts: a computer, a solution driver and delivery catheter, a pressure detection system, and an injection control system. The computer is used for data acquisition and analysis, and for operative control. The solution driver propels the DNA solution into the selected
Correspondence: Dexi Liu, Department of Pharmaceutical Sciences, University of Pittsburgh School of Pharmacy, 527 Salk Hall, Pittsburgh, Pennsylvania 15261, USA. E-mail: [email protected]
1098
www.moleculartherapy.org vol. 16 no. 6, 1098–1104 june 2008
© The American Society of Gene Therapy
Computer-controlled Gene Delivery
a
b IVC pressure (mm Hg)
Computer
L
A/D converter
Valve driver
d IVC pressure (mm Hg)
Amplifier
Pressure transducer
Three-way adaptor
Valve
Solution container
Solution driver
Figure 1 Schematic presentation of computer-controlled hydrodynamic injection device (A/D, analog-to-digital converter).
blood vessel of the target organ. A pressure transducer connected to an amplifier is inserted into the blood vessel through a threeway adaptor to collect and send an intravascular pressure signal to the computer. Before the injection, the solution container is filled with DNA solution and connected to the solution driver (CO2 gas cylinder). The switch of the CO2 cylinder is set at the desired injection pressure. A catheter and a pressure transducer are inserted into the selected blood vessel. The injection starts when the valve-driver activates the valve (switching on). The DNA solution is pushed into the blood vessel of the animal through the inserted catheter. The pressure transducer senses the pressure near the tip of the inserted catheter and sends the pressure signal to the amplifier. This signal is converted and processed in the computer. When intravascular pressure reaches the preselected pressure value programmed into the computer, the injection valve is automatically inactivated by the valve-driver, and the injection stops. Similarly, if the intravascular pressure falls below the preselected value, the injection resumes. With this design, the real-time control of the injection is achieved by regulating the opening and closing of the valve in the injection route through a negative feedback circuit loop from the pressure transducer through the computer. As determined by a pressure profile programmed into the computer, the device injects the DNA solution into the animal through an inserted catheter when the pressure is below the preset pressure value, and shuts down when that value is reached. By varying the catheter size, injection pressure, and switching valve, this device can perform hydrodynamic injection at speeds ranging from 0.1 to 100 ml/s. We have named the computer-controlled injection device the “hydrojector” to reflect the hydrodynamic nature of injection.
Pressure profiles of hydrojector-assisted injection Programmed injection was demonstrated in anesthetized mice using the experimental set-up (Figure 2a). A catheter with a pressure transducer inserted in it was directly placed into the inferior vena cava (IVC). Two clamps were placed on the IVC, one at the segment below the catheter insertion site and the other at the section immediately below the diaphragm, in order to direct the injected solution into the liver in retrograde. Figure 2b shows examples of pressure profiles obtained at preset peak pressures of 10, 15, and 20 mm Hg. By using appropriate injection pressures, different durations of time could be taken to reach the same intravascular preset peak pressure (Figure 2c). Continuous Molecular Therapy vol. 16 no. 6 june 2008
20 10 0
e
30
f
20 10 0 0
Catheter
c
30
10
20
30 0
10 20 Time (s)
30 0
10
20
30
Figure 2 Pressure profiles of hydrojector-assisted injection. (a) Experimental setup for computer-assisted hydrodynamic gene delivery into liver [arrowheads: clamping sites on the inferior vena cava (IVC)]. (b) Pressure profiles of hydrodynamic injection with the peak pressure set at 10, 15, or 20 mm Hg. (c) Pressure profiles of hydrodynamic injection with fixed peak pressure and different injection durations for reaching peak pressure. (d) Pressure profile of a prolonged injection at the peak pressure. (e) Pressure profile of an injection at peak pressure followed by a sustained injection at lower pressure. (f) Pressure profile of repeated injections with the same peak pressure and time intervals of 250 ms.
injection at the same (Figure 2d) or different (Figure 2e) peak pressure, or repeated injections with time intervals of 250 ms (Figure 2f) or longer can also be achieved. In practice, our computer-controlled injection system is fully programmable for the various requirements of hydrodynamic delivery.
