INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

Clinical studies concerning cystic fibrosis (CF) gene therapy showed, although promising results in vitro, only a limited efficacy (1). In these studies the cystic fibrosis transmembrane conductance regulator genes, either complexed to cationic liposomes (lipoplexes) (Figure 1) or inserted in the genome of cripple viruses, are nebulized or instilled into the airways of patients with CF. The reason for this discrepancy must be sought in the additional extracellular and intracellular barriers present in vivo. First, it is known that differentiated cells, as present in the airways of patients with CF, are more resistant to gene transfer than nondifferentiated cells used in most in vitro transfection experiments. This is probably due to differences in cell surface properties and endocytosis activity (2, 3). Second, patients with CF have a lot of mucus in their lungs, which covers the target cells. For CF gene therapy to be effective, gene complexes will first have to overcome this extracellular barrier; they have to remain intact upon permeating the CF mucus and cross the mucus layer before being cleared from the lung by the mucociliary transport. In a previous study we showed that CF sputum (sputum = expectorated respiratory mucus) dramatically retards the movement of particles having a size comparable with lipoplexes. For the largest particles (560 nm), a complete blockade was actually observed (4). Compared with normal airway secretions, CF mucus contains high amounts of DNA, mucin, albumin, phospholipids, and inflammatory products (5, 6). As the (cationic) lipoplexes and the negatively charged molecules present in CF mucus may interact electrostatically, destabilization of the lipoplexes may occur.

View larger version (38K): [in this window] [in a new window]   Figure 1.   Hypothetical schematic representation of the interaction between cationic (DOTAP:DOPE) liposomes and pDNA leading to the encapsulation of pDNA between a lipid double layer (A). Small amounts of linear DNA added to the lipoplexes also become surrounded by a lipid double layer (up to a linear DNA:pDNA concentration ratio of 1) (B). At higher linear DNA:pDNA concentration ratios (± charge ratio < 2) linear DNA binds to the surface of the lipoplexes resulting in a decrease in surface charge and the formation of large aggregates (C ). At concentration ratios above 1.6, small negatively charged particles are formed again because of linear DNA chains which form an anionic shield around the lipoplex (D). Above a concentration ratio of 1.6 the lipids in the outer double layer probably also start to rearrange themselves around the surface-bound linear DNA chains. This may lead to the destruction of the double layer, which may result in the liberation of pDNA from the lipoplexes.

Although clinical trials have been undertaken with lipoplexes, their stability in the presence of CF mucus has never been studied. Lipoplexes are built up by cationic lipids, which tightly bind the "therapeutic" plasmid DNA (pDNA) by electrostatic interactions. It is quite reasonable that "nontherapeutic" DNA, which is abundantly present in CF mucus and originates mainly from the lysis of leukocytes and epithelial cells (7, 8), will interact with the cationic lipoplexes, leading to charge neutralization and release of the "therapeutic" pDNA from the lipoplexes. For example, it has been shown that the addition of serum or albumin to lipoplexes may neutralize their positive charges, resulting in aggregation of the lipoplexes and release of the pDNA (9). Because of charge neutralization the gene complexes may show less binding to the negatively charged cell surface, which, consequently, may decrease gene transfer. Other studies have shown that pulmonary surfactants, which contain negatively charged phospholipids, can inhibit cationic liposome-mediated gene delivery (10). Ernst and coworkers attributed this to charge neutralization which results in release of pDNA from its carrier (10). Furthermore, Xu and Szoka showed that anionic lipids are also able to disassemble lipoplexes (13). All these reports are very interesting with regard to the development of more efficient gene delivery vehicles. However, most of them lack clinical relevance with regard to CF.

To support the clinical research on CF gene therapy, we examined to what extent CF mucus components (DNA, mucin, albumin, and phospholipids) and native CF sputa destruct lipoplexes. To perform the experiments under conditions as close as possible to the in vivo situation, we determined, for each CF mucus component investigated, the clinical concentration ratio which indicates the concentration of a particular mucus component to the concentration of "therapeutic" pDNA expected in the CF lung after lipoplex nebulization. To calculate the clinical concentration ratios we analytically determined the average concentration of mucin, DNA, albumin, and phospholipids in CF sputa. The concentration of the "therapeutic" pDNA expected in the mucus was estimated from the mass of "therapeutic" pDNA administered to the lungs of patients with CF in the clinical trial by Alton and coworkers (14).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

