¹Department of Biopharmaceutical Sciences, School of Pharmacy, University of California, San Francisco, San Francisco, CA, USA

²Department of Anatomy, School of Medicine, University of California, San Francisco, San Francisco, CA, USA

Correspondence to: F C Szoka Jr, University of California, Dept of Pharmaceutical Sciences, School of Pharmacy, Room 926, 513 Parnassus Ave, San Francisco, CA 94143-0446, USA

^aCurrent address: Regeneron Inc., 777 Old Saw Mill River Road, Tarrytown, NY 10591, USA

We introduce a lung inflation-fixation protocol to examine the distribution and gene transfer efficiency of fluorescently tagged lipoplexes using fluorescence confocal microscopy within thick lung tissue sections. Using this technique, we tested the hypothesis that factors related to lipoplex distribution were the predominant reason that intravenous (i.v.) administration of lipoplex was superior to intratracheal (i.t.) administration for gene transfer in the murine lung. Lipoplex distribution was analyzed using digitized images of overlapping fields, reconstructed to view an entire lung lobe. Intravenously administered lipoplexes were confined to the capillary network and homogenously distributed throughout the lung lobe. In contrast, i.t. administration resulted in regional distribution of lipoplex, concentrated around bronchioles and distal airways. Not all the bronchioles were stained with lipoplex, suggesting that the airway-administered solution became channeled through certain bronchiolar pathways. A fluorescent oligonucleotide was used as a marker for cytoplasmic release of nucleic acids. Quantification of the resulting fluorescent nuclei was used to define the relationship between cytoplasmic release of nucleic acids and gene expression. Endothelial cells were stained after i.v. administration, and epithelial cells were stained after i.t. administration. The delivery of nucleic acids was also more homogeneous with i.v. administration of lipoplex than with i.t. administration. After i.t. administration, it was notable that high concentrations of fluorescent nuclei correlated with low GFP expression. This suggested that toxicity was associated with high local concentrations of cationic lipoplexes. The ratio of GFP-expressing cells to fluorescent nuclei indicated that capillary endothelial cells were more efficient in gene expression per delivery event than were pulmonary epithelial cells. Thus, the greater gene expression efficiency of i.v. administered lipoplexes was due not only to the initial distribution but also to the greater efficiency of the vascular endothelial cells to appropriately traffic and express the foreign gene. Gene Therapy (2001) 8, 828-836.

cationic lipid; confocal microscopy; gene therapy; liposome

Introduction

Cationic lipids have become a commonly used vector to mediate gene transfer, due to their ability to provide a variety of functions for gene delivery. The lipids form multivalent, positively charged liposomes which, through electrostatic interactions, condense negatively charged DNA into a small particle termed a 'lipoplex'.¹ In this complexed form, plasmid DNA is protected from nuclease degradation in vivo and is also protected from shear degradation when administered through a nebulizer/aerosolizer.^2,3 The cationic lipid also functions to bind DNA to cell surface proteoglycans, thus enhancing the efficiency of gene transfer.^4,5,6

In spite of extensive efforts to unravel the in vivo mechanisms of action of lipoplexes, simple questions concerning the variation in gene transfer efficacy between two different routes of administration have not been thoroughly examined. For instance, in mice, gene transfer to the lung has been demonstrated through both intratracheal (i.t.) instillation and intravenous (i.v.) administration of a cationic lipoplex.^7,8,9,10 However, i.v. administration routinely promotes 10- to 100-fold greater gene expression levels than i.t. administration, regardless of the formulation. Although several factors may contribute to this difference in activity, the pulmonary distribution promoted by the route of administration is, perhaps, the most likely.

Intravenous administration of lipoplex leads to the binding of approximately 80% of the dose within the lung, leading to high levels of gene expression in this organ.^10,11 Internalization of i.v. injected lipoplex has been visualized within the vast network of pulmonary capillaries in the lung, although lipoplex was not observed within the alveolar space.¹² Administration through the airways has access to the lung epithelia and should also deliver lipoplex throughout the lung parenchyma. However, it has previously been demonstrated that airway instillation can result in variable distribution of DNA between lung lobes.^8,13 In addition, gene expression can be detected in the bronchioles of the lung as well as within the alveolar acini.^14,15,16,17 Therefore, the distribution of airway-instilled lipoplex may involve different regions of the lung than that of intravenously injected material.

Fluorescent markers have been used to study the distribution and uptake of DNA lipoplexes and oligonucleotide complexes after intravenous injection.^12,18 This route of administration led to uptake and expression in capillary endothelial cells.^12,13,18 Distribution of airway-instilled lipoplex has been characterized by isotope labeling,¹⁷ but not by direct visualization within the lung. The distribution of lipoplex has been inferred by the localization of gene expression.^{16,17,19,20,21} However, it is not known what determines the observed expression levels: the distribution of the lipoplexes, or the inherent gene expression activity of various cell types.

Our objective was to develop a protocol which would allow us to study the distribution, uptake and expression of cationic lipid/DNA lipoplexes in the lungs. Electron microscopy permits analysis at high resolution, but is labor intensive and gives poor tissue sampling. The utilization of multiple fluorescent markers with fluorescence microscopy allows the simultaneous observation of the lipid and nucleotide components of the lipoplex and gives a good sample of visual fields. In particular, the incorporation of a fluorescent oligonucleotide (15-mer) is extremely informative since release of oligonucleotide into the cell cytoplasm leads to rapid localization within the cell nucleus.^22,23 In this way, we utilized the nuclear localization of oligonucleotide as a marker for cytoplasmic release of both oligo and plasmid DNA. Since plasmid DNA is reported to be more restricted in its diffusion through the cytoplasm,²⁴ the nuclear localization of oligonucleotide does not indicate identical localization of plasmid in the nucleus. However, the use of the fluorescent oligonucleotide allowed us to differentiate the uptake, internalization, and endosomal release stages of lipoplex delivery. We show that i.t. delivery of lipoplex results in comparatively less homogenous distribution, leading to lower gene expression than that obtained by i.v. administration of lipoplex.