Comparison between computer-assisted and conventional hydrodynamic gene delivery Various parameters were considered in carrying out a comparison between conventional hydrodynamic injection through the tailvein and localized injection using the hydrojector. At a peak pressure of 30 mm Hg in mice, localized gene delivery by hydrojector required less than half the volume of DNA solution required in conventional hydrodynamic gene delivery (Figure 3a), for producing the same level of transgene expression (Figure 3b) and the same percentage of transfected liver cells (Figure 3c and d). Other than a small “shoulder” in the pressure profile because of the clamping on the IVC, the hydrojector generated a pressure profile in IVC identical to that of conventional hydrodynamic injection (Figure 3e), suggesting a better efficacy achieved by the hydrojector in establishing the hydrodynamic pressure. Whereas conventional hydrodynamic tail-vein injection requires large volumes of the injection to drive a retrograde flow of the injected solution into the liver, the hydrojector achieves a significant reduction in the volume of injection required, because it eliminates the volume needed to fill the entire IVC and the heart chambers before reaching the liver. Hydrojector-assisted gene delivery into rat liver (tenfold larger than that of a mouse) resulted in luciferase gene expression at the 2.9 × 108 relative light unit/mg level (single peak pressure: 50 mm Hg; injection time: 10 seconds; injection volume: 5.4 ± 0.2% of BW), thereby proving the effectiveness of the device in larger animals. Hydrojector-assisted gene delivery to muscle and kidney Hydrojector-assisted gene delivery into muscle and kidney was also performed. With the pressure profile shown in Figure 4a, 1099
© The American Society of Gene Therapy
c
15
Hydrodynamic
10
0
Local
Systemic
10
109
Kidney
107
Local
Systemic
Local
30 20
Immunohistochemistry
10 0 10 20 Time (s)
0 10
20
30
d
120
60
0 10 20 Time (s)
Luciferase activity (RLU/mg)
200
30
f
g
h
1010 108 6
10
2.5
4
10
102
8.0
10
10
108 106 4
10
8.0
10
Figure 5 Comparison of the levels of green fluorescent protein (GFP) expression in muscle and kidney. Fluorescence microscope images were taken (a,c) 4 days after computer-assisted hydrodynamic delivery, or (b,d) slow infusion of Ad-GFP vectors in (a,b) muscle, and (c,d) kidney. Immunohistochemical examination of the (e) nontargeted kidney sample, and targeted kidney sample (f) without or (g,h) with anti-GFP antibody staining. Clear positive signals are seen, both in the (g) medulla and (h) cortex of the surrounding glomerulus. Original magnifications are ×4 (a–d), ×10 (e,f), and ×40 (g,h).
.26
2
Systemic Local
Figure 4 Hydrojector-assisted hydrodynamic gene delivery into muscle and kidney in rats. (a) Intrafemoral vein pressure during a continued injection (10 seconds) with programmed peak pressure at 350 mm Hg. (b) Comparison between conventional (systemic) and computer-assisted (local) hydrodynamic gene delivery in respect of levels of luciferase gene expression in muscle. (c) Intrarenal vein pressure of computer-assisted hydrodynamic injection into the right kidney with intravascular pressure set at 100 mm Hg. (d) Comparison between conventional (systemic) and computer-assisted (local) hydrodynamic gene delivery in respect of the levels of luciferase gene expression in kidney. The values shown on the bars in b and d represent the injection volumes presented as a percentage of total body weight. The data are presented as mean values ± SD, n = 3. RLU, relative light unit.
the level of reporter gene expression in targeted muscle cells after hydrojector injection was 10,000-fold higher than that from hydrodynamic tail-vein injection (Figure 4b). More important, the volume of DNA solution required decreased from 8 to 2.5% of BW (values on the bars in Figure 4b). Similarly, hydrojectorassisted gene delivery through the renal vein at a peak pressure of 100 mm Hg (Figure 4c) resulted in 100 times more transgene product in rat kidney than could be achieved by conventional 1100
e 30
Luciferase activity (RLU/mg)
Femoral vein pressure (mm Hg) Renal vein pressure (mm Hg)
b
400
0
d
8
10
Figure 3 Comparison of conventional (systemic) and computerassisted (local) hydrodynamic gene delivery. (a) Comparison in respect of injection volume. (b) Comparison of levels of luciferase gene expression. (c,d) Comparison of densities of transfected cells in the liver. (e) Comparison of pressure profiles in the inferior vena cava (IVC) (blue line, computer-assisted injection; grey line, conventional hydrodynamic injection through the tail vein). RLU, relative light unit.