Chemicals

Mucin (from bovine submaxillary glands), linear double-stranded DNA (from salmon testes, indicated as linear DNA in this study), dipalmitoylglycerophosphocholine (DPPC), dipalmitoylglycerophosphoglycerol (DPPG), albumin (from bovine serum), 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (Hepes), and tris(hydroxymethyl)methylamine (Tris) were purchased from Sigma (Bornem, Belgium). Dioleoylglycerophosphoethanolamine (DOPE) and 1,2-dioleoyl-3-trimethylammoniumpropane (DOTAP) were purchased from Avanti Polar Lipids (Alabaster, AL). The Maxi Prep kit was obtained from Qiagen (Leusden, The Netherlands). Sybr Green I nucleic acid gel stain was from Molecular Probes (Leiden, The Netherlands). Pvu II restriction enzyme, Concert Gel Extraction System kit, BioNick Labeling System kit, Concert Rapid PCR Purification System kit, Blugene Nonradioactive Nucleic Acid Detection kit, and electrophoresis grade agarose were obtained from GibcoBRL, Life Technologies (Merelbeke, Belgium). Alveofact, a dispersion of pulmonary surfactants from bovine lung lavage, was a gift from Boehringer Ingelheim (Brussels, Belgium). It contains 54 mg surfactants/ml from which 81.7% were phosphatidylcholine and 11.5% phosphatidylglycerol. Pulmozyme ampules, containing 1 mg of recombinant human deoxyribonuclease I (rhDNase I E.C.3. 1.21.1) per ml, were a gift from Roche (Brussels, Belgium).

Buffers

Hepes buffer (20 mM Hepes pH 7.4); dithiothreitol (DTT) buffer (100 mM Na₂HPO₄ pH 7.0); sputum buffer (85 mM Na⁺, 75 mM Cl, 3 mM Ca²⁺, and 20 mM Hepes pH 7.4); electrophoresis buffer (EB) (10.8 g/L Tris base, 5.5 g/L boric acid, and 0.85 g/L ethylenediaminetetraacetic acid); loading buffer (50% weight to volume [wt:vol] sucrose in sputum buffer); denaturation buffer (1.5 M NaCl and 0.5 M NaOH); standard saline citrate (SSC) buffer (3 M NaCl, 0.3 M sodium citrate pH 7.0); Denhardt's solution (1% [wt:vol] ficoll, 1% [wt:vol] polyvinylpyrrolidone, and 1% [wt:vol] bovine serum albumin); prehybridization buffer (500 ml formamide, 250 ml SSC buffer, 100 ml Denhardt's solution, 500 mg heat-denatured salmon testes DNA, 3.1 g Na₂HPO₄, and 150 ml distilled water pH 6.5); hybridization buffer (450 ml formamide, 250 ml SSC buffer, 20 ml Denhardt's solution, 200 mg heat-denatured salmon testes DNA, 50 g dextran sulfate, 3.1 g Na₂HPO₄, and 280 ml distilled water pH 6.5); wash buffer 1 (0.1% [wt:vol] of sodium dodecyl sulfate [SDS] in 10-fold diluted SSC buffer); wash buffer 2 (0.1% [wt:vol] SDS in 100-fold diluted SSC buffer); wash buffer 3 (0.1% [wt:vol] SDS in 125-fold diluted SSC).

Preparation and Purification of Plasmid DNA

The pDNA consisted of 5,803 base pairs (bp) and contained as reporter gene secretory alkaline phosphatase under the control of the simian virus 40 promotor. The pDNA was amplified in Escherichia coli. For the isolation and purification of the pDNA from the bacteria by alkaline lysis, the Maxi Prep kit from Qiagen was used. After the final isopropanol precipitation the pDNA was dissolved in 10 mM Tris buffer at pH 8.0. The pDNA concentration was set at 1.0 mg/ml assuming that the absorption of 50 µg DNA/ml at 260 nm equals one. The purity of the plasmid DNA was controlled by determining the 260 nm/280 nm absorption ratio. The pDNA showed a high purity as this ratio was always between 1.8 and 2.0.

Preparation of Cationic Liposomes

Cationic liposomes containing DOTAP and DOPE (Figure 1) in a 1:1 molar ratio were prepared as follows. DOTAP and DOPE were dissolved in a 1:1 (volume) mixture of chloroform:methanol. Consequently, the solution of lipids was placed in a round-bottomed flask and the solvents were evaporated under vacuum at 30° C for approximately 30 min. The resulting lipid film was further dried under a flow of nitrogen for 1 h. The lipids were then resuspended in Hepes buffer (final concentration of 5 mM DOTAP and DOPE) and rehydrated overnight at 4° C to form cationic liposomes. The following day the liposomes were extruded at room temperature through a polycarbonate membrane, with a pore size of 0.1 µm, using the Avanti Polar Lipids Mini-extruder. The size and zeta potential () of the resulting cationic liposomes were routinely checked and equaled respectively 124 ± 9 nm and +50 ± 2 mV.

