Introduction

Delivery of exogenous genes to airway epithelium in vivo has been limited by several physiological barriers. For viral vectors such as adenoviral and adeno-associated viruses, the receptors appear to be located predominantly on the basolateral surface of epithelial cells, largely inaccessible to reagents delivered into the airway [reviewed in¹]. DNA complexed with cationic lipids is successful in accessing airway epithelial cells², but is less effective in delivering DNA to the nucleus of nondividing cells³,⁴. Since airway epithelium turns over slowly, few of the cells are dividing at any given time, so gene delivery by this method has been inefficient. Indeed, some reports indicate that expression of DNA delivered to airway epithelium by cationic lipids is no more efficient than expression from naked DNA alone⁵,⁶,⁷.

DNA compacted with polycations accesses the nucleus of nondividing cells more efficiently than noncompacted DNA or lipid–DNA complexes⁴,⁸ and thus affords promise in addressing the problem of gene expression in nondividing cells. However, complexes consisting of only polycation and DNA tend to aggregate in tissue fluids and are unstable at this ionic strength and sometimes are toxic. Addition of polyethylene glycol (PEG) 1Abbreviations used: poly K, poly-l-lysine; PEG, polyethylene glycol; PEI, polyethylenimine; SD, standard deviation; TFA, trifluoroacetate; NMR, nuclear magnetic resonance; 4-PDS, 4,4'-dithiodipyridine; BGH, bovine growth hormone; i.t., intratracheal; i.n., intranasal; RLU, relative light units; SEM, standard error of the mean. to poly-L-lysine (poly K)-based complexes was shown to influence compacted DNA particle characteristics including globular structure, steric stabilization, aggregation, and associative properties of the poly K to DNA⁹,¹⁰. Furthermore, PEG modification prevented aggregation of compacted DNA particles and enhanced transfection efficiency in vitro¹¹,¹²,¹³, as well as increasing the circulation time¹⁴,¹⁵,¹⁶ and gene transfer efficacy in liver¹⁷ in vivo when cross-linked compacted DNA complexes were delivered intravenously. We have developed stabilized nanoparticles consisting of poly K covalently linked to PEG, compacted with DNA, into nanoparticles less than 25 nm in their minor diameter. We tested these nanoparticles for their ability to deliver reporter genes to airway epithelium of mice following luminal application and compared their gene transfer efficacy to that obtained with naked DNA, as well as other methods of DNA compaction. We report that the addition of PEG is necessary for efficient delivery of compacted DNA to the airway epithelium and that gene transfer occurs in a dose-dependent fashion.

Results

Structure and stability of compacted DNA

When CK₃₀-PEG-10kDa was prepared, 4-PDS assay indicated that all free SH groups had reacted, so all CK₃₀ molecules were substituted on the cysteine with PEG in these experiments. Sedimentation assay showed that 109% of the original amount of DNA remained in solution. In turbidity assays, the slope was -4.4. Serum stability assays showed a half life of 4.0 h. Compacted DNA nanoparticles appeared as electron-dense particles by electron microscopy. When acetate was the counterion, particles formed into rods and toroids with a length average of 211 29 nm and a width of 10 3 nm (Fig. 1A). Compacted DNA had a potential of 4.13 mV (slightly positively charged) and had retarded migration on agarose gel electrophoresis (Fig. 1B, compacted DNA lane). Intact plasmid could be fully recovered by trypsinization of the lysine peptide (Fig. 1B, compacted DNA + T lane). Furthermore, compaction protected DNA against degradation in 75% mouse serum (Fig. 1B, compacted DNA + S + T lane).

Figure 1.

Structure and stability of compacted DNA stabilized with PEG. (A) Electron micrograph of pKCERlucSV plasmid DNA compacted with CK₃₀ poly K, substituted with PEG (acetate counterion), suspended in physiologic saline. (B) Agarose gel electrophoresis analysis of compacted DNA. Naked intact DNA is compared to compacted DNA, DNA recovered from compacted particles with trypsinization (compacted DNA + T), and DNA recovered, with trypsinization, from compacted particles that were incubated in 75% mouse serum at 37°C for 2 h (compacted DNA + S + T). Arrow marks the wells. (C) Expression levels obtained with intranasal administration of a fresh preparation of compacted DNA (fresh) versus an identical formulation prepared 23 months prior to the date of experiment and stored at 4°C (stored) and levels obtained from saline-treated animals (group 3).

Full figure and legend (124K)

This formulation was stable in physiologic saline at 4°C for 23 months, as assessed by the physical methods described above and by ability to transfect in vivo. Following prolonged storage at 4°C (23 months), compacted luciferase DNA, delivered by i.n. instillation, produced expression similar to an identical formulation prepared fresh the week of the experiment (Fig. 1C). Activity observed corresponded to expression levels of 13.59 1.53 pg luciferase/mg protein for fresh and 11.29 2.00 pg luciferase/mg protein for stored preparations.

The impact of the method of administration on expression levels and reproducibility

We used two methods of dose instillation, i.t. and i.n. The i.t. operative procedure resulted in 10% mortality, whereas with the i.n. procedure very few animals were lost. With both procedures, an average of 10% of mice that were dosed with 10–100 g compacted DNA failed to show reporter gene activity above background. Although mean gene transfer efficacy among the responders did not significantly differ whether nanoparticles were administered i.n. or i.t. (P > 0.05, Fig. 2A), the highest levels of luciferase activity were achieved with i.t. administration. Expression levels were 54.86 19.11 pg luciferase/mg protein for i.t. and 18.15 4.39 pg luciferase/mg protein for i.n. administration.

Figure 2.

