Skip to main content
NCBI home page
Molecular Therapy logoLink to Molecular Therapy
. 2009 Jan 20;17(3):491–499. doi: 10.1038/mt.2008.294

Image-guided, Lobe-specific Hydrodynamic Gene Delivery to Swine Liver

Kenya Kamimura 1, Takeshi Suda 1,*, Wei Xu 1, Guisheng Zhang 1, Dexi Liu 1
PMCID: PMC2680706  NIHMSID: NIHMS112994  PMID: 19156134

Abstract

Image-guided, lobe-specific hydrodynamic gene delivery to liver was assessed in pigs. The procedure involved image-guided insertion of a balloon catheter to the hepatic vein of the selected lobe from the jugular vein and hydrodynamic injection of plasmid DNA using a newly developed computer-controlled injection device. We demonstrated that the impact of the procedure was regional with minimal effects on neighboring lobes. Level of gene expression resulted from the procedure was 107 relative light units (RLU)/mg in the targeted lobes and 102–105 RLU/mg in the nontargeted lobes 4 hours after hydrodynamic injection of pCMV-Luc plasmids. Occlusion of blood flow in the inferior vena cava (IVC) or IVC plus portal vein (PV) was effective in elevating hydrodynamic pressure in the targeted vasculature but did not enhance gene delivery efficiency. Physiological examination on pigs with IVC occlusion revealed transient decreases of blood pressure and respiration rate. Removal of occlusion from IVC resulted in a rapid and transient increase in heart rate. Occlusion of the PV and hepatic vein showed no effect on physiological and cardiac activities. No major changes in serum composition were observed. These results suggest that (i) image-guided, lobe-specific hydrodynamic procedure is effective for regional gene delivery to liver, (ii) blockade in IVC should be avoided for hydrodynamic gene delivery to the liver, and (iii) clinical application of hydrodynamic gene delivery to liver is feasible.

Introduction

Nine years ago we and Zhang et al. published the method of hydrodynamic gene delivery as a simple and effective means to transfect mouse hepatocytes in vivo.1,2 Subsequently, this method has been used widely in research laboratories for gene expression, gene knockdown, and functional analysis of genetic elements in whole animals (for recent reviews, see refs. 3,4). More recently, the hydrodynamics-based procedure has been used to enhance the transduction efficiency of lenti,5 adeno,6,7 and adeno-associated8 viral vectors. While commonly used for gene therapy studies in rodents, the hydrodynamics-based procedure has not been considered feasibly for gene delivery in large animals or humans because injection of a volume of 8–10% of body weight in volume (~5.6–7 l for a 70-kg man) within a few seconds was considered impractical and unsafe. In fact, we have shown that an acute overload of the systemic circulation with a large volume of DNA solution can cause significant cardiovascular dysfunction.9 To overcome this problem, a few laboratories have made significant efforts in developing modified procedures for hydrodynamic gene delivery. For example, under X-ray guidance Eastman et al. have shown in rabbits that a volume of 15 ml/kg can be safely injected into an isolated rabbit liver.10 Kobayashi's group reported delivery of the green fluorescence protein gene into the left lateral lobe (LLL) of pig liver by catheterization and occlusion of the portal vein (PV).11 A similar study using the human α1-antitrypsin gene as a reporter was reported by Alino et al.12 More recently, Fabre et al.13 have investigated the possibility of hydrodynamic gene delivery to the whole liver from the inferior vena cava (IVC) with double occlusions at the section above and below the IVC–hepatic vein conjunction. Although the overall level of gene expression demonstrated by these studies was low compared to that of the conventional method of hydrodynamic tail vein injection in mice, these results have provided direct evidence in support that hydrodynamic gene delivery to the liver can be achieved in large animals. The challenges, however, are to identify the factors that influence hydrodynamic gene delivery efficiency in a human-size animal model and to devise a safe and effective procedure that is clinically applicable.

Toward this end, we have recently developed a computer-controlled injection device enabling hydrodynamic gene delivery to both small and large animals.14 The device uses high pressure from a gas cylinder to drive DNA into the animal and a computer-controlled on/off switch to regulate the injection. Ultimate control over the injection is achieved through a computer program that uses the vascular pressure transmitted by an internally inserted transducer as the start/stop signal for hydrodynamic injection. In a previous study, we demonstrated that this new computer-controlled injection system was effective in hydrodynamic gene delivery to the liver, kidney, and muscle cells in rodents and pigs.14 In the current study, we explored the possibility of regional hydrodynamic gene delivery to pig liver by combining the technology of image-guided catheterization with the computer-controlled hydrodynamic gene delivery. The objectives of this study were to assess the feasibility of image-guided regional hydrodynamic gene delivery, gene delivery efficiency, and the effects of the procedure on physiological and cardiac functions. The long-term goal of our research is to establish a minimally invasive, clinically applicable hydrodynamic procedure for human gene therapy.

