New gene delivery agents are needed that function by efficiently mediating targeted nonviral gene delivery in vivo. Some of the most successful delivery agents developed to date, such as PEI and cationic lipids, produce robust gene transfer in vitro, but fail to mediate significant gene expression in vivo.1–8 In contrast, hydrodynamic (HD) dosing and electroporation are proven physical methods that are highly efficient in delivering plasmid DNA to animals and may find utility in humans.9–15 However, due to the invasiveness of HD dosing, there is a need for new gene delivery agents that produce gene expression at levels comparable to HD dosing, but without the requirement of high-volume administration.

Most i.v. dosed nonviral gene delivery systems are cationic polyplexes or lipoplexes, the primary exception being anionic liposomes with encapsulated DNA.16 To increase their compatibility with blood, cationic polyplexes are modified with polyethylene glycol (PEG) to block albumin binding and to “stealth” their rapid endocytosis by cells of the reticuloendothelial system (RES).1, 17–20 Despite these attempts, most PEGylated cationic polyplexes still quickly biodistribute to the liver, and are taken-up primarily by Kupffer cells.21, 22 Few, if any, of these experimental i.v. dosed gene delivery systems have produced appreciable gene expression in the liver,23, 24 especially when compared to the level of luciferase expression from an equivalent dose of DNA delivered hydrodynamically.10, 25

To develop nonviral gene delivery agents that function with greater efficiency in vivo following i.v. dosing, it is necessary to control the size, charge and metabolic stability of polyplexes.26 The cationic surface charge of most polyplexes results from residual unpaired amines on the cationic polymer used to condense DNA. Anionic polyplexes may be more blood compatible, but would require a new mode of binding polymers to DNA, not based on electrostatic interaction.

The use of polyintercalation has been shown to increase the binding affinity of small polymers and peptides to DNA.27–33 Previous studies from our group have established that polyacridine peptides possessing four Lys(Acr) residues (ε amine of Lys is modified with acridine), spaced by either a Lys or Arg, produce high affinity binding for plasmid DNA.34 Modification of polyacridine peptides with either a fusogenic peptide (melittin) or an N-glycan (high-mannose) produced gene transfer agents that formed cationic polyplexes resulting in high level selective gene transfer in vitro.34, 35 PEGylated polyacridine peptides have been shown to bind to plasmid DNA and produce unique, metabolically stable, anionic open-polyplexes that are capable of mediating gene transfer in vivo when delivered by i.m.-electroporation.36

In the present study we prepared a panel of PEGylated polyacridine peptides to establish their utility in mediating gene expression following i.v. dosing. The results establish a clear structure-activity relationship by which protection of the DNA in blood and liver by a PEGylated polyacridine peptide results in the ability to stimulate high level gene expression in liver with a delayed hydrodynamic dose of saline. The results also establish the ability to extend the circulatory half-life of a significant percentage of i.v. dosed DNA, which is an important prerequisite of achieving targeted gene delivery to organs and tissues. The synthetic adaptability of polyacridine peptides, along with the ability to prepare either anionic-open or cationic-closed-polyplexes that mediate gene transfer in vivo, demonstrate the unique importance of poly-intercalative binding to achieve gene delivery.

Unsubstituted Wang resin, 9-hydroxybenzotriazole, Fmoc-protected amino acids, O-(7-Azabenzo-triazol-1-yl)-N,N,N′,N′-tetramethylurionium hexafluorophosphate (HATU), Fmoc-Lysine-OH, and N-Methyl-2-pyrrolidinone (NMP) were obtained from Advanced ChemTech (Lexington, KY). N,N-Dimethylformamide (DMF), trifluoroacetic acid (TFA), and acetonitrile were purchased from Fisher Scientific (Pittsburgh, PA). Diisopropylethylamine, piperidine, acetic anhydride, Tris(2-carboxyethyl)-phosphine hydrochloride (TCEP), 9-chloroacridine and thiazole orange were obtained from Sigma Chemical Co. (St. Louis, MO). Agarose was obtained from Gibco-BRL. mPEG-maleimide and mPEG-OPSS (5,000 Da) were purchased from Laysen Bio (Avab, AL). D-Luciferin and luciferase from Photinus pyralis were obtained from Roche Applied Science (Indianapolis, IN). pGL3 control vector, a 5.3 kb luciferase plasmid containing a SV40 promoter and enhancer, was obtained from Promega (Madison, WI). pGL3 was amplified in a DH5α strain of Escherichia coli and purified according to manufacturer’s instructions.