The effect of method of administration on reproducibility and efficacy. Mice received 100 g compacted pKCERlucSV or pKCPIRlucBGH by intratracheal or intranasal bolus administration. Similar reagents from different batches were used for the various experiments shown. Lungs were harvested 2 days following dosing and assayed for luciferase activity. (A) Luciferase activity levels achieved with i.t. vs i.n. administration. Plotted values reflect actual activity minus mean background activity obtained from lungs of animals that received saline alone (mean i.t. vs mean i.n., P = 0.074). (B) Expression levels observed with compacted DNA formulations prepared on 6 separate occasions and administered i.t. with plasmid from two different vendors. Plotted values reflect actual activity minus mean background activity obtained from lungs of animals that received saline alone. ANOVA shows no difference among experiments. (C) Luciferase activity in the lungs observed following intranasal administration of compacted DNA on 16 different occasions with two different plasmids—pKCERLucSV or pKCPIRLucBGH. Plotted values reflect actual activity minus mean background activity obtained from lungs of animals that received saline alone. ANOVA shows no differences among experiments.

Full figure and legend (100K)

To evaluate the reproducibility of our compacted DNA formulation, we compared luciferase expression obtained for nanoparticles prepared at different times with different batches of the same compacting reagents (Figs. 2B and 2C). For the i.t. experiments (Fig. 2B), the plasmid was obtained from two different sources. Mean expression levels did not differ significantly (P > 0.05) between 6 experiments. Similar results were observed for i.n. administration (Fig. 2C). No significant difference in expression levels was observed when two different luciferase plasmids were used in 16 different experiments.

Dose response and time course of expression

To evaluate the relationship between gene expression and compacted DNA dose, animals received 25 l intratracheal boluses of naked or compacted pKCERlucSV DNA ranging in concentration from 0.4 to 12 mg/ml (dose of 10, 30, 100, or 300 g of DNA). Luciferase enzyme activity measured 2 days after gene transfer revealed dose-related expression of both naked and compacted DNA, with a plateau reached at approximately 100 g of DNA. Compacted DNA generated approximately 100-fold higher levels of gene expression than naked DNA when administered in this fashion (Fig. 3). At the optimal dose (100 g DNA), compacted DNA resulted in expression of 186.39 96.90 pg luciferase/mg protein, compared to 7.18 4.24 pg luciferase/mg protein for naked DNA. For compacted DNA, one animal in each of the 10-, 30-, and 100-g dose groups (12.5%) failed to exhibit activity levels above background, defined as the mean value from the saline-treated animals plus three times the standard deviation of these values ("3SD"), whereas four animals in the 300-g dose group did not exceed this background level of activity (50%, Fig. 3A). A plot of the means for animals that exhibited luciferase activity above 3SD (Fig. 3B) revealed that maximal activity is achieved with approximately 100 g DNA. We used this dose for the remainder of our studies. We detected no luciferase activity above background in 50–88% of the animals treated with naked DNA. We observed no expression in the lungs of saline-treated controls or the livers of any of the animals.

Figure 3.

Dose-dependent gene transfer with compacted and naked DNA. Two days following intratracheal administration of 10, 30, 100, or 300 g of either compacted or naked DNA, lungs were harvested and assayed for luciferase activity. All doses were delivered in 25 l. (A) Activity levels for individual animals. Solid line represents mean of blank plus 3 standard deviations. (B) Values exceeding this mark were used to plot the mean activity levels for naked DNA and compacted DNA.

Full figure and legend (78K)

To evaluate the time course of transgene expression in lung, we administered 100 g of naked or compacted DNA to mice and measured the luciferase activity over a 12-day period. Activity levels remained significantly different from the background + 3SD on day 12 in mice that received compacted DNA (Fig. 4A), but not in mice that received naked DNA (Fig. 4B). Activity was highest on day 2, the earliest time point examined, and declined on days 5 and 12, with a half-life of 1.4 days (Fig. 4C).

Figure 4.

Time course of gene expression. Mice received 100 g compacted or naked pKCERlucSV DNA by intratracheal administration. Lungs were harvested at days 2, 5, and 12 after dosing. Individual activity levels for compacted DNA (A) and naked DNA (B) are shown. Solid line represents mean of saline-dosed control lungs plus 3 standard deviations. Values exceeding this limit were used to calculate mean expression levels and half-life of expression for compacted DNA (C).

Full figure and legend (73K)

No luciferase activity above background was detected in the livers of any of animals in our experiments. In the trachea, low-level (compared to lung extracts) mean expression significantly different from background was detected in some animals dosed with compacted and naked DNA. Maximal expression levels reached 10⁶ RLU/mg protein (30 pg/mg protein) for compacted DNA and 10⁴ RLU/mg protein (0.55 pg/mg protein) for naked DNA (data not shown). However, expression was highly variable, and mean levels for compacted DNA did not significantly differ from those for naked DNA.

The effect of PEG and compacting reagent on gene transfer efficacy

To test the importance of inclusion of PEG in the complex, we compared levels of luciferase activity achieved with DNA compacted with CK₃₀ to those obtained with CK₃₀PEG10kDa (Fig. 5A). We observed no activity above background in the lungs of animals that received compacted DNA lacking PEG. When the DNA was compacted with CK₃₀PEG10kDa, activity levels were significantly higher than those obtained with non-PEGylated CK₃₀ (P < 0.05). Longer chain polymers of lysine (K256) lacking PEG also failed to transfect the lungs of dosed animals (Fig. 5B). However, when we used linear polyethylenimine (PEI) of similar size (M_r 22 kDa) to compact 13 g DNA, efficient gene transfer was observed (Fig. 5C). Although this level of activity was different from background (P < 0.05), it did not differ from levels obtained in animals dosed with 13 g of CK₃₀PEG10kDa compacted DNA.

Figure 5.

Effect of PEG and compacting reagent on gene transfer efficacy. Luciferase activity levels obtained in mouse lungs, 2 days following i.n. administration of DNA compacted with different reagents. Solid lines represent means of blank plus 3 standard deviations. (A) Activity for untreated mice or ones that received pKCERlucSV (75 g) compacted with CK₃₀, pKCERlucSV (125 g) compacted with CK₃₀PEG10K, or pKCPIRlucBGH (125 g) compacted with CK₃₀PEG10K. (B) Luciferase activity levels for mice that received saline, pKCERlucSV (82 g) compacted with linear poly K (chain length 256 lysines), or pKCERlucSV compacted with CK₃₀PEG10K. (C) Luciferase activity in lung for animals that received pKCERlucSV (13 g) compacted with linear PEI (M_r 22 kDa) or with CK₃₀PEG10K.