c
c
d
0
0
b
Systemic
10
e
a
Slow
a Muscle
5
IVC pressure (mm Hg)
b
Luciferase activity (RLU/mg)
a
Injection volume (% of BW)
Computer-controlled Gene Delivery
hydrodynamic tail-vein injection (Figure 4d). Significantly, the injection volume for muscle and kidney was only 2.5 and 0.26% of BW (values on the bars in Figure 4b and d), respectively, substantially lower than the injection volumes reported earlier.21–24
Hydrojector-assisted delivery of viral vectors As in the case of conventional nonviral approaches, the efficiency of gene delivery by viral vectors into kidney and muscle is relatively low primarily because the endothelial barrier hinders the direct access of viral vectors into the parenchyma cells. Taking advantage of the permeabilization effect that hydrodynamic injection has on the endothelium,25,26 we examined the usefulness of the hydrojector in enhancing viral vector–mediated gene delivery, using rat muscle and kidney as the targets. For gene delivery into muscle, 5 × 1011 adenoviral particles containing the green fluorescent protein (GFP) gene were injected through the femoral vein. GFP expression was seen in the targeted hindleg 4 days after injection (Figure 5a), whereas none was detected in the hindleg of the same animal receiving slow infusion of the same number of viral particles (Figure 5b). Again using the hydrojector, the same number of viral particles was injected into rat kidney through the right renal vein. www.moleculartherapy.org vol. 16 no. 6 june 2008
© The American Society of Gene Therapy
Computer-controlled Gene Delivery
Figure 5c shows clear GFP expression in the targeted right kidney but not in the left kidney that received a slow infusion of the virus (Figure 5d). Immunohistochemical staining of kidney sections with anti-GFP antibodies revealed enriched GFP expression in both the cortex and the medulla of the hydrodynamically treated kidney (Figure 5g and h) (Figure 5e and f, negative control).
Hydrojector-assisted hydrodynamic gene delivery in pigs In order to evaluate the feasibility of using our computer-controlled injection device for gene delivery into liver and kidney in large (comparable in size to humans) animals we used swine as an animal model. A standard image-guided insertion procedure employed in the clinic was used for placing a balloon catheter into the right lateral hepatic vein or renal vein to target the liver or
b
d
150 100 50 0 0
e
10 20 30 Time (s)
40
7.4 × 106 3.7 × 106 6.1 × 106
300 200 100 0 0
10 20 30 Time (s)
Tissue damage assessment Tissue damage caused by computer-assisted hydrodynamic gene delivery into the liver, kidney, and muscle was evaluated in rats, using serum biochemistry (Figure 7) and histological examination. As found after hydrodynamic tail-vein injection, tissue-specific
40
Caudate lobe 2.9 × 103
Right lateral lobe 4.8 × 106 3.0 × 106 103
Right middle lobe
f
1.8 × 10 1.2 × 105 1.4 ×
3.1 × 103
Left middle lobe
5
105
ALT (U/l)
6.7 ×
a
Left lateral lobe
7.3 × 101
g
4.7 × 101
5.7 × 105
Figure 6 Image-guided hydrodynamic gene delivery to pig liver and kidney. (a,b) Fluoroscopic images of the pig liver and kidney showing the relative locations of the inserted balloon catheters [the injection balloon catheter (*), the occlusion balloon in portal vein (**), and the occlusion balloon in the IVC (***)]. The distribution of phase contrast medium injected through the injection catheter to verify obstruction of blood vessels by the balloon catheters is also indicated (****). (c) Lateral hepatic vein pressure profile during and soon after hydrodynamic injection with peak pressure set at 100 mm Hg. (d) Right renal vein pressure profile during and soon after hydrodynamic injection with peak pressure set at 240 mm Hg. (e) Photograph of hydrodynamically transfected pig liver in cross-section, showing the levels of luciferase gene expression at the approximate sites from which liver samples were collected. (f) Photograph of hydrodynamically transfected right kidney cut in half in the middle, showing the levels of luciferase gene expression at the sites from which tissue samples were collected. (g) Photograph of untreated left kidney cut in half in the middle, showing the background level of luciferase activity. The levels of luciferase gene expression were determined 6 hours after gene delivery. The data presented are from two pigs, one each for gene delivery to the liver and the kidney. The values in e–g represent luciferase activity in relative light unit/mg of extracted proteins.
