Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

DNA constructs used in gene therapy can be administered in vivo in a number of different forms, including as viral particles, as complexes with cationic macromolecules or liposomes, or simply as naked DNA (Anderson, 1998). The most common mode of administration is by direct injection at or near the site where a therapeutic response is desired. This limits the number of target organs available for gene therapy to those readily accessible by direct injection. Systemic administration has proved difficult because of the rapid clearance of DNA from the circulation and the generally poor levels of expression observed following this method. Nevertheless, a number of systems have been described where expression of reporter genes in peripheral tissues can be observed following systemic administration (Wu et al., 1991; Koeberl et al., 1997; Li and Huang, 1997; Griesenbach et al., 1998; Schiedner et al., 1998; Topf et al., 1998). This approach is attractive pharmacologically; it is simple and may provide access to tissues where direct injection of genetic material is not practical. Currently, targeting of exogenous DNA and its persistent expression at therapeutically useful levels in peripheral tissues following intravenous injection is limited (Minchin et al., 1999).

Delivery of foreign DNA to specific tissues in vivo is influenced by many factors, including the structure and size of the DNA, stability against nuclease degradation, and the pharmacokinetics of the delivery system (Ledley and Ledley, 1994; Takakura, 1998). After administration, plasmid DNA that enters the circulation is rapidly removed primarily due to uptake by the nonspecific scavenger receptors located on nonparenchymal cells of the liver (Emlen et al., 1988; Kawabata et al., 1995). DNA complexed to cationic liposomes also can be rapidly cleared from the circulation (Lew et al., 1995). Plasmid DNA has been shown to be extensively degraded in vivo, although exactly where this degradation takes place remains unknown (Emlen and Mannik, 1984; Cosio et al., 1987). This very rapid clearance may be advantageous when exogenous DNA is delivered locally using direct injection, since it will limit the likelihood of expression in nontarget tissues of DNA that inadvertently escapes the site of administration. However, for systems dependent on delivery via the circulation, rapid removal by the liver will limit the extent of gene delivery to peripheral tissues, especially those with relatively small blood flows. This was demonstrated recently in studies of the kinetics of gene delivery to the pancreas, a tissue that receives less than 5% of the cardiac output (Carpenter and Minchin, 1998). Linearized plasmid DNA attached to cholecystokinin as a targeting ligand was rapidly removed from the circulation following intra-arterial administration such that little or no targeting to the pancreas was evident. However, following intraperitoneal administration, approximately 25% of the dose accumulated in the pancreas.

By understanding the kinetics of DNA clearance in the liver, it may be possible to devise interventions that enhance circulation times and, therefore, delivery to extra-hepatic tissues. Although there are a number of reports on the pharmacokinetics of genomic DNA (Emlen and Mannik, 1978, 1984; Cosio et al., 1987), closed circular plasmids DNA (Kawabata et al., 1995), and small single-stranded oligonucleotides (Bijsterbosch et al., 1997), our interest has been in optimizing the delivery of linear plasmid DNA, either alone or complexed to cationic macromolecules, to various target tissues. In the present study, we have investigated the clearance of linearized plasmid DNA in an isolated perfused rat liver model and in vivo. We have also examined the effects of condensing the DNA, addition of liposomes, and inhibition of the hepatic scavenger receptors on plasmid DNA disposition.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals and Materials. Sprague-Dawley rats (approximately 200 g) were purchased from the Animal Resource Center, Murdoch, Western Australia. DNA-modifying enzymes and electrophoresis chemicals were obtained from Promega (Sydney, Australia). Radioisotopes and size exclusion gels were purchased from Amersham Pharmacia Biotech (Sydney, Australia). Lipids were purchased from Sigma Chemical Co. (St. Louis, MO). All other chemicals were of analytical grade.

Preparation of Plasmid DNA. Plasmid pSV-CAT (4750 base pairs; mol. wt. ~3.0 × 10⁶) was amplified in a DH5 strain of Escherichia coli and purified by gel chromatography (Edwards et al., 1996). The DNA was linearized with BglII and end-filled using Klenow fragment in the presence of [³⁵S]dCTP. The radiolabeled DNA was then desalted using a G-25 size exclusion column into phosphate-buffered saline (100 mM NaCl, 50 mM Na₂HPO₄, pH 7.4; PBS). Radiopurity (>95%) was confirmed by 0.8% agarose gel electrophoresis, and the final specific activity was 2000 Ci/mmol.