Full figure and legend (66K)

Distribution and localization of reporter genes

To examine tissue distribution of gene expression, we assayed organs from eight Balb/c mice that received i.n. administration of compacted luciferase DNA. Fig. 6 demonstrates average luciferase activity observed 2 days postadministration. Expression was significantly different from that of saline-treated matched controls only in the lung and trachea (Fig. 6). Total log-transformed mean luciferase activity for these organs was 7,413,102 5,698,810 RLU for the lungs and 3,162,277 1,690,670 RLU for the trachea. These activity levels corresponded to 263.4 206 pg luciferase per lung and 118.9 66.2 pg luciferase per trachea. Control levels for all organs, except esophagus, from saline-treated animals were <6000 RLU/organ. Although four of eight animals showed activity above background in the esophagus, and one animal had quite high luciferase activity, expression levels in this tissue were variable and the group mean did not significantly differ from that of saline-treated controls (which was 9000 RLU/mg protein, Fig. 6).

Figure 6.

Tissue distribution of luciferase activity following i.n. administration of compacted DNA. Mean luciferase activity levels observed in organs from animals that received 100 g compacted pKCPIRlucBGH DNA by intranasal administration (n = 8 except for ovary and testis, for which n = 4) compared to matched controls that received only saline (n = 2). Expression was assayed 2 days after treatment. ^**Significantly different (P < 0.01) from saline alone. Mean luciferase activity in any organ for saline-treated animals was <9200 RLU/mg protein.

Full figure and legend (60K)

To localize gene expression within the lung, we used naked and compacted pKCERgalSV delivered by i.t. instillation. We compared sections with adjacent sections subjected to the identical staining protocol but without any exposure to primary antibody, which never showed any specific staining. Patterns of expression varied from mouse to mouse and from region to region in the lungs. Two days following administration, we detected -galactosidase expression by immunohistochemistry in the epithelial cells of small airways and lung parenchyma (Figs. 7A, 7C, and 7E). We observed intense and uniform staining in three of eight mice in airway epithelial cells in the dependent portions of the right lung in animals that received compacted DNA (Fig. 7A), but more typically, expression was patchy (arrowheads, Fig. 7C) and not every cell in an airway section had been transfected (Fig. 7C, which is representative of five of the eight animals studied). Every animal treated with compacted DNA had some airway epithelial expression in the right lung, particularly in the right lower lobe. However, in the three (of eight) mice with intense staining on the right, modest expression was also seen in sections of the left lung. In two of these animals, we also observed expression in small and large blood vessels near the large airways, which was accompanied by mononuclear cell recruitment (arrowheads, Fig. 7E). We also detected expression in alveoli surrounding heavily transfected airways (Fig. 7A). In areas of high gene expression (typically, dependent portion of the right lung), gene transduction efficiency reached as high as 90% of airway epithelial cells, 30% of alveoli, and <10% of alveolar macrophages staining positive for bacterial -galactosidase. No intense -galactosidase expression was detected in animals that received naked pKCERgalSV DNA (Fig. 8A illustrates representative staining in these animals), saline alone (Fig. 8B), or adjacent sections not treated with the primary antibody (Figs. 7B, 7D, 7F, 8B, and 8D).

Figure 7.

Localization of gene expression obtained with compacted DNA in the lung. Lungs harvested from animals that received 100 g of compacted or naked pKCERgalSV DNA were fixed, sectioned, and immunohistochemically stained for the bacterial -galactosidase protein 2 days following administration. Representative sections of medium (A, 20 original magnification) and small (C, 40 original magnification) airways from right lungs of animals that received compacted pKCERgalSV DNA are shown. In two of eight mice treated with compacted DNA intense staining of small and medium pulmonary blood vessels was also observed (E, 20 original magnification). (B, D, and F) Sections adjacent to A, C, and E, respectively, which were not treated with primary antibody but otherwise were processed in identical fashion. Hematoxylin stain (blue) marks nuclei, while Nova red stain (reddish brown) marks -galactosidase protein expression.

Full figure and legend (649K)

Figure 8.

Immunohistochemical staining in the lungs of control animals. Lungs harvested from animals that received 100 g of naked pKCERgalSV DNA (A and B) or saline (C and D) were fixed, sectioned, and immunohistochemically stained for the bacterial -galactosidase protein 2 days following administration. Minimal staining was observed in sections of lungs from animals that received naked DNA (A, 20 original magnification): section shown is representative of staining identified in the eight animals that received naked DNA. Nonspecific background staining observed in animals that received saline alone (C, 10 original magnification) was minimal. (B and D) Sections adjacent to A and C, respectively, which were not treated with primary antibody but otherwise were processed in identical fashion. Hematoxylin stain (blue) marks nuclei, while Nova red stain (reddish brown) marks -galactosidase protein expression.

Full figure and legend (685K)

Discussion

Stabilized, compacted DNA nanoparticles transfect epithelial cells of the murine lung. At its peak, expression was more than 2 logs in excess of that observed in animals dosed with naked DNA. Expression was comparable to that achieved with complexes made with 22-kDa PEI. Expression was dose-related and was maximal shortly after administration, declining in log-linear fashion. Compacted DNA stabilized with PEG can be prepared without difficulty at concentrations up to 12 mg/ml, whereas naked DNA at this concentration became extremely viscous. Moreover, since compacted, stabilized DNA is prepared in normal saline, no stimulation of airway irritant receptors from either hyper- or hypotonic solutions is incurred.