A panel of PEGylated polyacridine peptides were prepared and tested for their ability to bind and transport DNA in vivo. The peptides were designed to test the influence of DNA binding affinity on gene transfer efficiency. To examine the influence of spacing amino acid, four peptides of the general structure (Acr-X)₄ were prepared, where X is either Lys, Arg, Leu or Glu (Table 1). As discussed in more detail below, a Lys spacing amino acid proved most efficient in gene transfer, which warranted an expansion of (Acr-Lys)_n to include repeats of 2 and 6 (Table 1). Each polyacridine peptide also possessed a C-terminal Cys residue that was modified with PEG. During the course of synthesis, the synthetic yields of polyacridine peptides were improved by using HATU to activate amino acids during coupling. Improved yields were also the result of applying triple coupling of the spacing amino acids, omitting N-capping to avoid a side reaction with acridine, and using ethanedithiol as a scavenger during workup. Under these conditions, (Acr-Lys)_n with repeats ranging from 2 to 6 were prepared in 20–37% isolated yield (Table 1). Each polyacridine peptide produced an ESI-MS consistent with the calculated mass (Table 1).

Conjugation of each polyacridine peptide with PEG was accomplished by reaction of the Cys residues with either PEG-maleimide or PEG-OPSS, resulting in PEGylated polyacridine peptides with a reducible disulfide or a non-reducible maleimide linkage (Fig. 1). The reaction was monitored by RP-HPLC in which the polyacridine peptide (Fig. 2A) was mostly consumed, resulting in formation of a later eluting PEG-peptide (Fig. 2B). Preparative RP-HPLC purification produced PEG peptides free of unreacted polyacridine peptide and PEG as established by analytical RP-HPLC and MALDI-TOF analysis (Fig. 2C). Each PEGylated polyacridine peptide produced a MALDI-TOF MS with multiple peaks due to the polydispersity of PEG_{5000 Da}, with average mass closely correlated to the calculated mass for the PEGylated peptide (Table 1).

The relative binding affinity of PEGylated polyacridine peptides for pGL3 was compared by determining the concentration of peptide that displaces a thiazole orange intercalator dye, resulting in decreased fluorescence. Comparison of PEGylated (Acr-X)₄ (X is either Arg, Lys, Leu or Glu) established the importance of cationic amino acids to increase binding affinity. Both the Lys and Arg analogue demonstrated high affinity by completely displacing thiazole orange at 0.4 nmol of peptide per μg of DNA. Conversely, the Leu analogue demonstrated weaker binding, resulting in full displacement at 1 nmol, and the Glu analogue was determined to have the lowest affinity by only achieving 40% displacement at 1 nmol (Fig. 3A).

A similar comparison of the binding affinities for PEGylated (Acr-Lys)_{2, 4 and 6} established a relationship of increasing affinity with increasing the number of Acr, such that PEGylated (Acr-Lys)₆ completely displaced thiazole orange at 0.2 nmols of peptide per μg of DNA (Fig. 3B). The importance of the spacing amino acid was further established by PEGylated (Acr-Lys)₂ which possessed greater DNA binding affinity than PEGylated (Acr-Glu)₄ (Fig. 3A and B).

The size and charge of DNA polyplexes are often a function of the stoichiometry of peptide bound to DNA. To examine this relationship for PEGylated polyacridine polyplexes, QELS particle size and zeta potential were measured as a function of peptide to DNA ratio. One unusual property of PEGylated polyacridine polyplexes was their QELS mean diameter remained constant throughout the titration (Fig. 4A–F). Each of the polyplexes had an apparent mean diameter of approximately 100–200 nm, with the exception of PEGylated (Acr-Leu)₄ and PEGylated (Acr-Glu)₄ which produced polyplexes of an 200–400 nm mean diameter (Fig. 4B and C).

A second unusual property of PEGylated polyacridine polyplexes was the observed zeta potential at each peptide stoichiometry (Fig. 4). Titration with PEGylated (Acr-Arg)₄ resulted in an increase in zeta potential from −15 mV to a maximum of −2 mV at 0.4 nmols of peptide per μg of DNA or higher (Fig 4A). Similar zeta potential titration curves were determined for PEGylated (Acr-Leu)₄, (Acr-Lys)₂ and (Acr-Lys)₄, each of which resulted in polyplexes with a negative zeta potential when fully titrated (Fig. 4B, D and E). In contrast, PEGylated (Acr-Glu)₄ produced no change in zeta potential during the titration (Fig. 4C), and titration with PEGylated (Acr-Lys)₆ produced zeta potentials that increased from −5 mV to a maximum of +9 mV during the titration (Fig. 4F). These data suggested that, unlike polylysine or PEI polyplexes that condense DNA into smaller polyplexes of 50–100 nm with charge of +25 mV,41, 42 polyacridine peptides bind without dramatically changing DNA size and can be, in certain cases, titrated to completion while maintaining a negative zeta potential.