Results

Retention of complex components after agarose inflation

A new procedure to stain, fix, and inflate the lung was used to visualize fluorescent lipoplex delivered via the lung airway. Since the perfusion and agarose inflation techniques also utilized the airway route, we were concerned with the potential for redistribution of airway-administered lipoplex during our protocol. Therefore, we radiolabeled the lipoplex and compared the amount retained in the lung, with and without the applied technique. Figure 1a shows that both the cationic lipid and DNA components were retained in the lung after airway perfusion and inflation protocols (lipid label = 93.8%, DNA label = 75.7%, n of 2), to the same extent as the no inflation control (lipid label = 100.2%, DNA label = 65.2%). In addition, a non-airway perturbing lung inflation technique was performed (vascular inflation via intra-cardiac puncture) and showed no significant removal of radioactive counts. These data indicated that the DOTAP:Chol-based lipoplex was essentially bound within the lung by 3 h after i.t. administration and would not be grossly redistributed by the airway perfusion and inflation. The cationic lipid component was important for this binding and retention in the lung, since uncomplexed DNA and anionic liposomes were washed out from the lung to a much greater extent than the complexed DNA (Figure 1a). Visual assessments of lung tissue indicated that agarose inflation, via intra-cardiac puncture, resulted in poor preservation of the airway structure (data not shown), and thus was not used in subsequent studies.

Dual fluorescence labeling of lung endothelium and epithelium

To demonstrate that the lung vasculature and airway could be differentially stained, we labeled the blood vessels with FITC-L. esculentum lectin (tail vein injection) and the airway cells with Texas Red-Wheat germ agglutinin lectin (airway perfusion). Since the lectins used were not specific to the particular cell type, the route of administration dictated whether endothelium or epithelium was stained. The dual lectin staining exhibited different characteristics (Figure 1b). The endothelial staining can be identified by its smooth surface and tubular appearance, similar to that found in previous work.¹² The epithelial surface appears to have a rough texture, and it does not overlap with the green vascular stain in serial confocal sections (data not shown). The vasculature and pneumocytes maintained distinct staining by this protocol, indicating that neither the vascular or airway compartments were significantly breached during the staining and inflation protocol.

Fluorescent lipoplex distribution after i.v. and i.t. administration

A rhodamine-PE lipid marker was incorporated into the cationic lipoplex and administered intravenously (Figure 2a) and intratracheally (Figure 3a). A counterstain of FITC-lectin was used to identify the endothelial and epithelial surfaces of the lung, and transverse sections of a lung lobe were reconstructed from digitized images acquired with fluorescence microscopy. Intravenous administration resulted in fluorescent lipid distribution throughout the lung lobe (Figure 2a, b). Although some regions appeared to have stronger fluorescence intensity than others, the lipid fluorescence was pervasive. This homogeneity was characteristic of all lung lobes and was consistent with lipoplex binding in capillaries as previously reported.^12,13,18 At higher magnification, the lipoplex appeared to be bound in small aggregates on the endothelial surface (see Figure 2b).

In contrast, intratracheal administration of lipoplex resulted in highly localized distribution within the lung lobe (Figure 3a, b). Fluorescent lipid staining was associated with certain airways and extended out towards the respiratory bronchioles. Not all bronchioles were stained with fluorescent lipid, suggesting that the gene delivery fluid volume became channeled through some fraction of bronchiole pathways. Some alveolar sacs were also stained by the lipoplex (see higher magnification image, Figure 3c). These images indicated that cationic lipoplex was delivered to both bronchiolar and alveolar epithelial cells.

Sequential sections of the right superior lobe demonstrated that lipoplex did not reach the rostral regions of the lung, but collected mostly in the middle and lower sections of the lobe (Figure 4).

Oligonucleotide delivery and GFP expression

When oligonucleotides are microinjected into the cell cytoplasm, they accumulate in nuclei within 5 min.²² Although a complete plasmid will probably not migrate to the cell nucleus in the same manner, the nuclear localization of fluorescent oligonucleotides was used as a surrogate marker of cytoplasmic release of polynucleotides.²³ The fluorescent oligonucleotide and GFP-encoding plasmid were simultaneously incorporated into lipoplexes and delivered both intratracheally and intravenously to assess the relative frequency between nucleotide delivery and gene expression with cationic lipoplex. We deliberately mixed oligonucleotide and plasmid DNA before addition to cationic liposomes to ensure the formulation of oligonucleotide and plasmid within the same lipoplex particles. The cells which exhibited fluorescent oligo localized within the nucleus would have, by inference, also released plasmid DNA into the cytoplasm.