Preparation of Liposomes. Lipids (10 mg of phosphatidylcholine, 4 mg of phosphatidylethanolamine, and 6 mg of cholesterol) dissolved in 1 ml of chloroform-methanol were evaporated to dryness in a rotary evaporator under reduced pressure. They were then freeze-dried for at least 2 h to remove any traces of organic solvents. PBS (3 ml) was added to the dried lipids and sonicated for 15 s on ice to form small unilamellar vesicles. After centrifugation at 2000g for 10 min, the liposomes in the resulting supernatant were sized using a submicron particle analyzer (Malvern Instruments, Worcestershire, UK). More than 80% of the vesicles ranged from 50- to 100-nm diameter. The liposomes were stored at 4°C until used.

Isolated Perfused Rat Liver. Male Sprague-Dawley rats (approximately 200 g each) were anesthetized with sodium pentobarbital (75 mg/kg i.p.), and the portal vein was cannulated. After severing the aorta, perfusion with Krebs-Henseleit buffer (pH 7.4, 37°C) was commenced at 10 ml/min. The liver was then removed and placed onto a perfusion platform above a collecting funnel. After 5 min of equilibration, ³⁵S-DNA (10 fmol in 100 µl) was administered into the portal vein as a bolus dose. Fractions (5 s) of the effluent were collected for 1 min, and the radioactive content was determined by liquid scintillation spectroscopy.

Iodination of Bovine Serum Albumin and Asialoglycoprotein. Neuraminidase-treated ₁-acid glycoprotein (ASGP) and bovine serum albumin (BSA) were iodinated by the chloramine T method as previously described (Edwards et al., 1996).

In Vivo Studies. Rats were anesthetized with sodium pentobarbital as described above, and a cannula was placed into the left carotid artery. ³⁵S-DNA (500 fmol in 100 µl) was injected as a bolus dose after which blood samples (~600 µl) were collected into EDTA at various times up to 20 min. In some experiments, liposomes (500 µg) and/or polyinosinic acid (100 µg) were premixed with the DNA and coadministered as a single bolus dose. Plasma was recovered by centrifugation, and the ³⁵S-DNA content was determined by liquid scintillation spectroscopy. The amount of degraded DNA in each sample was determined by TCA precipitation as outlined above. At the end of each experiment, tissues were removed, weighed, and homogenized in H₂O. An aliquot of the homogenate was used to determine the total ³⁵S content by liquid scintillation spectroscopy.

Data Analysis. Kinetic parameters for the disposition of ³⁵S-DNA in the isolated perfused rat liver preparation were estimated according to established methods (Lassen and Perl, 1979). Briefly, the single-pass extraction (E) of ³⁵S-DNA was calculated as:

(1)

where (dpm) is the sum of radioactivity in each fraction and D is the dose. The mean transit time () was calculated as:

(2)

where AUC is the area under the DNA efflux-time curve, and C_max is the amount of ³⁵S-DNA in the peak fraction. The volume of distribution (V_d) was calculated as the product of and the flow rate (10 ml/min) divided by the weight of the liver. For all calculations, the volume of the cannula was subtracted.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Disposition of Plasmid DNA in Isolated Perfused Rat Liver. To evaluate the clearance of linear plasmid DNA by the liver, an isolated perfused liver model was used. Figure 1 (upper panel) shows that, following a bolus dose of DNA (10 fmol), less than 40% was recovered in the effluent over 60 s after which little or no further radiolabel was recovered. A similar extraction was observed for ¹²⁵I-ASGP, which binds to the ASGP receptor on parenchymal hepatocytes (Fig. 1, middle panel). This high single-pass clearance of ASGP served as a positive control for the isolated perfused liver preparation (Emlen et al., 1988). By contrast, ¹²⁵I-BSA, which distributes into the vascular space of the liver, showed little or no hepatic clearance (Fig. 1, lower panel). From these data, the extraction ratio (E), mean transit time (), and apparent volume of distribution (V_d) for the plasmid DNA, ASGP, and BSA were calculated and are shown in Table 1. There was no significant difference in the estimates for and V_d for each of the three macromolecules.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Efflux of plasmid DNA, neuraminidase-treated ASGP and BSA from the isolated perfused rat liver following a single bolus dose (10 fmol). On the left axis, the percentage of the dose in each fraction (open bars) is shown; on the right axis, the cumulative efflux of radiolabel as percentage of dose is shown (closed circles). The plots are representative curves for each substrate.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Effect of dose on the single-pass extraction of linear plasmid DNA by the isolated perfused rat liver. Each point is a single independent experiment.