The shape and relative charge of PEGylated polyacridine polyplexes was further examined by atomic force microscopy (AFM) (Fig. 5). Polyplexes prepared at 0.8 nmol per μg of DNA using PEGylated (Acr-Arg)₄, (Acr-Lys)₄, (Acr-Leu)₄ and (Acr-Glu)₄, each produced anionic open-polyplexes that bound to cationically charged mica (Fig. 5A-D). The relative size and shape appeared as open-polyplexes of approximately 0.2–0.3μm, relative to naked plasmid DNA which also bound to cationic mica to produce a more open coiled structure of approximately 0.5 μm diameter (Fig. 5E). Examination of AFM images of PEGylated (Acr-Lys)₆ polyplexes revealed anionic open-polyplexes that bound to cationic mica at 0.2 nmol per μg of DNA (Fig. 5F). Alternatively, 0.8 nmol per μg of DNA produced cationic closed-polyplexes of slightly smaller size that bound to anionic mica (Fig. 5G). Cationic closed-polyplexes were only observed on anionic mica, whereas an attempt to achieve binding to cationic mica resulted in images devoid of visible polyplexes (Fig. 5H). These results illustrate the unique shape of anionic open-polyplexes as well as the ability to manipulate the charge of PEGylated (Acr-Lys)₆ polyplexes to be either cationic or anionic and change shape from open to closed-polyplexes depending on peptide stoichiometry. Anionic PEGylated open-polyplexes should minimize binding to serum proteins and extend the pharmacokinetic half-life of plasmid DNA, provided that the DNA remains protected from premature metabolism.

To investigate the metabolic stability of DNA polyplexes, PEGylated (Acr-Lys)_{2, 4, and 6} polyplexes were digested with DNase, followed by phenol-chloroform extraction to remove peptide and gel electrophoresis to determine the status of the plasmid DNA (Fig. 6). Unique band shifts occurred for PEGylated (Acr-Lys)_{2, 4, and 6} polyplexes prepared at either 0.2 or 0.8 nmols of peptide per μg of plasmid DNA (Fig. 6, lane 2 and 3 of panels A, B, C). However, phenol-chloroform extraction removed the PEG-peptide allowing recovery of the DNA which migrated coincident with control (Fig. 6, lane 4 of panels A, B and C). When challenged with a DNase digestion, PEGylated (Acr-Lys)_{2 or 4} polyplexes prepared at 0.2 nmol per μg of DNA failed to protect DNA from metabolism (Fig. 6, panel A and B, lane 5 and 6). In contrast, polyplexes formed at 0.2 nmols PEGylated (Acr-Lys)₆, protected the DNA from DNase resulting in the recovery of bands following extraction (Fig. 6, panel C, lanes 5 and 6). At a higher stoichiometry of 0.8 nmols, PEGylated (Acr-Lys)_{2, 4 and 6} each protected DNA from metabolism (Fig. 6, panel A–C, lanes 7 and 8). These results establish a clear correlation between polyacridine DNA binding affinity (Fig. 3) and metabolic stability of polyplexes. Most importantly, PEGylated (Acr-Lys)₆ is able to form DNase stable anionic open-polyplexes when prepared at 0.2, as well as cationic closed-polyplexes prepared at 0.8 nmol per μg of DNA.

PEGylated polyacridine peptides were also subjected to trypsin digestion to determine if they could be digested by a common serine protease, suggesting perhaps they could also be more easily cleared from cells. The results established that even as little as 100 mU of trypsin catalyzed the formation of a dipeptide of Lys-Lys(Acr) with mass of 451 g/mol, to a reaction completion within 1 hr. Thereby, the high-affinity DNA binding of PEGylated polyacridine peptides is greatly reduced upon proteolytic digestion.