Results

Anatomic features of the pig liver for hydrodynamic gene delivery

Swine have been utilized as an animal model for studies of human liver diseases and for liver transplantation.15 Swine liver consists of five lobes including the right lateral lobe (RLL), right medial lobe (RML), left medial lobe (LML), LLL, and caudate lobe (CL) (Figure 1a) and is wrapped in a tough fibrous capsule that branches and extends throughout the liver parenchyma. With the exception of the small CL that contains part of the IVC within its parenchyma and drains blood directly to the IVC, all other liver lobes have their own primary hepatic veins to drain their blood into the IVC. Figure 1b shows a cross-section of the pig liver demonstrating the structure and location of the major blood vessels including the right lateral hepatic vein (RLHV), right medial hepatic vein, left medial hepatic vein, and left lateral hepatic vein. Two major branches of the hepatic veins are seen in each lobe. Branching points into smaller blood vessels are clearly visible at the endothelial surface of the hepatic veins (Figure 1c) and the CL-wrapped IVC (Figure 1d). For image-guided regional hydrodynamic gene delivery to the liver, we aimed at insertion of a balloon catheter into a hepatic vein to deliver gene into cells of the target lobe.

Figure 1.

Figure 1

Open in a new tab

Anatomic features of pig liver. (a) Pig liver showing relative position of each of the five lobes and their conjunction. (b) Cross-section of pig liver showing major blood vessels in each lobe. (c) Endothelial surface of the hepatic vein with branching sites (arrowheads) for smaller blood vessel. (d) Endothelial surface of the inferior vena cava (IVC) inside the caudate lobe showing branching sites (arrowheads) for smaller blood vessels. LLHV, left lateral hepatic vein; LMHV, left medial hepatic vein; RLHV, right lateral hepatic vein; RMHV, right medial hepatic vein.

Regional impact of image-guided hydrodynamic injection

The RML was selected as the target to evaluate the impact of image-guided hydrodynamic injection. To maximize the effect of the hydrodynamically injected solution, occlusion balloon catheters were placed in the IVC and PV to prevent leakage of the injected solution into these two major blood vessels that receive blood from and provide blood to the liver, respectively. The balloon catheter utilized for injection prevented backflow of the injected solution and aided in establishing an elevated hydrodynamic pressure. Insertion sites of the injection balloon catheter and occlusion balloons are shown in Figure 2a. To visualize the distribution of the injected solution in the liver, we injected into the RML with a phase contrast medium. Image shown in Figure 2a clearly demonstrates that the injected solution was retained in the RML.

Figure 2.

Figure 2

Open in a new tab

Regional impact of lobe-specific hydrodynamic injection. Hydrodynamic injection was made to the right medial lobe (RML) (injection pressure, 300 psi; injection time, 15 seconds; injection volume, 600 ml) through the preinserted injection balloon catheter. (a) Fluoroscopic image showing location of balloon of injection catheter (arrow), occlusion balloon in inferior vena cava (IVC) (*) and in portal vein (**), and distribution of injected contrast medium in the RML. (b) Impact of hydrodynamic injection on lobe size before and 2.5, 15, and 25 seconds after injection. (c) Time-dependent volume change of the targeted RML and the nontargeted right lateral lobe (RLL).

A similar experimental setup to that described in Figure 2a was used to verify the regional impact of the image-guided hydrodynamic injection and to assess the potential impact of hydrodynamic injection on the neighboring lobes. Photo images in Figure 2b show the appearance of the RML (targeted) and RLL (nontargeted) at selected time points upon hydrodynamic injection of 600 ml in 15 seconds. Compared to the original size (time 0), the RML volume increased ~150% at 2.5 seconds, 270% at the end of injection (15 seconds), and 175% 10 seconds after the injection. The RML volume reached the maximum (~2.7-folds of the original) in ~10 seconds (Figure 2c) and started to recover soon after the injection. In contrast, the volume of the nontargeted RLL was largely unchanged. These results verify that the impact of the image-guided hydrodynamic procedure is regional and lobe specific.

Effect of the image-guided, lobe-specific hydrodynamic injection on physiological functions

The effects on selected physiological parameters and cardiac activity of the regional hydrodynamic injection into the four primary liver lobes were examined. Six pigs were used in the study and each received one injection into a selected lobe. Figure 3a–f show the insertion site of the balloon catheter inside the target lobe and the position of the occlusion balloons at the IVC and PV (Figure 3e,f). Pressure changes in each targeted hepatic vein are shown in Figure 3g–l. With the exception of injection to the RLHV (Figure 3g), which resulted in an irregularly shaped pressure profile reaching peak pressure (75 mm Hg) in 5 seconds and declining slowly thereafter, the pressure–time curves for the RML, LML, and LLL appear similar, reaching the peak pressure (75 mm Hg) soon after initiation of injection and remaining in the peak range until the end of injection. Occlusion of the IVC (Figure 3k) or IVC plus PV (Figure 3l) elevated the peak pressure to 100–125 mm Hg. No change in heart rate was induced by the injection (Figure 3m–r). However, animals with blood occlusion at the IVC or IVC plus PV showed rapid decreases in blood pressure (Figure 3w,x) and respiration rate (Figure 3z3,z4). On occlusion, systolic blood pressure decreased from 130 to 70 mm Hg and diastolic blood pressure from 76 to 38 mm Hg in 5 seconds, but both returned to normal range in 5 seconds after the injection and removal of the occlusion balloon. The respiratory rate followed the exactly same pattern, decreasing from ~30 to 4 times/min and returning to the normal range soon after the releasing of the occlusion.

Figure 3.