At 24 h after i.v. administration, small flat fluorescent nuclei were clearly discerned in the vascular endothelium (Figure 5b). Regions which showed strong GFP fluorescence were found in the lung parenchyma, typically in gas exchange regions where the capillaries have close contact with alveolar epithelium. The number of GFP-transfected cells and the number of RODN-positive nuclei were quantified in several 400 ´ fields. A distribution of the data is shown in Figure 6c. Fluorescent nuclei were prevalent, showing a distribution that ranged from 12 to 85 nuclei per field, with an average of 42 positive nuclei per field (Figure 6a and Table 1). Cells expressing the green fluorescent reporter protein were sparse, at most five positive cells per field. This observation was consistent with other reports of visually detected reporter gene activity resulting from i.v. administration of lipoplexes.^12,13,25

After i.t. administration, fluorescent oligonucleotide accumulated in nuclei that were larger and round in morphology, suggesting that a different population of cells had been targeted by this route of administration (Figure 5d). Intratracheal administration also led to greater variability in the efficiency nucleotide delivery, as determined the number of RODN positive nuclei observed per field (Figure 6a and Table 1). The number of fluorescent nuclei identified in a given visual field ranged from 0 to 210 RODN-positive nuclei per field, a much greater range than observed for i.v. administration of lipoplex (12-85 per field). However, fewer GFP-expressing cells were observed, at most two per field, than observed after i.v. administration (Figure 6a). In addition, there was no GFP expression associated with fields exhibiting greater than 102 RODN-positive nuclei per field (Figure 6c). The absence of gene expression in regions with large numbers of fluorescent nuclei was a compelling finding. This may have been due to a toxic effect from either the cationic lipid^26,27 or possibly from the large amount of polynucleotide delivered. The regions with greater than 102 fluorescent nuclei per field accounted for approximately 30% of the observed fluorescent nuclei, suggesting that a large proportion of the lipoplex dose was delivered to these cells but did not elicit gene expression.

A direct comparison of the oligonucleotide delivery and frequency of gene expression is made in Table 1. While the average number of fluorescent nuclei per field were similar for the two routes of administration, the frequency of GFP-expressing cells was six-fold greater with i.v. lipoplex administration (1.3 versus 0.2 GFP-positive cells per field). When total GFP expression was normalized for the number of fluorescent nuclei, the difference was nearly 10-fold (3.04 versus 0.36 GFP cells/100 fluorescent nuclei).

Free oligonucleotide distribution in the absence of cationic lipid

To control for nucleic acid transfer that was not due to the lipoplex, free oligonucleotide and GFP plasmid were co-administered without cationic lipid in the lung, via i.t. instillation. Four hours after i.t. administration, oligonucleotide fluorescence signal was widely distributed within the lung parenchyma and was prevalent within and at the surface of the cells (Figure 5e). At 24 h, oligonucleotide fluorescence was observed to be widely distributed within the lung tissue, but no GFP-positive cells were detected (Figure 5f). The intensity of the oligonucleotide nuclear fluorescence was less than when formulated with cationic lipid and delivered via the i.t. route (Figure 5c, d), but the distribution of the labeled nuclei was more homogenous. At higher magnification, fluorescent punctates appeared in the cytoplasm and near the surface of the cells, and in some cells RODN was observed within the nuclei. In other cells, the RODN are clearly not co-localized within the nucleus but in the periphery (Figure 5f).

Discussion

We have developed and validated a means to study the distribution, uptake and expression of cationic lipid/DNA lipoplex in the mouse lung. This technique permitted the preservation of lung structure such that confocal microscopy could be employed to follow the distribution and internalization of fluorescent lipoplex after both i.t. and i.v. distribution. Since the perfusion of fixative and inflation by agarose did not significantly remove the airway-instilled cationic lipoplex (Figure 1), we believed this technique was suitable for studying the fate of the lipoplex in the lung.

Previous studies have characterized the distribution of lipoplex using radioactive markers. After i.v. administration, approximately 80-90% of the lipoplex dose was typically retained in the lung within 5 min of injection, followed by a decline with a half-life of less than 1 h.^{10,11,28,29,30} Thus, the initial targeting of cationic lipoplex to the lung is sufficient for high levels of gene expression. In this study, we demonstrated that the spatial distribution of lipoplex, as indicated by the lipid marker, was homogenous throughout the lung parenchyma (Figure 2a). Thus, intravenous administration of cationic lipoplex appeared to be optimized for initially dispersing lipoplex throughout the lung.

The aggregates of lipoplex observed at higher magnification suggested the existence of specific binding domains at the endothelial surface (Figure 2). This pattern of binding had been previously described for other cationic particles within the lung capillaries, and agreed with previous observations using cationic lipoplexes.^12,13,18,31 The binding of lipoplex at the endothelial surface was attributed to electrostatic interactions with negatively charged molecules on the cell surface. Sulfated glycosaminoglycans, which are prevalent on the surface of capillary endothelium, have been previously shown to mediate lipoplex binding in vitro and in vivo.^4,6

A previous study from our laboratory demonstrated that the retention of lipoplex for approximately 60 to 90 min was important for high levels of expression.³² This finding suggested that the critical events for cellular uptake occurred within the initial 90 min. Other investigators have also noted that prolonged binding (half-lives) in the lung crudely correlates with subsequent gene expression levels.^10,11,28,33 Thus, to attain the highest levels of pulmonary gene expression with i.v. administered lipoplex, prolonging the duration of binding to the cell surface appears to be more critical than improving the spatial distribution of lipoplex.

By comparison, i.t. administration led to localized delivery of lipoplex within only some of the bronchioles and the immediately distal alveoli (Figure 3a). The distribution was even more disparate in the apical sections of the lung lobe (Figure 4). This localized distribution within the airways was also reported to occur in rats and hamsters after i.t. instillation of iron oxide particles,^34,35 which suggested that the pattern was common to the instillation procedure. We inferred from these image reconstructions that the aqueous delivery volume had become channeled into select bronchioles pathways.