To determine if polyacridine peptides would mediate gene transfer, we chose to examine the ability of PEGylated DNA polyplexes to produce luciferase expression in the liver of mice, 24 hrs following a 1 μg dose of pGL3. Hydrodynamic dosing of 1 μg of pGL3 produced 10⁸ photon/sec/cm²/sr determined using a calibrated bioluminescence assay (Fig. 7A, HD DNA). When pGL3 (1 μg in 50 μl) was dosed via the tail vein without hydrodynamic delivery, there was no detectable luciferase expression (not shown). In an attempt to stimulate gene expression, PEGylated polyacridine peptide polyplexes were administered via the tail vein (1 μg of pGL3 in 50 μl) and after 30 min a hydrodynamic dose (9 wt/v %, or approximately 2 ml) of saline was used to stimulate liver uptake and gene expression (Fig. 7A). In the subsequent studies described below, stimulated expression produced 10⁷–10⁹ photon/sec/cm²/sr, the magnitude being dependent on the structure of the PEGylated polyacridine peptide, the time delay between primary dose and stimulation dose, the charge ratio at which the polyplex was prepared, and the amount of DNA dosed.

At a fixed dose of 1 μg of pGL3 and a time delay of 30 min between primary and stimulatory dose, polyplexes prepared at 0.5 nmol of PEGylated (Acr-Arg)₄, in which the PEG was attached by a maleimide (Mal) or a reducible disulfide (SS) linkage, were administered to mice and analyzed by BLI after 24 hrs. Approximately 5-fold higher luciferase expression was observed for Mal versus SS, indicating that a reducible linkage offered no advantage for stimulated expression (Fig. 7A). Comparison of the Mal peptide prepared as either an acetate or TFA salt form (ion paired with Arg) resulted in an approximately 10-fold loss of gene transfer for the TFA form (Fig. 7A). These results justified the use of Mal peptides in the acetate salt form for further comparisons.

Several controls were applied to further establish that PEGylated polyacridine peptides were necessary to achieve stimulated expression. The administration of a 1 μg dose of pGL3 followed by a stimulatory dose of saline after 30 min resulted in no detectable luciferase expression at 24 hrs (Fig. 7A, IV DNA). Likewise, there was no detectable luciferase expression with a 1 μg dose of pGL3 prepared as a PEGylated-Cys-Trp-Lys₁₈ polyplex (Fig. 7A, cont 1),43 or PEGylated glycoprotein polyplex (Fig. 7A, cont 2).44 These controls confirmed that naked DNA was not responsible for the observed expression and also established that not all gene formulations were capable of mediating stimulated expression.

The stimulated expression mediated by PEGylated (Acr-Arg)₄ polyplexes (1 μg) was examined to determine the influence of the delay time between primary and stimulatory dose. Varying the delay from 5–120 min established a 10-fold loss in luciferase expression between 5 and 30 min, followed by a complete loss of stimulated expression at 60 min and longer (Fig. 7B). Interestingly, the magnitude of stimulated gene expression at a delay time of 5 min was approximately 5-fold greater than direct hydrodynamic dosing of 1 μg of pGL3 (Fig. 7A and B). Controls that either decreased the volume of the hydrodynamic stimulatory dose or the speed at which it was administered both led to no gene expression.

To determine how the spacing amino acid influence stimulated gene expression, PEGylated peptides of general structure (Acr-X)₄, possessing X as either Lys, Arg, Leu or Glu were used to prepare polyplexes that were dosed in mice and stimulated to express luciferase. While the DNA dose was fixed at 1 μg, and the time delay was 30 min, the stoichiometry of PEGylated peptide to DNA was adjusted based on the results of zeta potential titration (Fig. 4) to favor complete polyplex formation. The results established that only PEGylated peptides possessing an Arg or Lys spacing amino acid produced appreciable levels of luciferase expression (Fig. 8A), whereas peptides with Glu and Leu spacers were essentially inactive.

The stoichiometry of both Lys and Arg spaced PEGylated polyacridine peptide were varied from 0.2–1 nmol of peptide per μg of DNA, then dosed via tail vein and stimulated after 30 min. A 1 μg open polyplex prepared with 0.6 nmol of PEGylated (Acr-Arg)₄ or 0.8 nmol of PEGylated (Acr-Lys)₄ mediated maximal luciferase expression when assayed by BLI at 24 hrs (Fig. 8B). Both the Lys and Arg spaced PEGylated peptide mediated approximately 100-fold lower gene expression at 0.2 nmols of peptide compared to the stoichiometry of maximal expression (Fig. 8B). Likewise, at stoichiometries greater than maximal expression, a decline in expression was observed. These results suggested that a Lys spacer may have a slight advantage over Arg, which then prompted an investigation into varying the number of (Acr-Lys) repeats.