Figure 3

Open in a new tab

Effect of lobe-specific hydrodynamic injection on physiological functions. Hydrodynamic injection into right lateral lobe (RLL), right medial lobe (RML), left medial lobe (LML), and left lateral lobe (LLL) was performed separately (injection pressure, 300 psi; injection time, 15 seconds; injection volume; 600 ml) with occlusion of the inferior vena cava (IVC) or the IVC plus portal vein (PV), or without occlusion. The black dots in a–f show the balloon location of the injection catheter in each targeted lobe. Location of the occlusion balloons is marked in e and f (arrows). Time-dependent change upon hydrodynamic injection on intravascular pressure was detected via the transducer inserted inside the hepatic vein through the interior of the injection catheter. Changes in heart rate, systolic (upper line) and diastolic (lower line) blood pressures, and respiration rate upon hydrodynamic injection were shown in g, m, s, and y for injection to RLL; in h, n, t, and z for injection to the RML; in i, o, u, and z1 for the LML; in j, p, v, and z2 for injection to the LLL; in k, q, w, and z3 for the RML with IVC occlusion, and in l, r, x, and z4 for injection to the RML with occlusions at the IVC plus the PV, respectively. Data are representative of measurements on two pigs.

A further study was conducted to identify the direct cause for sudden drop in blood pressure and respiration rate in the test animals (Figure 3w,x,z3, and z4). We focused on the effect of IVC obstruction because both occlusion procedures (IVC alone and IVC plus PV) exhibited the same pattern of abnormality. An IVC occlusion was established by inserting a balloon catheter from the femoral vein to the IVC–hepatic vein conjunction followed by balloon inflation. On holding the occlusion for a total of 20 seconds without hydrodynamic injection, the respiration rate dropped from 30 to 4 times/min in the first 10 seconds and the animal stopped breathing thereafter for the remaining 10 seconds. Respiration returned to a rate of 30 times/min within 10 seconds after deflation of the occlusion balloon. Similarly, blood pressure dropped upon occlusion and returned to normal range in an identical time frame (data not shown). Occlusion of the PV alone did not induce any of the abnormal blood pressure and breathing patterns, suggesting that IVC occlusion is the direct cause for the abnormal blood pressure and irregular respiration.

The impact of rapid deflation of the balloon was also examined. Results in Figure 4 show the effect of occlusion removal from the IVC on cardiac activity. Electrocardiograms given in Figure 4a show that a rapid deflation of the IVC occlusion balloon (<3 seconds) induced an ST segment elevation (4–10 seconds after the deflation) and increase of heart rate. Heart rate increased from ~100 to 160 bpm in 25 seconds, but returned to the normal range 1 minute after the removal of IVC balloon (Figure 4b). Figure 4c shows the change in cardiothoracic ratio revealed by chest X-ray during and soon after balloon deflation. The cardiothoracic ratio value decreased slightly to 45% during deflation, increased to 68% at the end of deflation, and returned to 48% in 30 seconds after removal of the occlusion balloon. These results suggest a significant cardiac stress induced by acute overflow of blood into the heart. No abnormal cardiac activity was seen with obstruction of PV (data not shown). In addition, cardiac effects can be significantly reduced if balloon deflation was done slowly (>20 seconds).

Figure 4.

Figure 4

Open in a new tab

Effect of occlusion removal from the inferior vena cava (IVC) on cardiac activity. An occlusion balloon (8 Fr) was inserted from the femoral vein into the IVC and the balloon was inflated for 20 seconds. (a) Electrocardiogram at different time after deflation of balloon. (b) Time-dependent heart rate after removal of occlusion balloon. (c) Cardiothoracic ratio (CTR) before, during, and after removal of occlusion balloon. Data are representative of results obtained from three pigs.

Efficiency of regional hydrodynamic gene delivery

The level of luciferase gene expression in all hepatic lobes was examined 4 hours after image-guided hydrodynamic gene delivery to each of the four major hepatic lobes. Data summarized in Figure 5a–f show the average luciferase activity in each lobe. Among the targeted lobes, the lowest gene expression was seen in the RLL with an average luciferase activity at 104–105 relative light units (RLU)/mg, similar to the level in the CL and 100-folds higher than that of other nontargeted lobes (Figure 5a). Luciferase activity in the RML, LML, and LLL, when targeted, was ~107 RLU/mg (Figure 5b for RML, Figure 5c for LML, and Figure 5d for LLL). Levels of luciferase gene expression in nontargeted lobes were generally low, 1,000- to 100,000-fold less than that of the targeted lobe. Obstruction of the IVC or IVC plus PV did not result in increase in luciferase gene expression (Figure 5e,f) in both targeted and nontargeted lobes (Figure 5b versus Figure 5e,f). Immunohistochemistry of plasmid DNA–injected lobe and noninjected lobe was performed. The estimated number of positive cells identified by this method is <2% (Supplementary Figure S1a), and no cells were stained positive in the nontargeted lobe (Supplementary Figure S1b).

Figure 5.