In the absence of cationic lipid, DNA administered via the airways was easily displaced from the lung (Figure 1a) and was more widely distributed throughout the lung parenchyma (using fluorescent oligonucleotides as markers, Figure 5e and f). The formulation of plasmid DNA with cationic lipids clearly facilitated binding within the lung airway, but this binding occurred at the anionic surfaces encountered early within the distribution pathway, in the proximal regions of the lung. The velocity of administration, as well as the viscosity and surface effects of the lipoplex solution, may have been additional factors which affected the lipoplex distribution. Thus, for instilled lipoplex, the fluid flow into the first series of airway branches appeared to be an important determinant of the distribution of lipoplex within the lung.

We have proposed the use of fluorescent phosphorothioate oligonucleotides as a marker for cytoplasmic delivery of nucleic acids.²³ Phosphorothioate oligonucleotides are resistant to nuclease degradation, have a greater cytoplasmic diffusivity than plasmid DNA and will rapidly accumulate in the cell nucleus where they persist for up to 48 h.^22,24 Thus, when the fluorescent RODN were incorporated into the lipoplex, the number of fluorescent nuclei provided an estimate of the number of cells that had received nucleic acids, and represented the upper limit for the number of cells that could express a delivered plasmid. We believe that the fluorescent oligonucleotides are informative markers for assessing cytoplasmic release of lipoplex nucleotides and are a useful surrogate for estimating the cytoplasmic delivery of plasmid DNA.

Cell division has been shown to be an important factor leading to efficient gene transfer in vitro.^36,37 Unfortunately, we cannot assess the role of cell division upon in vivo DNA uptake, since nuclear localization of the phosphorothioate oligonucleotide would not differentiate between cells that had or had not yet divided. However, given the time-frame of these experiments (24 h) it is unlikely in the context of the turnover time for endothelial cells and epithelial cells that nuclear uptake of oligonucleotide was significantly influenced by cell division.

The pattern of fluorescent oligonucleotide delivery was similar to the localization of lipoplex: homogeneous for i.v. delivered lipoplex and concentrated around the bronchioles for i.t. delivered complex (Figure 5a, c). The difference in RODN fluorescence intensity at 24 h after i.v. administration (Figure 5a) compared with i.t. administration (Figure 5c) is due to the pharmacokinetics of lipoplex elimination from the lung. Only 5% of the lipoplex dose remains in the lung at 24 h after i.v. administration whereas 40% of the dose remains after i.t. administration.^10,38

With i.t. administration, a broad range in fluorescent oligonucleotide-nuclei localization (0 to 210 positive nuclei per field) demonstrated the great variability in i.t. gene delivery efficiency within the lung (Figure 6a, Table 1). The delivery expression histogram can be roughly divided into three domains: (1) visual fields with greater than 85 fluorescent nuclei per field, having high nucleotide delivery but little or no gene expression; (2) fields with 30-50 fluorescent nuclei per field, characterized by intermediate nucleotide delivery and effective gene expression; and (3) fields with both low nucleotide delivery and low gene expression expression. We interpreted the absence of gene expression at very high levels of DNA delivery as evidence of lipoplex toxicity due to high local concentrations of the lipoplex or its components. An alternative possibility is that the degradation-resistant RODN may have been internalized as the noncomplexed nucleotide, as seen in Figure 5e, giving a biased estimate of lipoplex-mediated delivery. Still, it is clear that the deposition of the lipoplex was concentrated around the bronchioles and the alveoli immediately distal to them. Therefore, one strategy to improving i.t. gene delivery and gene expression would be to achieve a more uniform delivery within the aiways, leading to lower toxicity and perhaps, more efficient gene expression.

Since airway administered lipoplex gene expression levels are lower than those of i.v. delivered lipoplex, we concluded that that the poor spatial distribution within the lung was a factor which limited the efficiency of i.t. gene delivery. Eastman and colleagues³⁹ have formulated cationic lipids with polyethylene glycol-conjugated lipid to reduce shear degradation of DNA upon aerosolization, and have demonstrated some gene expression activity in the lung.³⁸ While such charge-shielding formulation should have more homogeneity in lipoplex distribution in the lung, this has not yet been confirmed. Strategies such as this, which combine the gene transfer properties of cationic lipid with a better dispersing tactic, may be useful in attaining higher levels of gene transfer through the airway administration.

By comparing the morphology of nuclei stained by fluorescent oligonucleotide, the data presented in this report suggested that each route of administration targets a different population of cells (Figure 5b, d). The morphology of the fluorescent nuclei, after i.v. lipoplex, was consistent with internalization within vascular endothelial cells, in agreement with observations made by McLean and colleagues.¹² In contrast, the morphology of fluorescent nuclei after i.t. lipoplex administration suggested that type I and type II pneumocytes (by their rounded nuclei) were responsible for the majority of internalization. Expression of GFP was also consistent with these cell types (Figure 5d). These observations also agreed with other reports of i.t. lipoplex administration,^14,15,16 although some of these studies also observed expression in the bronchiolar epithelium.^14,17

These different cell types may have different propensities for trafficking and expression of the exogenous plasmid DNA. The data suggested that endothelial cells of the pulmonary capillaries had a greater frequency of gene expression, on average 3.04 GFP-expressing cells per 100 fluorescent nuclei, than did the epithelial cells (0.36 GFP cells/100 RODN-positive nuclei, see Table 1). Part of this 10-fold difference was attributed to the regions in the intratracheally dosed lung which had a high density of fluorescent nuclei but did not have corresponding GFP expression (Figure 6c, cut off at 85 positive nuclei per field). However, even when corrected for the nuclei found in these regions, a six-fold difference in expression/delivery could be still be demonstrated (0.525 GFP/100 RODN-positive nuclei). Since the nuclear delivery of RODN suggested equivalent cytoplasmic delivery, the higher level of expression may be due to a greater capacity for nuclear localization or to the gene expression capacity of the cell. Thus, pulmonary capillary endothelial cells appear to be more efficient at the nuclear transfer or gene expression process than the pulmonary epithelial cells.