These nanoparticles transfect predominantly airway epithelial cells, as demonstrated by immunohistochemistry for bacterial -galactosidase. Alveolar epithelial cells and macrophages, cells that might also be expected to come into direct contact with a complex administered into the airway, were also transfected, although with less intensity. In two animals, expression was detected in the endothelial cells of pulmonary veins, and this expression was associated with mononuclear cell infiltrates along these blood vessels. Therefore, in some cases, the nanoparticles can penetrate the epithelial barrier and transfect other cells. Whether the two animals that displayed such endothelial expression and surrounding inflammation had breaches in the epithelium that permitted systemic access, or whether they simply had sufficient transepithelial penetration of the complex to be detected at nonairway sites, is not certain. However, if the nanoparticles gained systemic access, we might expect reporter gene expression in other organs, such as the liver, which is well vascularized, or the heart, which is exposed directly to blood. No such expression was detected. In the tissue survey, outside the lung, only the esophagus gave a hint of expression (activity observed in only one of eight animals), which we speculate may have occurred in animals that swallowed the complex, which was administered via the nose and then aspirated. No inflammatory infiltrates were observed in animals in which reporter gene expression was detected only in the airways (six of eight animals in this experiment). No bronchial or alveolar infiltrates were detected in any animal. Further studies to understand the systemic access of these compacted DNA nanoparticles and the consequences thereof will be important to understand this potential mechanism of toxicity.

The cells that we observed to be transfected (epithelial cells, macrophages, and endothelial cells) all turn over slowly, and in all likelihood, most had intact nuclear envelopes at the time of transfection. This distribution of reporter gene expression suggests that at least in some cases, compacted DNA can access the nucleus of nondividing cells in vivo and be expressed. In vitro data from our laboratories indicate that the PEG-substituted DNA nanoparticles can access the nucleus after microinjection into the cytoplasm or delivery to growth-arrested cells⁸ and that complexes with fluorescent labels on either DNA or CK₃₀PEG10kDa can be detected in intact nuclei by 15 min after application of the complex to the apical surface of human airway epithelial cells in primary culture grown at the air–liquid interface¹⁸. This property may be a crucial advantage for this nonviral method of gene transfer. Microinjection data indicate that nuclear access depends on the minimum diameter of the nanoparticle, such that those particles smaller than 25 nm (the diameter of the nuclear pore) have the most activity⁸. Our complexes meet this criterion, since the rods have minor diameters of 7–14 nm. Previous reports on nuclear translocation of polycation DNA complexes in cultured airway epithelial cells stressed the requirement of targeting ligands, such as sugars¹⁹,²⁰,²¹,²², for this process to occur. Our PEGylated complexes do not require such a ligand. Moreover, our complexes did not include any endosomolytic moieties, nor did we dose the animals with drugs intended to disrupt endosomes. Others have reported that such reagents enhance gene transfer in the liver following intravenous administration of compacted DNA¹¹,¹⁴,¹⁵,¹⁶, and such modifications of our complexes may enhance our results.

Maximal luciferase expression occurred 2 days after transfection, at the earliest time point examined, declined steadily thereafter with a half-life of about 1.4 days, and remained above background levels even at day 12. This first-order, log-linear decline suggests steady degradation of the luciferase protein and mRNA, as if DNA to be expressed was delivered to the nucleus all at once, and there is no reservoir of luciferase DNA that feeds the nucleus to sustain expression over time. This observation correlates with reports of gene transfer duration with cross-linked PEG DNA complexes in the liver following intravenous administration¹⁷. In other experience in vitro with receptor-directed DNA–poly K complexes without PEG, prepared in high-salt solutions, reporter gene activity in vitro, with luciferase as transgene, persists at or near the initial levels for weeks²³. In vivo, transgene activity often increases or persists at the initial level for 10 days following complex administration²³,²⁴. Although the reporter genes in the in vivo studies were lacZ, human factor IX, and human CFTR, all of which have longer half-life than luciferase, still, the increase in activity over a 4- to 10-day period following transfection is striking. Differences in the duration of gene expression attained by our complexes may reflect multiple factors, including varying rates of nuclear uptake and promoter regulation.

Although our nanoparticles efficiently deliver genes to the majority of animals dosed, some animals did not exhibit gene transfer. Bolus administration probably accounts for some of this variability. We could not correlate the level of expression with the level of anesthesia of the mouse or with our observations on whether the mouse coughed following administration of the complex, but it seems likely that for some animals, the entire dose, or a portion of it, was simply expelled before it reached the lower airways. We speculate that this is the case since, for the 100-g dose of compacted DNA, there is not a continuum between the highest and the lowest expression recorded—rather, there is a cluster of animals with substantial expression and a few in which expression does not differ from background. The bolus intratracheal dose probably follows gravity and ventilation into the lungs, so the most dependent portions are the most heavily transfected, and the most direct bronchial route into the lower airways is followed preferentially. This probably accounts for the striking regional distribution of the expression in the immunohistochemical studies. Intranasal administration, which may produce more small droplets, is also susceptible to the mouse cough response, and some of the dose may be swallowed, as was evidenced by the occasional reporter gene expression detected in the esophagus and higher levels of tracheal expression, compared to i.t. administration, which is more likely to bypass the pharynx and much of the trachea as well. Following intranasal administration of complex, 10–15% of total activity is found in trachea and esophagus, while 85–90% is found in the lungs. Transfection of the esophagus is particularly interesting given the rapid directional transit of luminal contents, which would be predicted to limit exposure time of these nanoparticles to the cells of the esophagus. Thus, although 5–10% of activity is found in esophagus, more of the compacted DNA dose may have been delivered to that site, since transfection is probably less efficient. Although i.t. and i.n. administration gave relatively reproducible expression levels between experiments, we speculate that better methods of application of these nanoparticles to the airways, such as preformed aerosols, might improve their intrapulmonary distribution as well as their ability to transfect.