Figure 5

Open in a new tab

Level of reporter gene expression 4 hours after hydrodynamic gene delivery. The selected lobe was injected in 15 seconds with 600 ml of saline containing pCMV-Luc plasmid DNA (100 µg/ml) with the injection pressure at 300 psi. Four hours after gene delivery, the animals were killed and a total of 15 samples from each lobe (five samples from the caudate lobe because of its small size) were collected for luciferase activity analysis. Black bars represent the level of reporter gene expression of tissue samples collected from targeted lobe and the gray bars represent that in nontargeted lobes (mean ± SD). Animals were hydrodynamically injected with pCMV-Luc plasmid into the (a) right lateral lobe (RLL, n = 2); (b) right medial lobe (RML, n=5); (c) left medial lobe (LML, n = 2); (d) left lateral lobe (LLL, n = 2); (e) RML (n = 2) with occlusion of the inferior vena cava (IVC); and (f) the RML (n = 2) with occlusion of the IVC plus portal vein (PV).

Effect of multilobe hydrodynamic injection

To maximize the overall transgene expression in the liver, we explored the possibility of sequential image-guided hydrodynamic gene delivery to multilobes. To minimize the total volume injected to each animal, we used 10 seconds as injection time, 5 seconds shorter than that for single lobe injection. The total injection volume per injection was 400 ml. Two separate hydrodynamic injections to RML and LML were performed on anesthetized animals with a time interval of 40 minutes. Figure 6a shows fluoroscopic images of the inserted balloon catheter in the liver. The pressure profiles for each injection are shown in Figure 6b. For each injection, the electrocardiogram and respiratory pattern were recorded before, during, and after the injection, and appeared normal (Figure 6c). The chest X-ray of the heart did not show any obvious cardiomegaly (Figure 6d). Heart and respiration rate, and blood pressure were also normal (data not shown). These results suggest that sequential hydrodynamic injection into two separate lobes is safe and can be performed on the same animal.

Figure 6.

Figure 6

Open in a new tab

Effect of sequential hydrodynamic injection into the right medial and left medial lobes on cardiac activity. Sequential hydrodynamic injections were performed firstly into the right medial lobe (RML) and 40 minutes later into the left medial lobe (LML) of the same animal under anesthesia. (a) Fluoroscopy images of the liver showing the position of the inserted balloon in the targeted lobes. (b) Intravascular pressure change during hydrodynamic injection. (c) Electrocardiogram (upper line) and respiratory curve (lower line) of animals before, during, and after each hydrodynamic injection. (d) Cardiothoracic ratio (CTR) of the treated animal before and after the second injection. Data are representative of measurements from two pigs.

Site and level of reporter gene expression in the liver after sequential hydrodynamic gene delivery to RML and LML

Luciferase gene expression in different parts of the liver undergoing two hydrodynamic gene deliveries sequentially to the RML and LML was analyzed. Results given in Figure 7a show that luciferase gene expression in the targeted lobes was generally higher than that in the nontargeted lobes, but varied among the samples collected from different parts of the liver. The highest luciferase gene expression obtained was >107 RLU/mg and was found in both the RML and LML. Luciferase activity in the nontargeted lobes (RLL, LLL, and CL) was at least three to four orders of magnitude lower. No clear pattern is identifiable within the targeted lobe with regard to the type and location of liver cells transfected. Figure 7b shows the average luciferase gene expression in each of the five lobes. The average luciferase activity in the targeted RML and LML was ~106–107 RLU/mg, 100- to 1,000-fold higher than that seen in each of the nontargeted lobes (103–104 RLU/mg).

Figure 7.

Figure 7

Open in a new tab

Site and level of reporter gene expression in the liver after hydrodynamic gene delivery to right medial lobe (RML) and left medial lobe (LML). Sequential hydrodynamic injection into the RML and LML was performed with pCMV-Luc plasmid DNA (100 µg/ml) with an interval time of 40 minutes. The animal was killed 4 hours after the second injection and liver samples were collected from the sites displayed in a and luciferase activity was determined. (b) Average luciferase gene expression in the targeted lobes (black bars) and nontargeted lobes (gray bars) (mean ± SD). Data represent average from two pigs. CL, caudate lobe; LLL, left lateral lobe; RLL, right lateral lobe.

Assessment of hepatic toxicity of the regional hydrodynamic gene delivery

Careful examination was performed to assess whether the regional and lobe-specific hydrodynamic injection causes tissue damage. At the end of each experiment, organs inside the abdominal cavity including the liver, kidney, stomach, intestine, and spleen were collected and carefully examined for signs of tissue damage. No liquid accumulation or bleeding was seen in abdominal cavity of the animal 4 hours after hydrodynamic injection (length of experiment). A full panel of serum biochemistry tests was performed on each of the treated animals and the results are summarized in Table 1. Among the tests conducted included concentration determination for total proteins, globulin, albumin, alanine aminotransferase, aspartate aminotransferase, γ-glutamic aminotransferase, alkaline phosphatase, lactate dehydrogenase, total bilirubin, blood urea nitrogen, creatinine, glucose, sodium, potassium, and chloride. A slight elevation of serum concentration of aspartate aminotransferase was seen at 0.25, 2, and 4 hours after the hydrodynamic injection and the other serum biochemistry parameters remained unchanged before and after hydrodynamic injection. Microscopic examination of samples collected from various parts of the liver revealed no obvious damage in the targeted or nontargeted lobes (data not shown). These results appear to suggest that the image-guided procedure for regional hydrodynamic gene delivery is safe. However, further studies are needed to examine the long-term effect of the procedure.

Table 1.