Comparison of the stimulated gene expression (30 min) mediated by PEGylated (Acr-Lys)_{2, 4 and 6} polyplexes (1 μg pGL3) prepared at an optimized stoichiometry of 1, 0.8 and 0.2 nmols of peptide respectfully (based on zeta potential, Fig. 4), established that PEGylated (Acr-Lys)₂ was completely inactive. In contrast, PEGylated (Acr-Lys)₆ at 0.2 nmols matched the activity of PEGylated (Acr-Lys)₄ at 0.8 nmols (Fig. 8C). These results correlate with PEGylated (Acr-Lys)₆ having a higher binding affinity for DNA (Fig. 3), with a greater ability to protect DNA from metabolism (Fig. 4), and with the ability to mediate stimulated gene expression even at a low stoichiometry of 0.2 nmols per μg of DNA (Fig. 8).

When PEGylated (Acr-Lys)₆ polyplexes were administered to mice to determine how changing stoichiometry influenced gene expression, the results established that a 1 μg dose of pGL3 polyplex, stimulated at 30 min, produced equivalent gene expression across the range of 0.2–0.8 nmols of peptide (Fig. 9A). These results are in sharp contrast to those determined for PEGylated (Acr-Lys)₄ (Fig. 8B), in which the gene expression was highly dependent on the stoichiometry of peptide to DNA. Furthermore, since PEGylated (Acr-Lys)₆ was the only peptide to produce a zeta potential titration that started anionic (−5 mV) at 0.2 nmol and titrated to cationic (+9 mV) at 0.8 nmols (Fig. 4F), it appeared that the stimulated gene expression mediated by PEGylated (Acr-Lys)₆ is fully functional as both an anionic open polyplex (Fig. 5F), and as an cationic closed polyplex (Fig 5G).

To determine if the unique DNA binding properties of PEGylated (Acr-Lys)₆ would also extend the delay time allowed between primary and stimulatory dose, polyplexes (1 μg) were prepared with 0.2 nmols of peptide and administered with a time delay varying from 0 to 120 min. Delaying the stimulation from 5 to 60 min resulted in only a slight 2–3 fold decrease in expression, whereas with a 120 min delay the expression decreased 100-fold (Fig. 9B). Using a zero delay time resulted in a nearly 5-fold increase in gene expression relative to dose equivalent direct HD administration of pGL3 (Fig. 9B). These results demonstrate that PEGylated (Acr-Lys)₆ mediates significant high-level gene expression, even with a 60 min delay time, compared to PEGylated (Acr-Arg)₄ which failed to mediate detectable expression following a 60 min delay in stimulation (Fig. 7B).

To establish the dose-equivalency of direct HD administration of DNA, relative to the stimulated expression of PEGylated polyacridine polyplexes, a dose-response experiment was performed (Fig. 9C). The BLI detected expression at 24 hrs post-DNA delivery was determined for direct HD delivered pGL3, dosed at 0.1, 1 and 5 μg of DNA (Fig. 9C). This was compared with the 24 hr luciferase expression from identical DNA doses of PEGylated (Acr-Lys)₆ polyplexes prepared with 0.2 nmol of peptide and administered with a 5 min delay in stimulation. A nearly linear dose-response curve was identified for direct HD dosing of pGL3, whereas PEGylated (Acr-Lys)₆ polyplexes showed an increase from 0.1–1 μg, followed by a plateau at 5 μg dose (Fig. 9C). At a dose of 1 μg of DNA, PEGylated (Acr-Lys)₆ mediated approximately 5-fold greater gene expression (Fig. 9C).

To gain further mechanistic insight into the reason that PEGylated polyacridine polyplexes mediate stimulated gene expression, pharmacokinetic and biodistribution studies were performed. Following i.v. dosing of ¹²⁵I-DNA or PEGylated (Acr-Lys)₂ polyplexes resulted in a rapid 10-fold loss of radioactivity from the blood within 20–30 min, followed by an apparent long half-life (Fig. 10A). Isolation of ¹²⁵I-DNA from blood time points, followed by gel electrophoresis and autoradiography demonstrated that plasmid DNA was rapidly degraded to fragments within 1 min (Fig. 10C). Likewise, electrophoretic analysis of blood time points following dosing of PEGylated (Acr-Lys)₂ ¹²⁵I-DNA polyplexes demonstrated complete loss of plasmid DNA within 20 min (Fig. 10D), suggesting that the modest binding from PEGylated (Acr-Lys)₂ leads to slightly delayed metabolism.