Gene expression from naked DNA was variable, and in our experiments, expression was minimal. In some reports⁶,⁷, naked DNA is quite efficient in airway transfection, though in others, results have been disappointing⁴,²⁵. In our studies, naked DNA, delivered in normal saline, usually did not give sufficient expression of bacterial -galactosidase to be detected in immunohistochemical stains, and, although luciferase expression is significantly above background in mice given 30–100 g naked DNA, expression is variable and difficult to detect reliably, and many more animals do not exceed the control levels of luciferase activity than was true for those given compacted DNA. When we compared alternative formulations using CK₃₀ without PEG or longer polymers of lysine without PEG to compact our DNA, gene transfer efficacy declined to background levels. PEI complexed with the low dose of 13 g DNA administered to each mouse achieved expression levels comparable to 13 g of compacted DNA formulations. PEI has been reported to deliver genes to the nuclei of mammalian cells in vitro⁴ and to the airway in vivo²⁶. However, this gene transfer activity is coupled with a marked toxic effect that has been described in vitro²⁷ and in vivo²⁸,²⁹. Although prior descriptions of toxicity from aerosol administration of PEI indicate that toxicity is mild at an estimated lung dose of 5 g DNA²⁸, the bolus dosing method may have exacerbated the toxic effects. Therefore, we could not compare the maximally effective doses of DNA compacted with CK₃₀PEG10kDa with a comparable dose of PEI-complexed DNA.

The nanoparticles studied here have important advantages for gene therapy of airway epithelium. Because they are stable in normal saline, the nanoparticles can be administered without triggering airway responses to hypo- or hypertonic saline. High concentrations of DNA can be achieved, limiting the volume that must be delivered to the airway. The nanoparticles themselves are prepared from pure and uniform reagents (poly K chain length is uniform, substitution with PEG is uniform and at the same site on each molecule, and the number of substituted poly K molecules per complex is the same from batch to batch) and are stable after storage at 4°C in saline, for at least 23 months. An important, even critical, attribute of these nanoparticles is their ability to access the nucleus of nondividing airway epithelial cells from the apical surface. These advantages, together with the demonstrated success of delivery of reporter genes to the lung with these nanoparticles, make these reagents of interest for gene therapy directed at airway epithelial cells.

Methods

Materials

For intratracheal studies, poly K exactly 30 lysine residues in length with an N-terminal cysteine (CK₃₀), trifluoroacetate (TFA) salt, was custom prepared by solid-phase synthesis by Dr. S. Yadav at the Cleveland Clinic Foundation (Cleveland, OH). Purity and size of the product were verified by HPLC and mass spectrometry by the manufacturer. For intranasal administration studies, the CK₃₀ peptide was synthesized by PolyPeptide Laboratories (Torrance, CA). Purity, checked by HPLC, was 95%, and identity was confirmed by mass spectrometry. PEG, thiol-reactive methoxy-PEG-maleimide, M_r 10,000 (average), was custom synthesized by Shearwater Polymers, Inc. (Huntsville, AL), and contained 0.30 wt% impurities (assessed by ¹H NMR). Linear poly K (256-mer) used in comparative studies was obtained from Sigma Chemical Co. (St. Louis, MO). ExGen500 (linear PEI, M_r 22,000) used for comparative studies was purchased from Fermentas, Inc. (Hanover, MD). DNA plasmids were custom produced by BIO 101, Inc. (Carlsbad, CA) or Aldevron (Fargo, ND) and supplied in water. Tissue lysis and luciferase assay reagents were obtained from Promega Corp. (Madison, WI). Protein concentration assays were performed using the DC Protein Assay Kit from Bio-Rad Laboratories, Inc. (Hercules, CA). Immunohistochemistry reagents were obtained from Vector Laboratories, Ltd. (Burlingame, CA).

Preparation of polycation carrier

CK₃₀ was conjugated (1:1 molar ratio) to PEG via the sulfhydryl group of CK₃₀ and the reactive maleimide group on the PEG to generate a noncleavable CK₃₀–PEG (C–S bond). The CK₃₀ was dissolved in phosphate-buffered saline, pH 7.4, with 5 mM EDTA and mixed vigorously at room temperature with an equal volume of DMSO containing PEG-maleimide and left overnight. Completion of the reaction was verified using a 4-PDS assay for unreacted sulfhydryls³⁰. The product, CK₃₀PEG10kDa, was purified by size-exclusion chromatography (Sephadex G15, 2.5 50, MWCO 1500) with a 50 mM ammonium acetate buffer. This buffer allows for the exchange of TFA counterions with acetate. The absorbance of fractions at 220 nm was monitored. The high-molecular-weight peak was pooled and lyophilized. Since each PEG is attached to a single CK₃₀, and there are no reactive SH groups in the final complex preparation, each PEG must have been added at the cysteine, although maleimide is capable of reaction with amino groups as well (albeit with 1000-fold less avidity³¹).

Reporter genes and plasmid preparation

Three plasmids encoding luciferase or bacterial -galactosidase and kanamycin resistance were used, pKCERlucSV (5.9 kb), pKCPIRlucBGH (5.3 kb), and pKCERgalSV (7.5 kb), respectively. The pKCER-SV plasmids comprise a CMV enhancer, the elongation factor-1 promoter and first intron, the RU5 translational enhancer from HTLV, and the SV40 late polyadenylation signal. The pKCPIRlucBGH plasmid contains a CMV enhancer, the CMV promoter, and the BGH polyadenylation signal. Identity of the plasmids was confirmed by restriction endonuclease digestion, and purity was established by 1.0% agarose gel electrophoresis.

Preparation of compacted DNA

DNA (200 g/ml) was added to a tube containing twofold excess CK₃₀PEG10 polycation (+/- charge ratio). The compacted DNA was filtered through a 0.2-m Millipore filter and dialyzed overnight against 0.9% NaCl at 4°C. The DNA was then filtered again through the 0.2-m filter and concentrated using centrifugation (Ultrafree, Millipore, NMWL 100 kDa) to the desired concentration. Biophysical characterization on the final formulation included sedimentation, turbidity, gel, electron microscopy, and serum stability analyses as described below.

For experiments examining other formulations of compacted DNA, we used PEI (Exgen 500) and poly K₂₅₆ (256-mer of lysine) as compacting agents. The PEI is linear and is 22 kDa in molecular mass. DNA was mixed with the PEI just before the dosing of animals according to the manufacturer's suggestions: PEI (0.1 ml) was diluted in 5% glucose to 0.625 ml and added in 100-l aliquots to vortexed DNA (0.208 ml at 3 mg/ml). The final charge ratio of our PEI:DNA was 6:1 (+ to -). The poly K₂₅₆ (Br salt) DNA compaction was performed by stepwise addition of DNA (0.2 mg/ml) to vortexed poly K₂₅₆ (2.5 mg/ml). The final charge ratio of our poly K₂₅₆:DNA was 2:1 (+ to -).