Effects of Various Agents on DNA Extraction by the Liver. We next investigated the effect of a variety of agents on the disposition of linear plasmid DNA in the isolated perfused rat liver (Table 2). Addition of cationic liposomes composed of phosphatidylcholine, phosphatidylethanolamine, and cholesterol (10:4:6, w/w) significantly inhibited hepatic DNA clearance. Extraction decreased to 0.36 ± 0.12 with 50 µg of lipid. Greater amounts of liposomes did not further inhibit extraction. Neither mean transit time nor apparent volume of distribution was significantly altered. The inhibition of plasmid extraction by cationic liposomes was only seen if they were coadministered with the plasmid DNA (Fig. 3, upper panel). Subsequent doses of DNA were not affected.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Effect of cationic liposomes and polyinosinic acid on the extraction of linear plasmid DNA by isolated perfused rat liver. , 10 fmol of DNA alone; , 10 fmol of DNA plus either 50 µg of liposomes (upper panel) or 10 µg of polyinosinic acid (lower panel); , 10 fmol of DNA 5 min after either 50 µg of liposomes (upper panel) or 10 µg of polyinosinic acid (lower panel).

Disposition of Plasmid DNA in Vivo. Following an intra-arterial dose of linearized plasmid DNA, blood samples were collected over 20 min, separated into plasma and red blood cells, and the amount of degradation products in each sample was determined following TCA precipitation. Almost all of the radioactivity (>95%) in the blood was associated with the plasma compartment. The disposition of intact plasmid in plasma was best described by a one-compartment open model (Fig. 4). DNA was rapidly removed from the circulation with a half-life of less than 3 min and an apparent volume of distribution of 670 ml/kg (Table 3). The estimated clearance was 32 ml/min (154 ml/min/kg of body weight), which exceeds hepatic blood flow in the rat [~100 ml/min/kg of body weight (Vermeulen et al., 1983)].

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Disappearance of intact linear plasmid DNA from the plasma of rats in vivo following a single bolus dose of 500 fmol of DNA. , DNA alone; , DNA plus 250 µg of cationic liposomes; , DNA plus 100 µg of polyinosinic acid; , DNA plus 250 µg of cationic liposomes and 100 µg of polyinosinic acid.

Metabolism of Plasmid DNA in Vivo. Figure 5 shows the plasma profile for intact plasmid and degraded DNA following administration of DNA alone or in the presence of polyinosinic acid and/or liposomes. By 20 min after a bolus dose of DNA alone, all the radioactivity in plasma was DNA degradation products. By contrast, only 20% of the total radioactivity was degraded at the same time following DNA coadministered with polyinosinic acid (Fig. 5).

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Kinetics of intact linear plasmid DNA () and metabolites () following a single intra-arterial dose of 500 fmol of DNA either alone (upper left panel), with 500 µg of liposomes (upper right panel), with 100 µg of polyinosinic acid (lower left panel), or with a combination of both (lower right panel).

Effect of Polyinosinic Acid and Liposomes on the Distribution of Total Radioactivity in Vivo. The distribution of total radioactivity in various tissues of rats 20 min following a single dose of linearized plasmid is shown in Table 4. The highest concentration of radiolabel was seen in the liver and the urine. Urinary radiolabel represented DNA degradation products as it was all TCA soluble. If the DNA was administered along with cationic liposomes and polyinosinic acid, there was a marked decrease in the amount of radiolabel in the liver and an increase in most other tissues examined. Radiolabel in the urine was less than 5% of that seen when the DNA was administered alone indicating a decrease in renal excretion.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The rapid clearance of genomic DNA from the circulation was first reported by Tsumita and Iwanaga (1963) following intravenous administration to mice. Clearance involved uptake into the liver and kidneys as well as degradation. Since these early studies, similar findings have been reported for single-stranded DNA (Emlen and Mannik, 1978), closed circular plasmid DNA (Kawabata et al., 1995), and small single-stranded oligonucleotides (Bijsterbosch et al., 1997). In 1988, Emlen et al. (1988) showed that uptake of DNA by the liver involved specific receptors localized to the nonparenchymal cells. These receptors were subsequently shown to be members of the nonspecific scavenger receptor family (Kawabata et al., 1995).