Effect of image-guided regional hydrodynamic injection on serum biochemistry

graphic file with name mt2008294t1.jpg

Open in a new tab

Comparative liver histology of different animal species

In our previous study, we showed that the pressure for an optimal hydrodynamic gene delivery to mouse liver was <40 mm Hg,9 which is significantly lower than what we observed in pig liver (75–120 mm Hg, Figure 3g–l). Interestingly, the level of luciferase gene expression obtained in the current study is ~100-fold lower than what was obtained in mouse liver (107 RLU/mg in pigs versus >109 RLU/mg in mice1). These observations prompted us to examine whether there are structural differences between mouse and pig liver. To be inclusive in the study, we have also included liver samples from various animal species that were available to us. Standard histology was performed on all liver samples collected. Supplementary Figure S2 shows the results of Masson's trichrome staining16 (Supplementary Materials and Methods) by which the collagen fibers are stained blue and the background red for liver samples from mouse, rat, rabbit, pig, dog, and human. Evidently, the sample from pig liver is the only one whose lobule structure stained strongly positive. Under the experimental condition employed, no obvious staining of lobular fibers was seen in the tissue samples from other animal species examined. These results confirm a well-known fact that individual lobules of pig liver are surrounded by a thick layer of connective tissue and that pig liver is unique in its interlobular septa.

Discussion

As summarized in recent review articles,3,4 the hydrodynamics-based procedure is being employed in a number of organisms for various purposes such as gene function analysis, gene silencing and target validation (small interfering RNA delivery), analysis of genomes, as well as assessment of plasmid vectors. Because of its simplicity, efficiency, and lack of immunogenicity, the hydrodynamic procedure might be preferable for gene therapy over that of viral and synthetic vectors. The challenge in the past 9 years has been the development of a new technology that would allow hydrodynamic injection with a volume of a few hundred milliliters or more in vivo without causing tissue damage. In this study, we demonstrated for the first time that regional hydrodynamic gene delivery can be achieved with no obvious tissue damage in large animals. We also demonstrated that the approach of combining the image-guided catheter insertion with the newly developed computer-controlled injection device is effective in gene delivery and potentially applicable to humans.

Among many factors that are important for hydrodynamic gene delivery to pig liver, the size of the injection catheter appears critical. For example, when a 5 Fr catheter was used in the experiments described in Figure 3, vascular pressure peaked at 5 mm Hg and disappeared in ~1 second (data not shown). Insertion of an occlusion balloon from the femoral vein into the IVC to block the blood flow increased the peak pressure to ~25 mm Hg initially followed by a lower pressure (10–15 mm Hg) till the end of injection. Occlusion of both the IVC and PV resulted in a sustained peak pressure (25 mm Hg) for the entire injection time. In all of these cases, the level of luciferase gene expression in the targeted lobe was at the 102–103 RLU/mg level. In contrast, when a 10.5 Fr catheter was used, effective accumulation of injected solution (Figure 2), elevated hydrodynamic pressure (Figure 3), and a significant level of reporter gene expression (Figure 5) in the targeted lobes were achieved. Occlusion of the IVC or IVC plus PV helped in increasing hydrodynamic pressure (Figure 3k,l) but not in elevating the level of reporter gene expression (Figure 5). The high flow rate of the large size balloon catheter is most likely the reason for its effectiveness in establishing an elevated hydrodynamic pressure. Because blood circulation across the liver is open to both the PV and IVC, the volume injected with a small catheter would be pushed away from the injection site by the injection pressure and eventually into the IVC or/and PV without accumulation at the injection site. However, when large size catheter was used, the volume injected was sufficient to offset the leakage into the IVC and/or PV and the excess volume is able to accumulate near the injection site of the targeted hepatic vein, elevate vascular pressure, and result in successful gene transfer into liver cells.

Among the four liver lobes examined (RLL, RML, LML, and LLL), the RLL appears different from the other three in responding to image-guided hydrodynamic procedure. Figure 3g shows that the peak pressure in the RLHV was short-lived. Lack of sustained peak pressure for the entire injection period coincides with significantly lower luciferase gene expression in the RLL (104 RLU/mg in RLL versus 107 RLU/mg in other targeted lobes). In addition, an equal level of luciferase gene expression in the RLL and CL (Figure 5a) was seen with the RLL as the target lobe. These results suggest that the leakage of injected DNA solution from the RLHV to the CL and then to the IVC takes place due to the proximity between the RLL and CL and their vascular connection. Such leakage is responsible for a lower pressure in the RLHV and less luciferase gene expression in the RLL. These results also indicate that additional measures such as the IVC blockade may be needed in order to achieve effective hydrodynamic gene delivery to the RLL. However, it is important to point out that obstruction of blood flow in the IVC can cause detrimental effects including a decrease in blood pressure (Figure 3w,x) and respiration rate (Figure 3z3,z4), and a transient increase in heart rate on a rapid removal of the occlusion (Figure 4b).