Sedimentation analysis

A minimum of 50 l compacted DNA was used for centrifugation. DNA concentration was determined by OD₂₆₀ measurement. The compacted DNA was then centrifuged at 6000 rpm or 3406g for 1 min, and the DNA concentration was then reassayed. Aggregated DNA particles sediment and result in low OD₂₆₀ values. Acceptable samples gave a recovery of 80–120% of DNA measured in the supernatant.

Turbidity analysis

A 1-ml water blank quartz cuvette was inserted into a Molecular Devices SpectraMax spectrophotometer for path check. One hundred eighty microliters of compacted DNA samples was then loaded into 96-well plate wells containing 180 l of the blank. Absorbance was determined at wavelengths 330, 347, 364, 381, 398, and 415 nm, and the data were analyzed by the SoftMax Pro software. The slopes (turbidity parameter) of the curves produced by plotting the log of apparent absorbance vs log of wavelength for each sample correlated with the size and structure of the compacted DNA particles. Acceptable formulations have a turbidity parameter of approximately -4, with an apparent absorbance 0.040 at 330 nm, 7% standard error of slope, and 10% variability of duplicates.

Electron microscopy

Micrograph grids were prepared as previously described²⁴. Briefly, immediately after formation of compacted DNA, a drop of a solution (1:10 dilution of compacted DNA mixture in saline) was added to a 400-mesh electron microscope carbon grid, washed with water, stained with 0.04% (in methanol) uranyl acetate, washed in ethanol, and air dried. Samples were examined using a JEOL-100C electron microscope (magnification 40,000). Acceptable formulations contained at least 80% of particles within the following size range: rods 300 nm long and 20 nm wide, toroids 100 nm in outer diameter.

potential

Compacted DNA nanoparticles at a concentration of 0.2 mg/ml were evaluated using a dynamic light scattering instrument from Particle Sizing Systems, Model NICOMP 380 ZLS, using a run time of 3 min. Carboxylated latex microspheres from Bangs Laboratories were used to validate the potential measurements.

Gel analysis and serum stability

To test the integrity of the plasmid DNA used, compacted DNA samples were loaded on a 0.8% agarose gel following trypsinization to remove poly K or following a 2-h incubation in 75% mouse serum at 37°C followed by trypsinization.

Animal experiments

All animal protocols were approved by the Case Western Reserve University Institutional Animal Care and Use Committee. We studied 8- to 10-week-old C57BL/6j wild-type mice. In the case of the tissue distribution and long-term storage stability experiments, 8- to 10-week-old Balb/c mice were used. Following intraperitoneal injection of 150 l anesthetic cocktail (2.13 mg/ml Xylazine, 0.36 mg/ml Acepromazine, and 10.75 mg/ml ketamine) animals received a tracheostomy and DNA, either compacted or naked, or saline alone was administered intratracheally as a bolus (volume 25 l) through a 22-gauge catheter. An alternate method of administration was intranasal instillation of the dose (in 25 l) by aliquot to the nose bridge (2.5 l every 2–3 s). Intratracheal instillation was performed while the mice were affixed to a surgical board at a 30% incline. Intranasal instillation was performed while mice were held upright and reclined slightly backward. In all studies, the DNA plasmids or controls were prepared and coded by investigators who did not participate in dosing the mice or subsequent analyses, and the identity of the groups was revealed only after the data analysis was completed. The blind was preserved in all experiments except the dose–response study in which the highest concentration of naked DNA was noticeably more viscous than any other sample and was readily apparent to the investigators dosing the mice. For initial experiments testing the efficacy of compacted DNA in mice, treatment groups included animals that received 100 g compacted luciferase plasmid (pKCERlucSV) in 50 l saline, 100 g naked luciferase DNA in 50 l saline, or 50 l saline alone (n = 8 or 9 animals per group). Luciferase activity was assayed in lung, trachea, and liver extracts 2 and 4 days following gene transfer. Applying the dose in 25 l volume rather than 50 l resulted in lower variability in gene expression in experimental groups, so this volume was used for future experiments.

For experiments examining the dose response to compacted DNA, four groups of mice (n = 8 each) received 10, 30, 100, or 300 g of compacted pKCERlucSV in 25 l saline. Controls included 4 groups (n = 8 each) that received 10, 30, 100, or 300 g of naked pKCERlucSV in 25 l saline and a group (n = 6) that received 25 l saline alone. Expression in whole lung extracts was measured 2 days following gene transfer. To examine the time course of transgene expression, three groups of animals (one group for each time point) received 100 g compacted pKCERlucSV in 25 l saline (n = 8 each). Animals were then sacrificed and whole lung extracts collected on days 2, 5, and 12. Controls, sacrificed at the same time points, included three groups of animals (n = 8) that received 100 g naked pKCERlucSV in 25 l saline and two groups of two mice each (sacrificed on days 2 and 4) that received 25 l saline alone.

To test the effect of the compacting reagent on in vivo efficacy, we conducted three i.n. administration experiments. In the first we tested the importance of including PEG in the complex. Three groups of mice (n = 6 each) received pKCERlucSV compacted with non-PEGylated CK₃₀ (75 g DNA/mouse), CK₃₀PEG10kDa (130 g DNA/mouse), or pKCPIRlucBGH compacted with CK₃₀PEG10kDa (130 g DNA/mouse). The control group was untreated. A second experiment examined the efficacy of long-chain poly K (256 lysine chain). Mice were dosed with 87 g of pKCERlucSV DNA compacted with either K256 (n = 6) or CK₃₀PEG10kDa (n = 5) and compared with a saline-treated control group (n = 6). In the third experiment, we dosed mice with 13 g pKCERlucSV DNA compacted with either the ExGen500 reagent (linear 22-kDa PEI, compacted according to the manufacturer's instructions) or CK₃₀PEG10kDa. For this experiment, the low dose was used due to the toxicity of higher doses of PEI-compacted DNA particles. For all three experiments, mice were sacrificed 2 days following the i.n. administration, and their lungs were harvested, homogenized, and assayed for luciferase activity.