The pharmacokinetic profile for PEGylated (Acr-Lys)_{4 and 6} ¹²⁵I-DNA polyplexes were both very similar and distinct from those of ¹²⁵I-DNA and PEGylated (Acr-Lys)₂ ¹²⁵I-DNA polyplexes (Fig. 10A). The radioactivity in blood only decreased 3–4 fold in the first 20 min and proceeded with an apparent long half-life (Fig. 10A). Recovery of ¹²⁵I-DNA from PEGylated (Acr-Lys)_{4 and 6} polyplexes resulted in an agarose gel autoradiographic analysis that established the blood stability of plasmid DNA up to 30 min for PEGylated (Acr-Lys)₄ (Fig. 10E), and up to 182 min for PEGylated (Acr-Lys)₆ (Fig. 10F). Non-linear least squares analysis of the radioactivity in blood over time was used to calculate an apparent α-half-life of 2–3 min and a β-half-life of 65 min for PEGylated (Acr-Lys)₄ and 120 min for PEGylated (Acr-Lys)₆ (Table 2). For the purpose of comparison, complete metabolic stability of plasmid DNA was assumed in the pharmacokinetic calculation, even though the gel electrophoretic and autoradiographic analysis for PEGylated (Acr-Lys)₂ and (Acr-Lys)₄ establish the differential loss of intact DNA over time. The apparent clearance rate decrease and the mean residence time increased as the β-half-live increased, whereas the apparent volume of distribution remained nearly constant. This resulted in an increase in the AUC as the stability of the polyplex increased (Table 2).

The unique long pharmacokinetic β-half-life for PEGylated (Acr-Lys)_{4 and 6} polyplexes and the nearly coincident volume of distribution suggested there would be similar biodistribution to the organs. The experimental result established that the liver was the major site of biodistribution for PEGylated (Acr-Lys)_{2, 4 and 6} ¹²⁵I-DNA polyplexes (Table 3). In each of these experiments, it was not possible to extract ¹²⁵I-DNA from liver or other tissues to determine its metabolic status, as was performed for DNA in blood.

The biodistribution of ¹²⁵I-DNA and PEGylated (Acr-Lys)₂ ¹²⁵I-DNA polyplexes both reached a maximum of approximately 65% in the liver at 5 min, followed by a decrease to less than 10% by 120 min (Fig. 10B). In contrast, when dosing PEGylated (Acr-Lys)_{4 and 6} ¹²⁵I-DNA polyplexes the percent of radioactive dose in liver was approximately 55% at 5 min, which accumulated to over 60% at 15 min, followed by a decline to 30–40% at 120 min (Fig. 10B). By comparison, a maximum of 1–2% of the total radioactive dose distributed to lungs, heart, kidneys and intestine at all time points (Table 3). The exception was the spleen which showed a steady accumulation over time reaching 18% at 2 hrs when dosing PEGylated (Acr-Lys)₆ polyplexes. In addition to the liver and spleen, the blood accounted for the majority of the radioactive dose, which ranged from 65–40% in the circulation during 2 hrs following dosing of PEGylated (Acr-Lys)₆ polyplexes.

PEGylated polyacridine peptides were designed to bind to plasmid DNA for the purpose of making it more blood compatible as a first step toward ultimately adding a targeting ligand and sub-cellular targeting peptides, needed to complete the delivery system and achieve significant in vivo gene expression following i.v. dosing, without the requirement of an additional stimulation. To accomplish this goal, an optimal polyacridine peptide would need to bind to DNA with sufficient affinity to protect from DNase metabolism in the circulation, while being able to release the DNA inside the cell to gain access to the nucleus.

It is likely there are many unique sequences of polyacridine peptides that accomplish this goal. The aim of the structure-activity study described was to determine a relationship between the number of Acr, the binding affinity for DNA, the polyplex pharmacokinetic half-life, and the magnitude of stimulated expression in mice. Initially, a nine amino acid peptide allowed the incorporation of four Acr spaced by a hydrophobic (Leu), anionic (Glu), or cationic (Lys and Arg) residue, along with a C-terminal Cys for modification. Prior studies from our group established that spacing amino acids that lacked a bulky side-chain or protecting group (Gly or Ala) resulted in low peptide yields. During the course of this study it was necessary to further optimize peptide yield to prepare (Acr-Lys)₆-Cys. While it is possible to prepare even longer polyacridine peptides of this design, we found the yields began to diminish. Each of the polyacridine peptides was coupled to PEG_5000Da, resulting in PEGylated polyacridine peptide that produced a single symmetrical peak on RP-HPLC and a MALDI-TOF MS that verified the structure.