One interesting but puzzling observation made in this study is the significantly higher hydrodynamic pressure detected in the targeted hepatic vein and relatively low level of reporter gene expression compared to what has been reported in mice.1,2 While the low luciferase gene expression could be due to relative short time allowed for gene expression (4 hours here versus 8 hours in mice), potential differences in strength of cytomegalovirus promoter and differences in stability of gene product between pig and mouse cells in the liver, the high vascular pressure and resistance against hydrodynamic injection seen in pig liver are likely due to differences in hepatic structure between pig and mouse. Abundant collagen fibers surrounding each lobule are clearly demonstrated in Supplementary Figure S2 and suggest that the high pressure seen in pig liver could be due to their ability to restrict the lobules to distend. This restricted distention property of the lobules in pig liver could (i) restrain the expansion of the sinusoids and the circumferentially situated hepatocytes, (ii) limit the hydrodynamic impact on increasing the permeability of liver endothelium and plasma membrane of hepatocytes, and (iii) result in low gene delivery efficiency and reporter gene expression. From an experimental standpoint, the results in Supplementary Figure S2 would question whether pig liver is the most appropriate model for human liver for hydrodynamic gene delivery studies.

Nevertheless, the procedure established in this study represents an important milestone toward the clinical use of hydrodynamic gene delivery. Results shown in this report support the notion that image-guided catheterization through the jugular or femoral vein that has been well established and proven to be safe can be effectively combined with our computer-controlled injection device for effective hydrodynamic gene delivery to the liver. For clinical application, targeted gene delivery to selected liver lobes as shown in Figures 2, 3, and 5 offers a high degree of flexibility in dealing with situations where only part of the liver is available for gene delivery due to disease-caused liver damage. The fact that sequential gene transfer to different lobes can be performed in the same animal (Figure 7) adds significant value to this procedure because high gene product level can be obtained by sequential gene delivery to different lobes. Although further studies are needed, the safety profile demonstrated in Figure 6 and results summarized in Table 1 suggest that animals respond well to the volume (600–800 ml, 20–25-kg pig) and high injection speed (40 ml/s). Future studies need to focus on fine-tuning the hydrodynamic procedure to satisfy the needs of individual applications.

In summary, the procedure of image-guided hydrodynamic gene delivery to the liver has been established using pig liver as a model for large animals. Unlike previous studies using direct infusion of DNA solution to the hepatic vein or IVC and occlusion of blood flow, this procedure allows regional and lobe-specific gene delivery with a 100- to 1,000-fold better efficiency than what has been previously achieved.11,12,13 Relying on a computer-assisted system to control the injection and the image-guided catheter insertion to achieve site specificity, this procedure provides simplicity, effectiveness, and reliability. Further studies to reveal the mechanism by which gene delivery efficiency is determined in different animal species will bring the principle of hydrodynamic gene delivery closer to use in the treatment of human disease.

Materials and Methods

Materials. DNA plasmid (pCMV-Luc) was purified by CsCl-ethidium bromide density gradient ultracentrifugation and kept in Tris–EDTA buffer. Purity of the plasmids was verified by absorbency at 260 and 280 nm and 1% agarose gel electrophoresis. The luciferase assay kit was from Promega (Madison, WI). The introducer SET for image-guided catheter insertion was from COOK (Bloomington, IN) and the 12 Fr sheath and guide wire (ZIP wire) were from Boston Scientific (Natick, MA). Injection balloon catheters (10.5 Fr) were tailor-made by the Clinical Supply (Kakamigahara, Japan). Occlusion balloon catheters (8 Fr) were purchased from Medtronic (Minneapolis, MN). The MIKRO TIP catheter transducer was from Miller (Houston, TX). Contrast medium (OXILAN) was from Guerbet (Bloomington, IN). SIRECUST 961 Siemens for monitoring physiological parameters on animals was from Wittelsbacherplatz (Munich, Germany). Pigs (female, 20–25 kg) were from Wally Whippo (Enon Valley, PA).

Animal catheterization. All experiments performed on pigs were approved by the Institutional Animal Care and Use Committee, University of Pittsburgh, Pittsburgh, Pennsylvania. Under general anesthesia, the animal was placed on the table of fluoroscopy machine YSF-100 from Shimadzu (Nakagyo-ku, Japan). A skin incision was made to expose the jugular or/and femoral vein and an 18-G peripheral catheter from TERUMO (Shibuya-ku, Japan) was inserted. A 0.035-inch hydrophilic guide wire was inserted through the peripheral catheter and replaced later by a short sheath (12 Fr). The injection balloon catheter (10.5 Fr) was inserted through the sheath to the targeted liver lobe followed by insertion of a pressure transducer. To obstruct blood flow, we inserted two occlusion balloons (8 Fr), one from the femoral vein into the IVC and the other from the superior mesenteric vein into the main portal trunk. The latter involved a midline incision to expose the abdominal cavity. Inflation of the balloons was achieved by injecting a small volume of phase contrast medium into the balloon. Obstruction of blood flow was verified by injecting a small volume of phase contrast medium into the vasculature through the injection balloon catheter.

Hydrodynamic injection. Injection of saline containing either pCMV-Luc plasmid DNA (100 µg/ml) or phase contrast medium was made via the HydroJector14 under control of HydroJector600 software. Injection was driven by the pressure provided by a gas (CO2/O2) cylinder with its pressure regulator set at 300 psi. Intravascular pressure and physiological parameters (electrocardiogram, blood pressure, respiration rate, and heart rate) were continuously monitored and recorded during and after the injection. Volume of DNA solution before and after each injection was measured. The injection volume was calculated by subtracting the remaining volume in solution reservoir after the injection from the total volume before the injection.