Molecular Therapy vol. 16 no. 6 june 2008
104 103 102 101 100
b
2
Crt (mg/dl)
Hepatic vein pressure (mm Hg)
c
* ***
**
Right renal vein pressure (mm Hg)
*
*** *
** *
1
c
300
CPK (U/l)
a
kidney, respectively. Figure 6a and b demonstrates the relative positions of the inserted balloon catheters. In order to minimize leakage of the injected DNA solution into the portal vein and/or IVC, occlusion balloons were inserted from the superior mesenteric vein to the portal vein for the liver, and/or from the right femoral vein to the IVC for liver (located between diaphragm and hepatic vein) and kidney (covering the renal–IVC junction). Complete occlusion was confirmed by injection of a small amount of phase-contrast medium (gray area in Figure 6a beyond the balloon of the injection catheter). Figure 6c and d shows the pressure profiles of injection into the right hepatic vein and right renal vein, respectively. The total injection volume was 800 ml delivered in 20 seconds (40 ml/s) for the liver, and 55 ml in 11.7 seconds (4.7 ml/s) for the kidney. Figure 6e shows reporter gene expression in the liver in both targeted (right lateral and caudate) and nontargeted (right medial, left medial, and left lateral) lobes. Figure 6f and g reveals luciferase activity in the treated and untreated kidneys, respectively. Luciferase gene expression was seen in both targeted and nontargeted liver lobes because of the inner-lobular vascular connection. The level of luciferase gene expression seen in the targeted right lateral lobe, however, was 1,000-fold higher than that in the nontargeted lobe. A significant level of luciferase gene expression was seen in the hydrodynamically treated kidney (Figure 6f), but not in the untreated one (Figure 6g).
200
0
100 0
0
100 Time (h)
200
Figure 7 Assessment of tissue damage after hydrodynamic gene delivery using the computer-controlled injection device. Serum concentrations of tissue-specific marker enzymes in rats [(a) alanine aminotransferase (ALT) for liver, (b) creatinine (Crt) for kidney, and (c) creatinine phosphokinase (CPK) for muscle] at 6, 24, 48, 72, 120, and 168 hours after (closed circles) computer-assisted hydrodynamic gene delivery, or after (open circles) a slow (60 seconds) infusion of the same volume of saline. Open bars in each plot represent the serum concentration of marker enzyme in untreated animals. The data are presented as mean values ± SD, n = 3.
1101
Computer-controlled Gene Delivery
enzyme concentration was initially high after hydrojector-assisted injection, and decreased rapidly with time. At 24 hours, the serum concentration of alanine aminotransferase, a liver-specific marker, was three- to fourfold higher in hydrodynamically treated animals than in untreated animals (225 ± 131 U/l versus 62 ± 16 U/l) (Figure 7a). Similarly, the concentrations of the kidney- and muscle-specific markers, creatinine and creatine phosphokinase, 24 hours after hydrodynamic treatments, were 0.3 ± 0.04 mg/dl and 82 ± 27 U/l, respectively (Figure 7b and c), which was similar to the levels in control animals that had received no treatment (open bar insert). All the elevated values of enzyme concentrations in treated animals returned to the normal ranges within 3 days, which is in full agreement with results obtained with conventional hydrodynamic delivery.1,27,28 Histological evaluation of the rat livers revealed no obvious tissue damage (data not shown). Examination of the targeted kidney revealed no obvious bleeding, tubular cell detachment, apoptosis, necrosis, or inflammatory cell infiltration in either the cortical or medullary regions. The interlobular vein also showed no obvious abnormality. Muscles in the targeted hindleg revealed no obvious tissue damage such as edema, torn bundles, or inflammatory cell infiltration. Muscle fibers with multiple nuclei and regular cross-striation patterns were clearly observed, indicating mature muscles without obvious sarcomere unit destruction. All these results confirm previous findings in mice, that the structural impact of hydrodynamic gene delivery is transient, and the procedure is safe.1,26