The primary structural features that influence the level of gene expression mediated by PEGylated polyacridine peptides was the presence of at least four Acr residues combined with cationic (Lys or Arg) spacing amino acid. PEGylated polyacridine peptides possessing four Acr residues spaced by either Glu or Leu demonstrated weak binding to DNA (Fig. 3A), resulting in no stimulated gene expression (Fig. 8A), presumably due to the inability to protect DNA from metabolism in vivo. Conversely, PEGylated (Acr-Lys)₄ or (Acr-Arg)₄ possessed higher affinity for binding with DNA (Fig. 3A), which correlated with their ability to produce stimulated expression (Fig. 8A), due to their ability to stabilize DNA from premature metabolism in vivo (Fig. 10E). We previously reported that PEGylated Cys-Trp-Lys₁₈ formed polyplexes that readily dissociate in the blood, resulting in rapid DNA metabolism.22 It was therefore anticipated that stimulation of i.v. dosed PEGylated Cys-Trp-Lys₁₈ polyplexes would not produce gene expression (Fig. 7A, cont 1). Thus, a combination of polyintercalation and ionic binding to plasmid DNA results in improved affinity and stable polyplexes, where either functionality alone is insufficient.

One of the more unique aspects of PEGylated polyacridine polyplexes is their open polyplex structures that resemble naked DNA (Fig. 5). PEGylated (Acr-Lys)₄ and (Acr-Arg)₄ produced anionic open-polyplexes across the titration range of 0.2–0.8 nmols of peptide (Fig. 4A and E). When i.v. dosed and allowed to circulate for 30 min, the level of stimulated expression was dependent upon the stoichiometry of peptide to DNA, resulting in a parabolic peptide dose-response curve (Fig. 8B). These results suggested that the DNA stability afforded by PEGylated (Acr-Lys)₄ and (Acr-Arg)₄, was dependent on peptide stoichiometry, with higher stoichiometries resulting in charge neutral polyplexes (Fig. 4A) that stabilize DNA from metabolism (Fig. 6B) and demonstrate diminished gene transfer efficiency (Fig. 8B).

PEGylated (Acr-Lys)₆ binds to DNA with higher affinity compared to all other peptides studied (Fig. 3). It also possessed an unusual zeta potential titration curve that produced anionic open-polyplexes at 0.2 nmols of peptide, which convert to cationic closed-polyplexes at 0.8 nmols and higher (Fig. 4F and Fig 5F & G). The DNase stability afforded to PEGylated (Acr-Lys)₆ polyplexes at both 0.2 and 0.8 nmols of peptide (Fig. 6C) resulted in equivalent, high-level stimulated gene expression at each stoichiometry from 0.2–0.8 nmols (Fig. 9A). The greater affinity also led to greater stimulated expression for longer delay times compared to PEGylated (Acr-Arg)₄ polyplexes, such that following the administration of 1 μg of DNA, a 1 hour delayed stimulation produced high-level gene expression, and at 2 hour delay measurable expression was still detected (Fig. 9B). These results established that PEGylated (Acr-Lys)₆ polyplexes mediated equal gene expression as either closed or open-polyplex structures, and suggest that enhanced binding affinity translates to higher expression at longer delay times due to postponing metabolism.

Pharmacokinetic and biodistribution studies were used to gain insight into the underlying mechanism of how delayed hydrodynamic stimulation caused PEGylated (Acr-Lys)₆ polyplexes to mediate gene expression in the liver. The most striking result was the apparent long pharmacokinetic half-life of PEGylated (Acr-Lys)₄ and PEGylated (Acr-Lys)₆ polyplexes, compared to the rapid loss of ¹²⁵I-DNA and PEGylated (Acr-Lys)₂ polyplexes (Fig. 10A). Electrophoretic analysis of the ¹²⁵I-DNA recovered from blood clearly established that PEGylated (Acr-Lys)₆ stabilized DNA in the circulation for 2 hrs (Fig. 10F), whereas PEGylated (Acr-Lys)₄ stabilized open circular DNA bands for at least 30 min (Fig. 10E), and PEGylated (Acr-Lys)₂ for no more than 20 min (Fig. 10D). Most importantly, approximately 40% of the radioactive dose remained in the blood after 2 hrs when dosing PEGylated (Acr-Lys)₆ ¹²⁵I-DNA polyplexes (Table 3), and the pharmacokinetic profile appears to be the same whether dosing anionic-open or cationic-closed polyplexes (Fig. 10A).