The fate of DNA in vivo has recently been revisited because of the development of gene therapy protocols (Rappaport et al., 1995; Nomura et al., 1997; Qian et al., 1997; Mahato et al., 1998). DNA delivery to target tissues via the circulation is limited by its rapid removal by nontarget tissues, in particular the liver. Therefore, in the present study, we investigated several means to minimize the first-pass clearance of linear plasmid DNA in the liver. In the isolated perfused liver, extraction was similar to that reported for closed circular DNA as was the relationship between dose and uptake (Yoshida et al., 1996). However, the volume of distribution was significantly less (0.29 versus 0.6 ml/g). The closed circular DNA distributed into a volume similar to that for total water (Chou and Rowland, 1997), whereas the volume of distribution for linearized DNA was similar to that for albumin (Fig. 1) or dextran, which distribute into the vascular space of the liver (Mehvar et al., 1991; Yoshida et al., 1996). These results indicate either that the tightly packed closed circular DNA is able to diffuse more widely in the liver or that it undergoes significant reversible binding to hepatic tissue.

Polyinosinic acid, a potent blocker of hepatic nonspecific scavenger receptors, almost completely prevented uptake of DNA by the isolated perfused liver. Polyinosinic acid binds to most classes of the nonspecific scavenger receptor present in the liver (Yoshida et al., 1996; Bijsterbosch et al., 1997; Kamps et al., 1997), on macrophages (Falcone, 1989; Kobzik, 1995) and endothelial cells (Adachi et al., 1997; Nakamura et al., 1998), and in renal tissue (Sawai et al., 1996). Importantly, it was able to prevent uptake whether coadministered with the DNA or administered before it. This is consistent with the action of the polyinosinic acid being mediated by binding with high affinity to the tissue receptors and not by direct interaction with the DNA. A similar finding was observed if a large dose of unlabeled DNA (300 µg) was administered 10 min before the labeled plasmid indicating tight, almost irreversible binding of DNA in the liver (data not shown).

We also found that cationic liposomes significantly decreased the uptake of DNA in the perfused liver. However, this was only observed if the liposomes and DNA were administered concurrently, suggesting that the liposomes exerted their effect by direct interaction with the DNA. Charge neutralization possibly prevented recognition of the DNA by the scavenger receptors, which specifically bind anionic ligands (Yamamoto et al., 1997). However, when the DNA was condensed with polylysine at a ratio that also neutralized the negative charge of the DNA, little effect on extraction was seen, indicating that charge neutralization alone cannot account for the effects of cationic lipids on plasmid DNA extraction.

Linearized DNA administered in vivo was rapidly removed from the circulation partly due to extensive binding in peripheral tissues (large volume of distribution) and partly due to degradation. Degradation products were excreted into the urine within 20 min. The elimination half-life of the DNA was not affected by the addition of cationic liposomes but was increased by polyinosinic acid. Both agents decreased the volume of distribution substantially. Polyinosinic acid also appeared to inhibit the rate of degradation of the plasmid DNA resulting in higher ratios of intact DNA to metabolites in the plasma at 20 min and a marked decrease in the extent of urinary excretion of radioactivity. The metabolism of the DNA did not take place in the plasma, and we were unable to detect degradation in the isolated liver, confirming a similar observation reported by Yoshida et al. (1996) for closed circular DNA. Emlen and Mannik (1978) reported little or no DNA nuclease activity associated with the plasma or blood of mice. One possibility is that the nuclease activity is associated with the endothelial cells of the vasculature. This would account for the rapid metabolism seen in vivo and a clearance that exceeds blood flow to any of the major organs.

The intra-arterial DNA dose used in the present study was 500 fmol. Since approximately 20 to 25% of cardiac output is distributed to the liver, it can be estimated that, at most, an initial dose of 100 to 125 fmol would enter the liver during the first pass. This is within the range of doses that showed a high extraction in the isolated perfused liver studies (Fig. 2).

Combining plasmid DNA with polyinosinic acid and cationic liposomes resulted in a 5-fold increase in the half-life of the DNA, a substantially reduced clearance, and reduced uptake into the liver. This resulted in a higher distribution of radiolabel to nonhepatic tissues. Since DNA metabolism was reduced by this combination, the radiolabel associated with the peripheral tissues may have been intact linear plasmid. However, we were unable to satisfactorily extract the radiolabel to confirm this. Of interest was the apparent increase in uptake of DNA in the pancreas. We have previously reported the synthesis of a linear plasmid complex bound to cholecystokinin for gene targeting to pancreatic acini (Carpenter and Minchin, 1998). We originally found little or no targeting when the complex was administered intra-arterially because of its rapid removal from the circulation. The present study suggests that attenuation of the uptake and degradation pathways involved in DNA clearance may significantly enhance delivery of foreign genes to nonhepatic target tissues in vivo. We are presently utilizing the combination of polyinosinic acid and liposome mixtures to determine whether both targeting and expression in the pancreas can be enhanced by this approach.