DISCUSSION We have demonstrated in this study the successful development of a computer-controlled injection device for hydrodynamic gene delivery. The system uses negative feedback from the vascular pressure at the injection site to precisely control the injection, based on a preselected pressure profile for maximal gene transfer efficiency and minimal tissue damage. The most significant feature of our design is to use intravascular pressure as the key regulator for the injection. The injection shuts down automatically when the desired intravascular pressure is reached. Similarly, when the pressure falls below a preselected level, the injection automatically resumes. Intravascular pressure at the injection site can thereby be maintained within any desired range over the duration of the injection. Because the injection is controlled by the intravascular pressure, another important feature of this injection device is that the volume needed in order to develop a sufficiently elevated pressure, and thereby achieve a successful gene transfer, is self-adjusting based on the size and anatomical structure of the selected organ of an individual. This system can be used for both viral and nonviral gene delivery, and provides a potential method for safe, reliable, and efficient in vivo gene delivery. One important finding made during this study was that the pressure profiles that are effective for gene delivery vary from one target organ to another, and also from one animal species to another. Larger animals tend to require higher pressure for successful gene transfer to a given target organ. Although such interspecies differences could be caused by differences in the structure of the endothelium and parenchyma cell membrane of the target organs, it is likely that higher pressure required for larger organs is because of the low elasticity of the vasculature and parenchyma 1102
© The American Society of Gene Therapy
cells. The significantly high pressure and relatively low luciferase activity associated with kidney and muscle appears to bear out such a notion. Compared to the liver, both kidney and muscle have much lower elasticity. More research is needed to determine whether the turbulence generated by the injection inside the vasculature influences the efficiency of gene delivery. As far as the applications of the hydrojector system are concerned, future studies must establish optimal pressure profiles for each targeted organ so as to enable the development of computer programs that will guide safe and efficient gene delivery. With continuing efforts, and with new data being generated, it is foreseeable that an effective and clinically applicable hydrodynamic procedure will be established.
MATERIALS AND METHODS All experiments performed on mice, rats, and pigs were approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh (Pittsburgh, Pennsylvania). Conventional hydrodynamic gene delivery was performed in accordance with the procedure published earlier.2 Pressure changes in the IVC during conventional hydrodynamic injection through the tail vein were measured according to the method described by Zhang et al.27 Materials. Plasmids of pCMV-Luc and pCMV-LacZ were purified using
CsCl-ethidium bromide density gradient ultracentrifugation and stored in Tris–EDTA buffer. The purity of the plasmids was verified by absorbency at 260 and 280 nm and 1% agarose gel electrophoresis. Ad-GFP vectors were generously provided by Xiao Xiao (University of North Carolina at Chapel Hill). A single batch of high titer adenovirus stock (1012 viral particles/ml) was used throughout the experiments. The luciferase assay kit was from Promega (Madison, WI). CD-1 mice (18–20 g, female) and Wister rats (180–200 g, female) were from Charles River (Wilmington, MA), and pigs (20–30 kg) were from Wally Whippo (Enon Valley, PA). The 12 F sheath for image-guided catheter insertion was from COOK (Bloomington, IN), and guidewire was from Terumo Medical (Somerset, NJ). The 9 F balloon catheter for injection in pigs was custom-made by Clinical Supply (Gifu, Japan). The occlusion balloon catheters were purchased from Medtronic (Minneapolis, MN). Mikro Tip catheter transducer was from Millar Instruments (Houston, TX). Construction of the computer-controlled injection device. The injection
device was built according to the illustration shown in Figure 1. The computer program running the entire system was developed in-house to monitor pressure every 100 ms and operate the opening and closing of the valve at intervals of 250 ms. A regular CO2 tank, commonly used for cell culture, was used as the solution driver. The miniature pressure transducer (SPR671 or SPR-407) was from Millar Instruments. A solenoid or pneumatic valve was placed between the solution container and the catheter, and the opening and closing of the valve were controlled through activation and inactivation of a power supply (RK-80; Matsusada Precision, Shiga, Japan), regulated by transistor-transistor-logic signals generated from a signal generator (ADA16-32/2(CB) F; Contec, Osaka, Japan). Computer-assisted gene delivery to the liver. A midline incision was made
in the abdomen of anesthetized mice and rats to expose the internal organs. A 22G peripheral catheter (Terumo Medical, Somerset, NJ) was inserted into the IVC near the bifurcation of the iliac veins, and advanced further, above the confluence with the left renal vein, at which point a vessel clamp (AROSurgical, Newport Beach, CA) was placed to prevent backflow of the injected solution. A pressure transducer was inserted by means of a threeway adaptor into the IVC through the inserted catheter. Saline with or without plasmid DNA (pCMV-Luc, 10 μg/ml) was injected, and the pressure profile and injection volume were recorded. During injection, the IVC www.moleculartherapy.org vol. 16 no. 6 june 2008