These results cannot be directly compared with several prior studies that reported long circulating DNA or siRNA formulations by analyzing the pharmacokinetics of polyplexes by incorporating a fluorophore or radiolabel into the carrier component.16, 45–47 Three prior studies are more closely related in their direct analysis of pharmacokinetics using radiolabeled DNA48, 49 or by quantifying DNA in blood and tissues by PCR.50 These reports establish the ability to extend the half-life of DNA in the circulation, but stop short of establishing the metabolic status of recovered DNA. This is the first study to demonstrate stabilized plasmid DNA circulating in mice by gel electrophoresis and autoradiography. The metabolic stability of DNA is directly related to the mode and affinity of binding achieved by PEGylated (Acr-Lys)₆. A similar PEGylated polylysine peptide (PEG-Cys-Trp-Lys₁₈) also binds with high affinity to DNA through ionic interaction and forms cationic-closed polyplexes.43 However, despite being a longer peptide than PEGylated (Acr-Lys)₆, PEG-Cys-Trp-Lys₁₈ polyplexes rapidly dissociate following i.v. dosing leading to the complete metabolism of plasmid DNA in the circulation within 3 min.22 Thereby, it is the combination of polyintercalation and cationic binding that appears to be essential to achieve DNA stability during circulation, whereas PEGylated peptides that bind DNA by either mode individually are insufficient. This unique feature suggests that PEGylated (Acr-Lys)₆ polyplexes may find application in targeting DNA to tissues outside the liver.

Analysis of the tissue distribution of ¹²⁵I-DNA polyplexes over time established the liver as the major site of distribution accounting for approximately 54–69% of the dose within the first 5 min, with only minor (<1%) distribution to other organs. PEGylated (Acr-Lys)₄ and PEGylated (Acr-Lys)₆ polyplexes produced a distinct liver distribution and metabolism profile, with maximal accumulation of 60% at 20 min followed by a decrease to 30% over two hrs (Fig. 10A). The liver biodistribution profile of ¹²⁵I-DNA and PEGylated (Acr-Lys)₂ ¹²⁵I-DNA polyplexes were coincident with maximal accumulation of 65% at 5 min, followed by a decrease to 10% over 2 hrs. The stability of plasmid DNA in the blood afforded by PEGylated (Acr-Lys)₆ appears to extrapolate to similar stability in the liver. Taken together, the pharmacokinetic and biodistribution data support a hypothesis in which the DNA that distributes to the liver is primarily responsible for the stimulated expression. However, liver distribution is not sufficient, the DNA must also be sufficiently metabolically stabilized to mediate significant stimulated expression with delay times up to an hr. The hypothesis that liver associated DNA, and not the DNA in blood, is responsible for stimulated expression is deduced by comparison of the level of stimulated expression from PEGylated (Acr-Lys)₆ polyplexes (Fig. 9B), which decrease by nearly 100-fold between 60 and 120 min while the level of intact DNA in the blood remains constant over that time period (Fig. 10F). It is unlikely that PEG (Acr-Lys)₆ spontaneously dissociates from DNA upon entering the cell or nucleus. Perhaps, PEG (Acr-Lys)₆ possess affinity for chromosomal DNA, which acts as a sink, providing a mechanism for the release of plasmid DNA in the nucleus.

In conclusion, this is the first report of hydrodynamically stimulated gene expression from a nonviral delivery system that mirrors the level of expression produced by the same dose administered by direct hydrodynamic dosing. The unique attributes of PEGylated polyacridine peptides establish their ability to form open or closed polyplex structures that stabilize DNA from metabolism in mice and allow a stimulation of a high volume rapid dose of saline to complete the gene transfer, even after a 1 hour delay following the primary dose. This is still only a starting point toward the development of a nonviral delivery system that produces high level expression in animals without stimulation. However, given the modularity of polyacridine peptides, it should be possible to build multi-component gene delivery systems that drive the DNA further toward the nucleus, to ultimately achieve this aim.