To examine the tissue distribution of transgene expression, 100 g compacted pKCPIRlucBGH DNA was administered by intranasal dosing to eight mice (four female and four male). Lung, liver, trachea, spleen, esophagus, brain, heart, kidney, intestines, pancreas, stomach, and testes or ovaries were harvested and tested for luciferase activity 2 days after administration. Control animals (n = 4) received saline alone. For localization of transgene expression, we used the bacterial -galactosidase reporter gene. Animals received 100 g compacted pKCERgalSV and lungs were collected 2 days following gene transfer (n = 8). On the day of sacrifice, animals were exsanguinated. Lungs were then collected intact and inflation fixed in 2% paraformaldehyde at 4°C overnight. Following fixation, lungs were embedded in paraffin blocks, sectioned at 5-m intervals, mounted on glass slides, and stained for -galactosidase protein. Controls included two groups of animals that received 100 g naked pKCERgalSV DNA (n = 7) in 25 l saline or were uninjected (n = 4).

Assays

Assay for luciferase activity

After sacrifice, mouse tracheas, lungs, and livers were harvested and frozen at -80°C. Organs were then thawed and homogenized on ice in freshly prepared cell lysis reagent using a Polytron homogenizer (Model PT 1200). Mouse tracheas and lungs were homogenized in 1 ml of lysis reagent, while livers were homogenized in 2 ml. Homogenates were then centrifuged, and the supernatants were collected and stored at -65°C or colder. For luciferase activity analysis, all solutions were prepared fresh. eighty-microliter aliquots of luciferin solution were added to 20 l of each sample and luminescence was detected using a Berthold Lumat LB9507 luminometer over 10 s. Each sample was read in duplicate. Data are shown as RLU/mg protein.

Immunohistochemical detection of -galactosidase expression.

Mounted tissue sections were deparaffinized, hydrated, permeabilized in methanol at -20°C for 10 min, and incubated for 30 min in 0.3% H₂O₂ to quench endogenous peroxidase activity. Slides were blocked in goat serum for 1 h at room temperature and incubated with rabbit anti-bacterial -galactosidase (1:200 dilution, Research Diagnostics, Inc., Flanders, NJ) at 4°C overnight. The secondary antibody was biotinylated goat anti-rabbit IgG (Vector Laboratories, Burlingame, CA). Samples were incubated with VECTASTAIN Elite ABC Reagent for 30 min at room temperature, washed, and stained with Vector Nova red (Vector Laboratories) for 15 min. Each stained section was matched by an adjacent section not treated with the primary antibody. Sections were counterstained with hematoxylin. Transmitted image analysis was performed with an Olympus Camedia C-2020 Z digital camera attached to an inverted light microscope.

Statistical analysis

Data are expressed as log-transformed mean SEM. Both t test and ANOVA assume that the tested populations all have the same variance (and standard deviation). We found that different sets of luciferase data have different variances. We also found that standard deviation is proportional to the mean. However, after log transformation, the variances are not significantly different, so the data could be analyzed using parametric t tests and ANOVA³². For some analyses, animals having luciferase activity < blank + 3SD were assumed not to have received the dose and were not included in statistical analysis; however, all data points are shown in the figures.