While gene therapy to the airways is appealing for treating certain lung diseases such as cystic fibrosis and squamous cell carcinomas, these results suggest that gene expression after i.t. administration could still be relatively poor, even with improved delivery to the airway epithelium. Targeting the lung capillary endothelium via systemic administration will probably yield more promising results for gene therapy in the lung. Thus the diseases which can best be addressed using gene therapy to the lung would be those affecting the lung endothelium or those which can utilize the lung as a producing organ for the desired protein. In the case of secreted proteins, such as alpha-1 antitrypsin, for which the specificity of the expressing cell is less important, the lung remains a good target organ for gene therapy.⁷

In conclusion, the differences in gene expression levels between lipoplexes which are administered by the i.v. and i.t. routes can be attributed to a combination of non-uniform distribution of lipoplex observed with i.t. administration, and to the greater ability of endothelial cells to traffic and express the delivered gene.

Materials and methods

Reagents

1,2-Dioleoyl-3-trimethylammonium-propane (DOTAP), L--phosphatidylethanolamine-N-fluorescein (F-PE), and 1,2-dioleoyl-sn-glycero-3-phosphatidylethanolamine-N-lissamine rhodamine B sulfonyl (N-Rh-PE) were purchased from Avanti Polar Lipids (Alabaster, AL, USA). Cholesterol, Hepes, phosphate-buffered saline (PBS), glucose, formalin, and paraformaldehyde were purchased from Sigma (St Louis, MO, USA). Vectashield mounting medium and fluorescein isothiocyanate (FITC)-Lycopersicon esculentum (tomato) lectin were purchased from Vector Laboratories (Burlingame, CA, USA). Texas Red-labeled Wheat germ agglutinin (WGA) was purchased from EY Laboratories (San Mateo, CA, USA). Low melting point agarose was purchased from Gibco BRL Products, Life Technologies (Grand Island, NY, USA). Anesthetics used were Ketamine HCl (100 mg/ml; Abbott Laboratories, Abbott Park, IL, USA), Xylazine HCl (100 mg/ml; Fermenta Veterinary Products, Kansas City, MO, USA), Acepromazine Maleate (10 mg/ml; Boehringer Ingelheim Vetmedica, St Joseph, MO, USA), Forane (isofluorane for inhalation, Baxter Healthcare Corporation, Deerfield, IL, USA).

Oligonucleotide and plasmid DNA

The phosphorothioate oligonucleotide used in these studies was purchased from Oligos Etc (Wilsonville, OR, USA). A 15-mer phosphorothioate oligonucleotide (5'-TCA/TAC/TGT/TGA/GCA-3') was synthesized and conjugated at the 5' end with a C6 amino linker and labeled at the 3' end with rhodamine. The amine terminus permitted the oligonucleotide to be fixed in the cell and prevented artifactual re-localization of oligo in the nucleus after formalin treatment. High purity plasmid DNA, containing less than 5 microU of endotoxin per mg DNA (assayed by Limulus Amebocyte Lysate Assay kit; BioWhittaker, Walkersville, MD, USA) was a generous gift of Valentis (Burlingame, CA, USA). The plasmid expression vectors utilized in our studies included pCMV-GFP, pCMVCAT, and pCMVLuc. In some studies, a luciferase containing plasmid under control of the CMV promoter and bGH polyadenylation sequence from Valentis was used. Radiolabeled DNA was made by incorporating [125]-I-dCTP into plasmid DNA by nick translation followed by ligation and ethanol precipitation (for details see 'Validation of inflation protocol using radiolabeled complex'). For simultaneous RODN and GFP expression studies, fluorescent oligo and plasmid were premixed in 5 mM Hepes (pH 7.4) at a ratio of 5 g RODN to 30 g DNA per mouse.

Cationic lipid/DNA complex preparation

Cationic liposomes, DOTAP, DOTAP:DOPE (10:9 molar ratio), DOTAP:cholesterol (10:9 mol ratio) were prepared as previously reported.¹⁰ Briefly, the lipids were mixed in chloroform, evaporated to dryness on a rotary evaporator and placed under high vacuum for at least 3 h. Fluorescent 1,2-dioleoyl-sn-glycero-3-phosphatidylethanolamine-N-lissamine rhodamine B sulfonyl (N-Rh-PE) was incorporated at 0.5 mole percent to the lipids before rotary evaporator drying. The lipid film was hydrated with 5 mM Hepes (pH 7.4) for i.t. administration, and then briefly agitated using a vortex mixer. Small unilamellar vesicles were formed by extruding five times through a 200 nm polycarbonate membrane (Nucleopore Corporation, Pleasanton, CA, USA) using a hand held extrusion device (Avestin, Ottawa, Canada). The plasmid DNA, also in 5 mM Hepes, was added to the liposomes in equal volumes, to give a charge ratio of 3 to 1 (positive to negative), based on the number of DOTAP amines and DNA phosphates. For i.v. administration, lipids were hydrated with 10 mM Hepes (pH 7.4) in a 10% glucose solution and also extruded five times through a 200 nm membrane. Plasmid DNA, prepared in sterile water, was added to the liposomes in an equal volume to form 5 to 1 (+/-) charge ratio complexes. The lipoplex was allowed to stand at room temperature for at least 10 min before administration.