© The American Society of Gene Therapy
was manually compressed using a cotton swab at the section immediately below the diaphragm. The total occlusion time was 40 seconds including the injection time. At the end of the procedure, hemostasis was established by applying gelatin sponge and n-butyl cyanoacrylate adhesive (3M Health Care, St Paul, MN) at the insertion site, and the midline incision was sutured layer-by-layer using 4-0 silk. Computer-assisted gene delivery to kidney. Similar to the procedure
undertaken for the liver, a midline incision was made in the abdomen of anesthetized rats. A peripheral catheter (22G) was inserted through the sidewall of the IVC into the right renal vein. After insertion of the pressure transducer through the inserted catheter, the right renal vein was taped using 4-0 silk and temporarily tied off. The peak pressure was set at 100 mm Hg, and the injection was administered, consisting of saline with plasmid DNA (100 μg/ml) or Ad-GFP viral vectors (5 × 1011 viral particle in volume of 0.25% BW). The same procedure as that used for liver was employed for hemostasis and incision closure. Computer-assisted gene delivery to muscle. On an anesthetized rat, a
midline skin incision was made in the lower abdomen. A 22G peripheral catheter was inserted into the exposed femoral vein in a hindleg near the bifurcation from the external iliac vein, and advanced into the medial saphenous vein. After inserting a pressure transducer through the inserted catheter, both the femoral vein and artery were taped and temporarily tied off using a 4-0 silk. Hydrodynamic injection of saline containing pCMVLuc (100 μg/ml) or Ad-GFP viral particles (5 × 1011 viral particles in volume of 3% BW) in phosphate-buffered saline was administered. At the end of the procedure, hemostasis was achieved by simple compression with sterile gauze, and the incision was closed. Image-guided hydrodynamic gene delivery to liver and kidney in swine.
For gene delivery to the liver, the animal was put under general anesthesia, and laid out on a fluoroscopy table. Sterile drapes were placed appropriately, and a midline skin incision was made near the mandibular arch to expose the jugular vein. An 18G peripheral catheter was inserted over a stylet into the exposed jugular vein. The stylet was removed and a 0.035-in hydrophilic guide wire was inserted and advanced through the catheter into the IVC under fluoroscopic guidance. Retaining the guide wire in place, the peripheral catheter was replaced with a 12 F-size short sheath. After removing the guide wire and washing out blood from the sheath with saline/heparin, a 9 F-size balloon catheter was inserted through the sheath into the right lateral lobe of the liver, followed by insertion of the pressure transducer. Two occlusion balloons (8 F, 46-mm diameter when fully inflated) were inserted, one to occlude passage from the femoral vein to the IVC and the other to occlude passage from the superior mesenteric vein into the main portal trunk, in order to block leakage of injected solution into the IVC and the portal vein, respectively. The occlusion balloons were inflated by infusion of phase-contrast medium. A small volume of contrast medium was injected to verify a complete occlusion. Saline containing plasmid DNA (100 μg/ml) was injected through the hydrojector with peak pressure set at 100 mm Hg. For gene delivery to the kidney, the catheter insertion was made into the right renal vein following the same procedure, with the exception that only the IVC occlusion balloon was utilized, and the peak pressure was set at 240 mm Hg. Intravascular pressure was continuously monitored and recorded during and soon after the injection. The volume actually injected was calculated on the basis of the amount of solution in the solution container before and after injection. Reporter gene expression assay. The level of reporter gene expression was
determined using procedures established earlier.1,27 In the case of mouse liver, mouse kidney, or rat kidney for the purpose of luciferase assay, the whole organ was collected whereas, for rat liver, pig liver, and pig kidney, tissue cubes (~200 mg) were obtained from the respective organs. In animals that had received hydrodynamic treatment in the legs, tissue samples were obtained from the quadriceps femoris, biceps femoris, tibialis anterior, and Molecular Therapy vol. 16 no. 6 june 2008
Computer-controlled Gene Delivery