References

1. Boucher, R. C. (1999). Status of gene therapy for cystic fibrosis lung disease. J. Clin. Invest. 103: 441–445. | PubMed | ISI | ChemPort |
2. Lee, E. R., et al. (1996). Detailed analysis of structures and formulations of cationic lipids for efficient gene transfer to the lung. Hum. Gene Ther. 7: 1701–1717. | PubMed | ISI | ChemPort |
3. Zabner, J., Fasbender, A. J., Moninger, T., Poellinger, K. A. and Welsh, M. J. (1995). Cellular and molecular barriers to gene transfer by a cationic lipid. J. Biol. Chem. 270: 18997–19007. | Article | PubMed | ISI | ChemPort |
4. Pollard, H., Remy, J. S., Loussouarn, G., Demolombe, S., Behr, J. P. and Escande, D. (1998). Polyethylenimine but not cationic lipids promotes transgene delivery to the nucleus in mammalian cells. J. Biol. Chem. 273: 7507–7511. | Article | PubMed | ISI | ChemPort |
5. Cockett, M. I. (1999). Technology evaluation: cystic fibrosis therapy, Genzyme. Curr. Opin. Mol. Ther. 1: 279–283. | PubMed | ChemPort |
6. Knowles, M. R., et al. (1998). A double-blind, placebo controlled, dose ranging study to evaluate the safety and biological efficacy of the lipid–DNA complex GR213487B in the nasal epithelium of adult patients with cystic fibrosis. Hum. Gene Ther. 9: 249–269. | PubMed | ChemPort |
7. Zabner, J., et al. (1997). Comparison of DNA–lipid complexes and DNA alone for gene transfer to cystic fibrosis airway epithelia in vivo. J. Clin. Invest. 100: 1529–1537. | PubMed | ISI | ChemPort |
8. Liu, G., et al. (2003). Nanoparticles of compacted DNA transfect postmitotic cells. J. Biol. Chem. 278: 32578–32586. | Article | PubMed | ISI | ChemPort |
9. Mannisto, M., et al. (2002). Structure–activity relationships of poly(L-lysines): effects of pegylation and molecular shape on physicochemical and biological properties in gene delivery. J. Controlled Release 83: 169–182. | Article | ChemPort |
10. Park, S. and Healy, K. E. (2003). Nanoparticulate DNA packaging using terpolymers of poly(lysine-g-(lactide-b-ethylene glycol)). Bioconjugate Chem. 14: 311–319. | Article | ChemPort |
11. Lee, H., Jeong, J. H. and Park, T. G. (2002). PEG grafted polylysine with fusogenic peptide for gene delivery: high transfection efficiency with low cytotoxicity. J. Controlled Release 79: 283–291. | Article | ChemPort |
12. Oupicky, D., Konak, C., Ulbrich, K., Wolfert, M. A. and Seymour, L. W. (2000). DNA delivery systems based on complexes of DNA with synthetic polycations and their copolymers. J. Controlled Release 65: 149–171. | Article | ChemPort |
13. Choi, Y. H., Liu, F., Park, J. S. and Kim, S. W. (1998). Lactose-poly(ethylene glycol)-grafted poly-L-lysine as hepatoma cell-targeted gene carrier. Bioconjugate Chem. 9: 708–718. | Article | ChemPort |
14. Kwoh, D. Y., et al. (1999). Stabilization of poly-L-lysine/DNA polyplexes for in vivo gene delivery to the liver. Biochim. Biophys. Acta 1444: 171–190. | PubMed | ChemPort |
15. Ward, C. M., Pechar, M., Oupicky, D., Ulbrich, K. and Seymour, L. W. (2002). Modification of pLL/DNA complexes with a multivalent hydrophilic polymer permits folate-mediated targeting in vitro and prolonged plasma circulation in vivo. J. Gene Med. 4: 536–547. | Article | PubMed | ChemPort |
16. Oupicky, D., Carlisle, R. C. and Seymour, L. W. (2001). Triggered intracellular activation of disulfide crosslinked polyelectrolyte gene delivery complexes with extended systemic circulation in vivo. Gene Ther. 8: 713–724. | Article | PubMed | ISI | ChemPort |
17. Kwok, K. Y., Park, Y., Yang, Y., McKenzie, D. L., Liu, Y. and Rice, K. G. (2003). In vivo gene transfer using sulfhydryl cross-linked PEG-peptide/glycopeptide DNA co-condensates. J. Pharm. Sci. 92: 1174–1185. | Article | PubMed | ChemPort |
18. Kube, D. M. and Davis, P. B. (2003). Intracellular trafficking of nontargeted stabilized molecular conjugates in human airway epithelial cells. Mol. Ther. 7: S371. | ISI |
19. Fajac, I., Briand, P., Monsigny, M. and Midoux, P. (1999). Sugar-mediated uptake of glycosylated polylysines and gene transfer into normal and cystic fibrosis airway epithelial cells. Hum. Gene Ther. 10: 395–406. | Article | PubMed | ChemPort |
20. Kollen, W. J., et al. (1999). High-efficiency transfer of cystic fibrosis transmembrane conductance regulator cDNA into cystic fibrosis airway cells in culture using lactosylated polylysine as a vector. Hum. Gene Ther. 10: 615–622. | Article | PubMed | ChemPort |
21. Klink, D. T., Chao, S., Glick, M. C. and Scanlin, T. F. (2001). Nuclear translocation of lactosylated poly-L-lysine/cDNA complex in cystic fibrosis airway epithelial cells. Mol. Ther. 3: 831–841. | Article | PubMed | ChemPort |
22. Ziady, A. G., Kelley, T., Milliken, E., Ferkol, T. and Davis, P. B. (2002). Functional evidence of CFTR gene transfer in nasal epithelium of CF mice following apical application of SEC receptor targeted DNA complexes. Mol. Ther. 5: 413–419. | Article | PubMed | ChemPort |
23. Ziady, A. G., Ferkol, T., Dawson, D. V., Perlmutter, D. H. and Davis, P. B. (1999). Chain length of the polylysine in receptor-targeted gene transfer complexes affects duration of reporter gene expression both in vitro and in vivo. J. Biol. Chem. 274: 4908–4916. | Article | PubMed | ChemPort |
24. Ziady, A. G., Ferkol, T., Gerken, T., Dawson, D. V., Perlmutter, D. H. and Davis, P. B. (1998). Ligand substitution of receptor targeted DNA complexes affects gene transfer into hepatoma cells. Gene Ther. 5: 1685–1697. | Article | PubMed | ChemPort |
25. Felgner, P. L., et al. (1995). Improved cationic lipid formulations for in vivo gene therapy. Ann. N. Y. Acad. Sci. 772: 126–139. | PubMed | ChemPort |
26. Gautam, A., Densmore, C. L., Golunski, E., Xu, B. and Waldrep, J. C. (2001). Transgene expression in mouse airway epithelium by aerosol gene therapy with PEI–DNA complexes. Mol. Ther. 3: 551–556. | Article | PubMed | ChemPort |
27. Gebhart, C. L. and Kabanov, A. V. (2001). Evaluation of polyplexes as gene transfer agents. J. Controlled Release 73: 401–416. | Article | ChemPort |
28. Gautam, A., Densmore, C. L. and Waldrep, J. C. (2001). Pulmonary cytokine responses associated with PEI–DNA aerosol gene therapy. Gene Ther. 8: 254–257. | Article | PubMed | ChemPort |
29. Chollet, P., Favrot, M. C., Hurbin, A. and Coll, J. L. (2002). Side-effects of a systemic injection of linear polyethylenimine–DNA complexes. J. Gene Med. 4: 84–91. | Article | PubMed | ISI |
30. Grassetti, D. R. and Murray, J. F. Jr (1967). Determination of sulfhydryl groups with 2,2'- or 4,4'-dithiodipyridine. Arch. Biochem. Biophys. 119: 41–49. | Article | PubMed | ISI | ChemPort |
31. Hermanson, G. T. (1966). Bioconjugate Techniques. Academic Press: New York. 148.
32. Armitage, P., Berry, G. and Matthews, J. N. S. (2002). Statistical Methods in Medical Research. Blackwell Sci: London. 307–308.

Acknowledgements

We are grateful for support from the Cystic Fibrosis Foundation (ZIADY00F0 and R447-CR02); the National Institutes of Health via Grants T32 HL07415 (A.G.Z.), P30 DK27651 (P.B.D.), RO1 DK52981 (P.B.D.), and RO1 DK58318 (P.B.D.); and Copernicus Therapeutics, Inc. Drs. Davis and Ziady hold equity in Copernicus Therapeutics, Inc.