In vivo administration of lipoplex

Female CD-1 (3-6 week old) mice were purchased from Charles River Laboratories (Hollister, CA, USA). All animals were studied in accordance with guidelines established by the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals, and with the approval of the Committee on Animal Research at the University of California, San Francisco. Mice receiving lipoplex i.v. were injected in the tail vein with 100 l of lipoplex in 5 mM Hepes/5% glucose. Mice receiving lipoplex via i.t. administration were first anesthetized with ketamine (65 mg/kg), acepromazine (0.65 mg/kg), and xylazine (3.3 mg/kg) followed by inhalation with isoflurane. The trachea was visualized using a midline incision, and the mice were intubated using a blunted 18G needle, and threaded with a PE-10 tubing which extended 1-2 mm past the tip of the blunted needle.³⁸ A 0.5 cc syringe, connected to the PE-10 tubing, was used to dispense the lipoplex in 100 l doses directly into the airways. After administration, the tubing was removed and an empty syringe was connected to the blunted needle and used to deliver gently 0.3 ml of air to disperse the dose throughout the lung.

Agarose inflation of the mouse lung

Inflation of the mouse lung was achieved by a protocol developed by McLean and colleagues.¹² At the time of death, mice were anesthetized with an intraperitoneal injection of ketamine (200 mg/kg), acepromazine (2 mg/kg), and xylazine (10 mg/kg) followed by inhalation with isoflurane. A midline incision was reopened to visualize the trachea, followed by intubation with an 18G blunted needle. To label the airway epithelium fluorescently, a 1.0 ml solution of FITC-Lycopersicon esculentum lectin (0.3 mg/ml in PBS) or of Texas Red-Wheat germ agglutinin (0.3 mg/ml in PBS) was directly instilled into the mouse lung using a 3 cc syringe, followed by a 2.5 ml 'wash out' with 10% formalin in PBS (using a 3 cc syringe). All solutions were warmed to 42°C before perfusion. The mouse thoracic cavity was cut open, exposing the lungs and heart. Excess perfusate was permitted to drain back out of the 18G needle before instilling 0.8 ml of 3% low melting point agarose in PBS, also warmed to 42°C. Warm PBS was used as a lavage in the open chest cavity and was important to keep agarose solution flowing into the entire lung. An ice-cold PBS lavage was then used to solidify the agarose. The lungs and trachea were removed, en bloc, and stored in fixative (10% formalin in PBS) until tissue sectioning. Lung lobes were dissected and set in 7% low melting point agarose, and subsequently sectioned at 200 microns with a Vibratome 1000 (Technical Products International, St Louis, MO, USA).

Lectin staining of the lung epithelium and endothelium

FITC-Lycopersicon esculentum lectin was solubilized in phosphate-buffered saline (0.3 mg/ml), and injected as a 100 l bolus into the mouse tail vein. After 5 min, the mouse was anesthesized and intubated as described above. Texas Red-Wheat germ agglutinin (0.3 mg/ml in PBS, 1.0 ml) was perfused into the airways, using a 3.0 cc syringe. A 2.5 ml volume of 10% formalin was subsequently perfused to wash out the airways, and the chest cavity was opened. All perfusates were warmed to 42°C. At this point, the mouse was exsanguinated and then infused, through the right ventricle, with 1.0 ml of 10% formalin. The lung was then inflated with 3% low melting point agarose, as described above.

Validation of inflation protocol using radiolabeled complex

To determine the retention of plasmid DNA and cationic liposomes following i.t. inflation, [125]-I labeled DNA and [125]-I labeled cationic liposomes were utilized. Iodinated para-hydroxybenzamidine-phosphatidylethanolamine (BPE) was included in the standard preparation of DOTAP/Chol liposomes (3:1 charge ratio); approximately 200 000 DPM of [125]-I BPE or 200 000 DPM of [125]-I DNA were administered per mouse. BPE was radiolabeled in our laboratory as previously described and has been demonstrated to be an effective marker for distribution studies.^8,10,40 Intact radiolabeled plasmid DNA was prepared by nick translating 100 g of plasmid DNA with [125]-I-dCTP (NEN Research Products, Boston, MA, USA), 20 pg of DNase I (Gibco-BRL, Gaithersburg, MD,USA), and 5 units of DNA polymerase (Gibco-BRL). Unincorporated nucleotides were separated from labeled plasmid DNA by spin column chromatography, and labeled DNA was incubated with T4 ligase overnight at 16°C. Twenty micrograms of unlabeled plasmid were added to 200 000 DPM of labeled plasmid. Plasmid DNA alone or complexed at a 3:1 lipid/DNA charge ratio with DOTAP:cholesterol liposomes was then administered intratracheally to mice; 3 h after complex administration, lungs were inflated and levels of radioactivity measured. Levels of radioactivity were determined in the lung following gamma counting (Beckman Gamma 8000, Palo Alto, CA, USA).

Confocal microscopy

Images were obtained on a BioRad laser scanning confocal microscope (MRC-1024, with a Nikon Diaphot 200 inverted microscope), using LaserSharpe software. Simultaneous scans for Texas Red/rhodamine and FITC/GFP detection were made using a laser intensity of 3%, with a gain set to '1486' and black level of -3, per software settings. The confocal iris was fixed at '2.2' (maximum of 8, per LaserSharpe). Reconstructions of the entire lobe sections were obtained using a 10 ´ Nikon objective (Plan 10X, air objective NA = 0.30), with merged projections of 200 micron sections. The 512 ´ 512 pixel images were then imported into a paint/draw program (Corel Photo-Paint 4.0; Corel Corporation, Ottawa, Ontario, Canada) and visually pieced together. Higher magnification images were obtained using either 400 ´ (Nikon Fluor Plan 40X, NA = 1.30) or 600 ´ (Nikon Plan 60X, NA = 1.40) magnification, as noted.