Visualization of the external appearance of the liver upon hydrodynamic injection. A midline incision was made on an anesthetized animal to expose the abdominal cavity followed by insertion of occlusion balloons and injection balloon catheter according to the procedure described above. The change in liver appearance before, during, and soon after hydrodynamic injection was videotaped using an SD300 digital camera. Eight markers were patched onto the surface of each lobe for estimation of the volume expansion of each lobe. The interior and exterior orientation parameters and target coordinates in the images were calculated based on a bundle adjustment method.17,18 The volume of two tetrahedrons defined by the four target markers was calculated based on Euler's formula that defines the volume of a tetrahedron.

Analysis of luciferase gene expression. Animals were killed and the liver removed 4 hours after hydrodynamic gene delivery. The liver was photographed and 15 liver samples (~200 mg each) covering the entire area of each lobe were collected. Tissue samples were immediately frozen on powdered dry ice and kept at −80°C until use. For the luciferase assay, 1 ml of lysis buffer was added to each sample and it was thawed on ice and homogenized using a tissue Tearor (1 minute, max speed). The tissue homogenate was centrifuged for 10 minutes in a microcentrifuge and the supernatant was collected. Ten microliters of supernatant was taken for luciferase and protein assay according to the previously established procedure.1 The level of reporter gene expression was presented as relative light units per milligram of total protein. Immunohistochemical staining of luciferase is described in Supplementary Materials and Methods.

Assessment of tissue damage. Blood samples were collected from each animal before (time = 0) and at 0.25, 2, and 4 hours after the hydrodynamic injection. Automated concentration determination was performed using an IDEXX VetTest Chemistry Analyzer (Westbrook, ME). For microscopic examinations, liver tissue samples were collected and fixed immediately in 10% formalin. Tissue embedding, sectioning (5 µm), and staining (H/E or trichrome staining16) were done in the pathology lab of the University of Pittsburgh, Department of Pathology. Microscopic examination was performed and photographed under a regular light microscope.

Supplementary MaterialFigure S1. Distribution of luciferase gene expression in the liver. Hydrodynamic gene delivery was performed followed the same procedure as described in Figure 5 legend. Liver samples from the targeted and non-targeted liver lobes were collected, fixed, and sectioned. The sections of hepatic tissue collected from the edge of targeted or and non-targeted lobe were immuno-stained with anti-luciferase antibody. (a) Representative image of tissue sections of one liver sample from the targeted lobe; (b) Representative image for non-targeted lobe. Scale bar represents 100 μm (40x).Figure S2. Comparative liver histology of different animal species. Liver samples from different animal species were fixed and sectioned according to the standard procedure, and Masson's trichrome staining was performed. Arrow indicates connective tissue surrounding the lobular structure of lobule in pig liver. Dark staining in the regions near the blood vessels such as the portal triad are also identifiable in some of the sections. Scale bar represents 100 μm (40x).Materials and Methods.

Supplementary Material

Figure S1.

Distribution of luciferase gene expression in the liver. Hydrodynamic gene delivery was performed followed the same procedure as described in Figure 5 legend. Liver samples from the targeted and non-targeted liver lobes were collected, fixed, and sectioned. The sections of hepatic tissue collected from the edge of targeted or and non-targeted lobe were immuno-stained with anti-luciferase antibody. (a) Representative image of tissue sections of one liver sample from the targeted lobe; (b) Representative image for non-targeted lobe. Scale bar represents 100 μm (40x).

Click here for supplemental data (1.3MB, tiff)
Figure S2.

Comparative liver histology of different animal species. Liver samples from different animal species were fixed and sectioned according to the standard procedure, and Masson's trichrome staining was performed. Arrow indicates connective tissue surrounding the lobular structure of lobule in pig liver. Dark staining in the regions near the blood vessels such as the portal triad are also identifiable in some of the sections. Scale bar represents 100 μm (40x).

Click here for supplemental data (752.5KB, tiff)
Materials and Methods.
Click here for supplemental data (52KB, doc)

Acknowledgments

We thank Joseph E. Knapp for critical reading of the manuscript. This work was supported in part by the National Institute of Health grant RO1EB007357.