gastrocnemius muscles. The tissues were immediately frozen in powdered dry ice and stored at −80 °C until use. Histochemical β-galactosidase assay on mouse liver was performed in accordance with the method described earlier.1 Immediately after tissue sectioning, the expression of GFP was evaluated under a fluorescence microscope.25 Immunohistochemistry for GFP expression in kidney was performed on 4-μm thick formalin-fixed paraffin-embedded kidney sections using biotin–avidin enzyme conjugates and rabbit anti-GFP antibody of ab290 (Abcam, Cambridge, MA) at a 1,000-fold dilution overnight at 4 °C. The distribution of the primary antibody was visualized through avidin–biotin complex amplification using a Vectastain Elite ABC Kit (Vector, Burlingame, CA) with 3, 3ʹ-diaminobenzidine as chromaogen (Vector). Assessment of tissue damage. Saline was injected using the hydrojector to
target the muscle, kidney, and liver in separate rats. At 6, 24, 48, 72, 120, and 168 hours after the injection, blood samples were collected from the lateral tail vein for determination of creatinine and creatine phosphokinase concentrations, and from the retro-orbital plexus for quantification of alanine aminotransferase. Serum concentrations of the marker enzymes were determined using commercial kits (Stanbio Laboratory, Boerne, TX) in accordance with the manufacturer’s instructions. Histological evaluation was performed using formalin-fixed paraffin-embedded sections of the target organ obtained 24 hours after the injection. Hematoxylin and eosin staining was performed for the liver and muscles, whereas periodic acid-Schiff stain was employed for the kidney.
ACKNOWLEDGMENTS We thank Joseph E. Knapp for his critical reading of the manuscript. This work was supported in part by grants from National Institutes of Health (EB2002946 and HL075542).
REFERENCES
1. Liu, F, Song, Y and Liu, D (1999). Hydrodynamics-based transfection in animals by systemic administration of plasmid DNA. Gene Ther 6: 1258–1266. 2. Zhang, G, Budker, V and Wolff, JA (1999). High levels of foreign gene expression in hepatocytes after tail vein injections of naked plasmid DNA. Hum Gene Ther 10: 1735–1737. 3. Maruyama, H, Higuchi, N, Nishikawa, Y, Hirahara, H, Iino, N, Kameda, S et al. (2002). Kidney-targeted naked DNA transfer by retrograde renal vein injection in rats. Hum Gene Ther 13: 455–468. 4. Zhang, G, Song, YK and Liu, D (2000). Long-term expression of human α1-antitrypsin gene in mouse liver achieved by intravenous administration of plasmid DNA using a hydrodynamics-based procedure. Gene Ther 7: 1344–1349. 5. Miao, CH, Thompson, AR, Loeb, K and Ye, X (2001). Long-term and therapeutic-level hepatic gene expression of human factor IX after naked plasmid transfer in vivo. Mol Ther 3: 947–957. 6. Stoll, SM, Sclimenti, CR, Baba, EJ, Meuse, L, Kay, MA and Calos, MP (2001). EpsteinBarr virus/human vector provides high-level, long-term expression of α1-antitrypsin in mice. Mol Ther 4: 122–129. 7. Yew, NS, Przybylska, M, Ziegler, RJ, Liu, D and Cheng, SH (2001). High and sustained transgene expression in vivo from plasmid vectors containing a hybrid ubiquitin promoter. Mol Ther 4: 75–82. 8. Alino, SF, Crespo, A and Dasi, F (2003). Long-term therapeutic levels of human α-1 antitrypsin in plasma after hydrodynamic injection of nonviral DNA. Gene Ther 10: 1672–1679. 9. Chen, ZY, He, CY, Ehrhardt, A and Kay, MA (2003). Minicircle DNA vectors devoid of bacterial DNA result in persistent and high-level transgene expression in vivo. Mol Ther 8: 495–500. 10. Score, PR, Belur, LR, Frandsen, JL, Guerts, JL, Yamaguchi, T, Somia, NV et al. (2006). Sleeping Beauty-mediated transposition and long-term expression in vivo: use of the LoxP/Cre recombinase system to distinguish transposition-specific expression. Mol Ther 13: 617–624. 11. Dai, C, Yang, J and Liu, Y (2002). Single injection of naked plasmid encoding hepatocyte growth factor prevents cell death and ameliorates acute renal failure in mice. J Am Soc Nephrol 13: 411–422. 12. Chen, L and Woo, SL (2005). Complete and persistent phenotypic correction of phenylketonuria in mice by site-specific genome integration of murine phenylalanine hydroxylase cDNA. Proc Natl Acad Sci USA 102: 15581–15586. 13. Suda, T and Liu, D (2007). Hydrodynamic gene delivery: its principles and applications. Mol Ther 15: 2063–2069. 14. Herweijer, H and Wolff, JA (2007). Gene therapy progress and prospects: hydrodynamic gene delivery. Gene Ther 14: 99–107. 15. Zhang, X, Dong, X, Sawyer, GJ, Collins, L and Fabre, JW (2004). Regional hydrodynamic gene delivery to the rat liver with physiological volumes of DNA solution. J Gene Med 6: 693–703. 16. Inoue, S, Yakamata, Y, Kaneko, M and Kobayashi, E (2004). Gene therapy for organ grafts using rapid injection of naked DNA: application of the rat liver. Transplantation 77: 997–1003.