Acknowledgements

We gratefully acknowledge financial support from the following agencies: NIH DK-46052-05 (FCS), The Cystic Fibrosis Foundation UCSF RDP, and the California Statewide Tobacco RDP 6RT-0109. Dr Szoka has a financial interest in and serves as a consultant to Valentis, Inc., a biotechnology company developing gene medicines. We thank Donald McDonald for advice and access to a vibrotome for use in this project.

1 Felgner PL et al. Nomenclature for synthetic gene delivery systems. Hum Gene Ther 1997; 8: 511-512, MEDLINE

2 Van der Woude I et al. Parameters influencing the introduction of plasmid DNA into cells by the use of synthetic amphiphiles as a carrier system. Biochim Biophys Acta 1995; 1240: 34-40, MEDLINE

3 Eastman SJ et al. Optimization of formulations and conditions for the aerosol delivery of functional cationic lipid: DNA complexes. Hum Gene Ther 1997; 8: 313-322, MEDLINE

4 Mislick KA, Baldeschwieler JD. Evidence for the role of proteoglycans in cation-mediated gene transfer. Proc Natl Acad Sci USA 1996; 93: 12349-12354, MEDLINE

5 Belting M, Petersson P. Intracellular accumulation of secreted proteoglycans inhibits cationic lipid-mediated gene transfer. Co-transfer of glycosaminoglycans to the nucleus. J Biol Chem 1999; 274: 19375-19382, MEDLINE

6 Mounkes LC et al. Proteoglycans mediate cationic liposome-DNA complexed-based gene delivery in vitro and in vivo. J Biol Chem 1998; 273: 26164-26170, MEDLINE

7 Brigham KL et al. In vivo transfection of murine lungs with a functioning prokaryotic gene using a liposome vehicle. Am J Med Sci 1989; 298: 278-281, MEDLINE

8 Meyer KB et al. Intratracheal gene delivery to the mouse airway: characterization of plasmid DNA expression and pharmacokinetics. Gene Therapy 1995; 2: 450-460, MEDLINE

9 Zhu N, Liggitt D, Liu Y, Debs R. Systemic gene expression after intravenous DNA delivery into adult mice. Science 1993; 261: 209-211, MEDLINE

10 Barron LG, Meyer KB, Szoka FC. Effects of complement depletion on the pharmacokinetics and gene delivery mediated by cationic lipid-DNA complexes. Hum Gene Ther 1998; 9: 315-323, MEDLINE

11 Mahato RI et al. Physicochemical and pharmacokinetic characteristics of plasmid DNA/cationic liposome complexes. J Pharm Sci 1995; 84: 1267-1271, MEDLINE

12 McLean JW et al. Organ-specific endothelial cell uptake of cationic liposome-DNA complexes in mice. Am J Physiol 1997; 273: H387-404. MEDLINE

13 Deshpande D et al. Target specific optimization of cationic lipid-based systems for pulmonary gene therapy. Pharm Res 1998; 15: 1340-1347, MEDLINE

14 Griesenbach U et al. Comparison between intratracheal and intravenous administration of liposome-DNA complexes for cystic fibrosis lung gene therapy. Gene Therapy 1998; 5: 181-188, MEDLINE

15 Gorman CM et al. Efficient in vivo delivery of DNA to pulmonary cells using the novel lipid EDMPC. Gene Therapy 1997; 4: 983-992, MEDLINE

16 Blezinger P et al. Intratracheal administration of Interleukin-12 plasmid-cationic lipid complexes inhibits murine lung metasteses. Hum Gene Ther 1999; 10: 723-731, MEDLINE

17 Oudrhiri N et al. Gene transfer by guanidinium-cholesterol cationic lipids into airway epithelial cells in vitro and in vivo. Proc Natl Acad Sci USA 1997; 94: 1651-1656, MEDLINE

18 Li S et al. Dynamic changes in the characteristics of cationic lipidic vectors after exposure to mouse serum: implications for intravenous lipofection. Gene Ther 1999; 6: 585-594, MEDLINE

19 Stribling R et al. Aerosol gene delivery in vivo. Proc Natl Acad Sci USA 1992; 89: 11277-11281, MEDLINE

20 Scheule RK et al. Basis of pulmonary toxicity associated with cationic lipid-mediated gene transfer to the mammalian lung. Hum Gene Ther 1997; 8: 689-707, MEDLINE

21 McDonald RJ et al. Aerosol delivery of lipid: DNA complexes to lungs of rhesus monkeys. Pharm Res 1998; 15: 671-679, MEDLINE

22 Chin DJ et al. Rapid nuclear accumulation of injected oligodeoxyribonucleotides. New Biol 1990; 2: 1091-1100, MEDLINE

23 Zelphati O, Szoka FC. Mechanism of oligonucleotide release from cationic liposomes. Proc Natl Acad Sci USA 1996; 93: 11493-11498, MEDLINE

24 Lukacs G et al. Size-dependent DNA mobility in cytoplasm and nucleus. J Biol Chem 2000; 275: 1625-1629, Article MEDLINE

25 Li S, Huang L. In vivo gene transfer via intravenous administration of cationic lipid-protamine-DNA (LPD) complexes. Gene Therapy 1997; 4: 891-900, MEDLINE

26 Smith JG et al. Characterization and in vivo testing of a heterogeneous cationic lipid-DNA formulation. Pharm Res 1998; 15: 1356-1363, MEDLINE

27 Xu Y , Hui S-W, Frederick P, Szoka FC. Physicochemical characterization and purification of cationic lipoplexes. Biophys J 1999; 77: 341-353, MEDLINE

28 Liu Y et al. Factors influencing the efficiency of cationic liposome-mediated intravenous gene delivery. Nat Biotechnol 1997; 15: 167-173, MEDLINE

29 Hofland HE et al. In vivo gene transfer by intravenous administration of stable cationic lipid/DNA complex. Pharm Res 1997; 14: 742-749, MEDLINE

30 Mahato RI et al. Biodistribution and gene expression of lipid/plasmid complexes after systemic administration. Hum Gene Ther 1998; 9: 2083-2099, MEDLINE

31 Simionescu M, Simionescu N. Endothelial surface domains in pulmonary alveolar capillaries. In: Ryan US (ed.). Pulmonary Endothelium in Health and Disease. Marcel Dekker: New York, 1987, pp 35-62.