REFERENCES

  1. Liu F, Song Y., and , Liu D. Hydrodynamics-based transfection in animals by systemic administration of plasmid DNA. Gene Ther. 1999;6:1258–1266. doi: 10.1038/sj.gt.3300947. [DOI] [PubMed] [Google Scholar]
  2. Zhang G, Budker V., and , Wolff JA. High levels of foreign gene expression in hepatocytes after tail vein injections of naked plasmid DNA. Hum Gene Ther. 1999;10:1735–1737. doi: 10.1089/10430349950017734. [DOI] [PubMed] [Google Scholar]
  3. Suda T., and , Liu D. Hydrodynamic gene delivery: its principles and applications. Mol Ther. 2007;15:2063–2069. doi: 10.1038/sj.mt.6300314. [DOI] [PubMed] [Google Scholar]
  4. Herweiier H., and , Wolff JA. Gene therapy progress and prospects: hydrodynamic gene delivery. Gene Ther. 2007;14:99–107. doi: 10.1038/sj.gt.3302891. [DOI] [PubMed] [Google Scholar]
  5. Condiotti R, Curran MA, Nolan GP, Giladi H, Ketzinel-Gilad M, Gross E, et al. Prolonged liver-specific transgene expression by a non-primate lentiviral vector. Biochem Biophys Res Commun. 2004;320:998–1006. doi: 10.1016/j.bbrc.2004.06.044. [DOI] [PubMed] [Google Scholar]
  6. Brunetti-Pierri N, Palmer DJ, Mane V, Finegold M, Beaudet AL., and , Ng P. Increased hepatic transduction with reduced systemic dissemination and proinflammatory cytokines following hydrodynamic injection of helper-dependent adenoviral vectors. Mol Ther. 2005;12:99–106. doi: 10.1016/j.ymthe.2005.03.001. [DOI] [PubMed] [Google Scholar]
  7. Brunetti-Pierri N, Stapleton GE, Palmer DJ, Zuo Y, Mane VP, Finegold MJ, et al. Pseudo-hydrodynamic delivery of helper-dependent adenoviral vectors into non-human primates for liver-directed gene therapy. Mol Ther. 2007;15:732–740. doi: 10.1038/sj.mt.6300102. [DOI] [PubMed] [Google Scholar]
  8. Su LT, Gopal K, Wang Z, Yin X, Nelson A, Kozyak BW, et al. Uniform scale-independent gene transfer to striated muscle after transvenular extravasation of vector. Circulation. 2005;112:1780–1788. doi: 10.1161/CIRCULATIONAHA.105.534008. [DOI] [PubMed] [Google Scholar]
  9. Zhang G, Gao X, Song YK, Vollmer R, Stolz DB, Gasiorowski JZ, et al. Hydroporation as the mechanism of hydrodynamic delivery. Gene Ther. 2004;11:675–682. doi: 10.1038/sj.gt.3302210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Eastman SJ, Baskin KM, Hodges BL, Chu Q, Gates A, Dreusicke R, et al. Development of catheter-based procedures for transducing the isolated rabbit liver with plasmid DNA. Hum Gene Ther. 2002;13:2065–2077. doi: 10.1089/10430340260395910. [DOI] [PubMed] [Google Scholar]
  11. Yoshino H, Hashizume K., and , Kobayashi E. Naked plasmid DNA transfer to the porcine liver using rapid injection with large volume. Gene Ther. 2006;13:1696–1702. doi: 10.1038/sj.gt.3302833. [DOI] [PubMed] [Google Scholar]
  12. Alino SF, Herrero MJ, Noguera I, Dasi F., and , Sanchez M. Pig liver gene therapy by noninvasive interventionist catheterism. Gene Ther. 2007;14:334–343. doi: 10.1038/sj.gt.3302873. [DOI] [PubMed] [Google Scholar]
  13. Fabre JW, Grehan A, Whitehorne M, Sawyer GJ, Dong X, Salehi S, et al. Hydrodynamic gene delivery to the pig liver via an isolated segment of the inferior vena cava. Gene Ther. 2008;15:452–462. doi: 10.1038/sj.gt.3303079. [DOI] [PubMed] [Google Scholar]
  14. Suda T, Suda K., and , Liu D. Computer-assisted hydrodynamic gene delivery. Mol Ther. 2008;16:1098–1104. doi: 10.1038/mt.2008.66. [DOI] [PubMed] [Google Scholar]
  15. Kirk AD. Crossing the bridge: large animal models in translational transplantation research. Immunol Rev. 2003;196:176–196. doi: 10.1046/j.1600-065x.2003.00081.x. [DOI] [PubMed] [Google Scholar]
  16. Lefkowitch JH. Special stains in diagnostic liver pathology. Semin Diagn Pathol. 2006;23:190–198. doi: 10.1053/j.semdp.2006.11.006. [DOI] [PubMed] [Google Scholar]
  17. Abdel-Aziz YI., and , Karara HM. Proceedings of the ASE/UI Symposium on Close-range Photogrammetry. University of Illinois Press,Urbana; 1972. Direct linear transformation from comparator coordinates into object space coordinates in close-range photogrammetry; pp. 1–18. [Google Scholar]
  18. Maas HG. Concepts of real-time photogrammetry. Hum Mov Sci. 1997;16:189–199. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1.

Distribution of luciferase gene expression in the liver. Hydrodynamic gene delivery was performed followed the same procedure as described in Figure 5 legend. Liver samples from the targeted and non-targeted liver lobes were collected, fixed, and sectioned. The sections of hepatic tissue collected from the edge of targeted or and non-targeted lobe were immuno-stained with anti-luciferase antibody. (a) Representative image of tissue sections of one liver sample from the targeted lobe; (b) Representative image for non-targeted lobe. Scale bar represents 100 μm (40x).

Click here for supplemental data (1.3MB, tiff)
Figure S2.

Comparative liver histology of different animal species. Liver samples from different animal species were fixed and sectioned according to the standard procedure, and Masson's trichrome staining was performed. Arrow indicates connective tissue surrounding the lobular structure of lobule in pig liver. Dark staining in the regions near the blood vessels such as the portal triad are also identifiable in some of the sections. Scale bar represents 100 μm (40x).

Click here for supplemental data (752.5KB, tiff)
Materials and Methods.
Click here for supplemental data (52KB, doc)

Articles from Molecular Therapy: the Journal of the American Society of Gene Therapy are provided here courtesy of The American Society of Gene & Cell Therapy

RESOURCES