1103
Computer-controlled Gene Delivery
17. Eastman, SJ, Baskin, KM, Hodges, BL, Chu, Q, Gates, A, Dreusicke, R et al. (2002). Development of catheter-based procedures for transducing the isolated rabbit liver with plasmid DNA. Hum Gene Ther 13: 2065–2077. 18. Yoshino, H, Hashizume, K and Kobayashi, E (2006). Naked plasmid DNA transfer to the porcine liver using rapid injection with large volume. Gene Ther 13: 1696–1702. 19. Alino, SF, Herrero, MJ, Noguera, I, Dasi, F and Sanchez, M (2007). Pig liver gene therapy by noninvasive interventionist catheterism. Gene Ther 14: 334–343. 20. Fabre, JW, Grehan, A, Whitehorne, M, Sawyer, GJ, Dong, X, Salehi, S et al. (2007). Hydrodynamic gene delivery to the pig liver via an isolated segment of the inferior vena cava. Gene Ther 15: 452–462. 21. Budker, V, Zhang, G, Danko, I, Williams, P and Wolff, J (1998). The efficient expression of intravascularly delivered DNA in rat muscle. Gene Ther 5: 272–276. 22. Su, LT, Gopal, K, Wang, Z, Yin, X, Nelson, A, Kozyak, BW et al. (2005). Uniform scale-independent gene transfer to striated muscle after transvenular extravasation of vector. Circulation 112: 1780–1788.
1104
© The American Society of Gene Therapy
23. Hagstrom, JE, Hegge, J, Zhang, G, Noble, M, Budker, V, Lewis, DL et al. (2004). A facile nonviral method for delivering genes and siRNAs to skeletal muscle of mammalian limbs. Mol Ther 10: 386–398. 24. Danialou, G, Comtois, AS, Matecki, S, Nalbantoglu, J, Karpati, G, Gilbert, R et al. (2005). Optimization of regional intraarterial naked DNA-mediated transgene delivery to skeletal muscles in a large animal model. Mol Ther 11: 257–266. 25. Brunetti-Pierri, N, Palmer, DJ, Mane, V, Finegold, M, Beaudet, AL and Ng, P (2005). Increased hepatic transduction with reduced systemic dissemination and proinflammatory cytokines following hydrodynamic injection of helper-dependent adenoviral vectors. Mol Ther 12: 99–106. 26. Suda, T, Gao, X, Stolz, DB and Liu, D (2007). Structural impact of hydrodynamic injection on mouse liver. Gene Ther 14: 129–137. 27. Zhang, G, Gao, X, Song, YK, Vollmer, R, Stolz, DB, Gasiorowski, JZ et al. (2004). Hydroporation as the mechanism of hydrodynamic delivery. Gene Ther 11: 675–682. 28. Kobayashi, N, Nishikawa, M, Hirata, K and Takakura, Y (2004). Hydrodynamics-based procedure involves transient hyperpermeability in the hepatic cellular membrane: implication of a nonspecific process in efficient intracellular gene delivery. J Gene Med 6: 584–592.
www.moleculartherapy.org vol. 16 no. 6 june 2008