32 Barron LG, Gagne L, Szoka FC. Lipoplex-mediated gene delivery to the lung occurs within 60 minutes of intravenous administration. Hum Gene Ther 1999; 10: 1683-1694, Article MEDLINE

33 Templeton NS et al. Improved DNA:liposome complexes for increased systemic delivery and gene expression. Nat Biotechnol 1997; 15: 647-652, MEDLINE

34 Brain JD, Knudson DE, Sorokin SP, Davis MA. Pulmonary distribution of particles given by intratracheal instillation or by aerosol inhalation. Environ Res 1976; 11: 13-33, MEDLINE

35 Schanker LS, Mitchell EW, Brown RAJ. Species comparison of drug absorption from the lung after aerosol inhalation or intratracheal injection. Drug Metab Dispos 1986; 14: 79-88, MEDLINE

36 Coonrod A, Li F-Q, Horwitz M. On the mechanism of DNA transfection: efficient gene transfer without viruses. Gene Therapy 1997; 4: 1313-1321, MEDLINE

37 Zelphati O, Liang X, Hobart P, Felgner PL. Gene chemistry: functionally and conformationally intact fluorescent plasmid DNA. Hum Gene Ther 1999; 10: 15-24, Article MEDLINE

38 Meyer KEB, Thompson MM, Barron LG, Szoka FC. Cationic liposome formulations containing cholesterol increase plasmid persistence and gene expression in the mouse lung after intratracheal administration. J Liposome Res (submitted)

39 Eastman SJ et al. A concentrated and stable aerosol formulation of cationic lipid: DNA complexes giving high-level gene expression in mouse lung. Hum Gene Ther 1997; 8: 765-773, MEDLINE

40 Abra RM, Schreier H, Szoka FC. The use of a new radioactive-iodine lipid marker to follow in vivo disposition of liposomes: comparison with an encapsulated aqueous space marker. Res Commun Chem Pathol Pharmacol 1982; 37: 199-213, MEDLINE

Figure 1 Validation of the lung inflation and staining protocols. (a) Retention of intratracheal lipoplex components 3 h after instillation. The percentage of radioactive counts remaining after lung perfusion with counterstain solution, PBS, and agarose inflation are shown. Comparison of lipoplex component removal after intratracheal (i.t.) perfusion, no perfusion or inflation (Control), intracardiac inflation (IC); anionic liposomes are shown as a non-cationic control for washing out unbound material. (b) Intravascular (red) and intratracheal (green) staining using fluorescent lectins.

Figure 2 Distribution of lipoplex, 5 min after i.v. injection. (a) Reconstructed view of the right superior lobe, a transverse section. Rhodamine-PE (red) indicates distribution of the lipoplex against a FITC-lectin (green) counterstain for the lung epithelium. The lectin was introduced via tail vein injection just before death. (b) Original magnification at 600 ´, showing red lipoplex uptake in endothelium.

Figure 3 Distribution of lipoplex, 3 h after i.t. administration. (a) Reconstructed view of the right superior lobe, transverse section. Rhodamine-PE (red) indicates lipoplex distribution with a FITC-lectin (green) counterstain. The lectin was introduced via airway perfusion as described in the Materials and methods section. (b) An example of the digitized fields used to reconstruct the lobe image; original magnification at 100 ´. (c) Original 400 ´ magnification of lipoplex on epithelial surface after i.t. administration.

Figure 4 Distribution of lipoplex in sequential lung sections, Nos 7 to 12, reconstructed from digital images obtained from low magnification Bio-Rad confocal MRC-1024. Transverse sections were cut to 200 microns in thickness and numbered, starting from rostral to caudal ends of the right superior lung. Rhodamine (red) lipid localizes the lipoplex deposition, and FITC (green) lectin counter-stains the entire epithelial surface.

Figure 5 Simultaneous fluorescent oligonucleotide (red) localization and GFP (green) expression in the mouse lung after i.v. lipoplex administration (a, b), i.t. lipoplex administration (c, d) and i.t. administration without cationic lipid (e, f). Low magnification (a, c, e at 100 ´) and high magnification (b at 400 ´, d and f at 600 ´). GFP expression shown by arrow heads. In panels a and b a fluorescein-PE marker was also used and can be distinguished as green punctates, distinct from GFP expression (arrow heads). The green reticular fluorescence observed in panels b and f is due to autofluorescence that is visualized at the gain used to image the section. No fluorescent lectin was used in these studies.

Figure 6 Quantification of oligonucleotide-positive nuclei and GFP-expressing cells in the mouse lung 24 h after i.v. and i.t. administration. (a) Quantification of the number of RODN fluorescent nuclei per 400 ´ visual field. (b) The distribution of visualized fields based upon the number of GFP-expressing cells viewed per field. (c) Composite distribution showing both the number of RODN fluorescent nuclei and GFP-positive cells viewed in each field.

Table 1 Quantification of oligonucleotide nuclear localization and GFP expression; quantified per visual field