Preparation of Lipoplexes

Plasmid DNA was first diluted in Hepes buffer to a concentration of 0.41 mg/ml. Subsequently, the diluted pDNA was added to an equal volume of cationic liposomes (5 mM DOTAP) resulting in a final ± charge ratio of 4. Immediately after the addition of pDNA to the cationic liposomes, Hepes buffer was added until the final concentration of pDNA in the system was 0.126 mg/ml. This mixture was then vortexed and incubated at room temperature for 30 min. This resulted in the formation of lipoplexes. The size and  of the lipoplexes were routinely determined and equaled respectively 279 ± 14 nm and +50 ± 2 mV.

Determination of the Concentration of Electrolytes, DNA, Mucin, Phospholipids, and Proteins in CF Sputum

Approval for the collection of CF sputum was obtained from the ethics committee of the University Hospital of Ghent. The sputum was expectorated spontaneously by patients with CF during chest physiotherapy sessions in the CF center Zeepreventorium in De Haan, Belgium. The mucin and DNA concentrations in the sputa were determined using a protocol previously described (4). For the determination of the phospholipid, protein, Na⁺, Cl, and total calcium concentrations, we mixed 150 µl of CF sputum with 150 µl DTT solution (6 mM DTT dissolved in DTT buffer). DTT liquefies the sputum by cleaving the intramolecular disulfide bridges between the mucin subunits and therefore facilitates the diluting of the sputa. These mixtures were incubated at room temperature for 1 h. As a control we also mixed 150 µl of double distilled water with 150 µl of the DTT solution. The phospholipid, protein, Na⁺, Cl, and total calcium concentrations in the control solution were used as blank values in the determination of the concentrations in CF sputa. These measurements were performed by means of an automatic analyzer (Hitachi 747; Hitachi, Tokyo, Japan). The protein concentration was determined colorimetrically based on the biuret reaction (15). The concentrations of Na⁺ and Cl in the mixtures was measured using ion-selective electrodes. The total calcium concentration was determined colorimetrically after complexation with o-cresol-phthalein (16). An enzymatic assay was used to measure the phosphatidylcholine concentration (17). From this concentration and the phosphatidylcholine to phosphatidylglycerol concentration ratio present in Alveofact we calculated the phosphatidylglycerol concentration.

Besides native CF sputum samples, we also made use of synthetic sputum (SS). To prepare the SS, 115 mg mucin, 21 mg linear DNA (from salmon testes), 125 mg albumin (from bovine), and 300 µl Alveofact were dissolved in 5.0 ml sputum buffer (for composition, see RESULTS) and mixed overnight at 4° C. We used albumin as a model protein because this is the major protein in CF sputum (18). Salmon testes DNA was used to represent sputum DNA. As we did not dispose of CF mucins, we made use of mucins from bovine submaxillary glands.

Determination of the Clinical Concentration Ratios

Lipoplexes nebulized into the lungs of patients with CF will encounter the CF mucus layer and may interact with the different mucus components. To consider the interactions of lipoplexes with CF mucus components as close as those occurring in vivo, we determined for each CF mucus component the clinical concentration ratio: clinical concentration ratio = (average concentration of the component in the mucus after nebulization [mg/ml])/(concentration of pDNA (complexed to cationic liposomes) expected in the mucus after nebulization [mg/ml]).

The concentration of pDNA in the mucus after nebulization depends on the amount of pDNA delivered to the mucus-covered airways, the aerosol volume, and the volume of mucus present in the CF lung. Estimating the volume of CF mucus in the lung is difficult. On the basis of a publication of Lopez-Vidriero (19) we assumed a value of 20 ml of CF mucus in the calculations of the clinical concentration ratios. The total mass of nebulized pDNA and aerosol volume in the clinical trial of Alton and coworkers (14) was 34 mg and 16 ml, respectively. In the calculation of the clinical concentration ratios, we assumed that all of the pDNA was deposited on the mucus-covered airways.

Agarose Gel Electrophoresis Experiments

Lipoplexes (50 µl) were incubated for 30 min at room temperature, with solutions (50 µl) containing rising concentrations of a sputum component dissolved in sputum buffer. This resulted in different concentration ratios of sputum component to pDNA. One of these ratios was the clinical concentration ratio.

To study the release of pDNA from the cationic liposomes by CF sputum, 50 µl of lipoplexes was incubated (also for 30 min) with 50 µl 13.5-fold in sputum-buffer-diluted sputum. As the sputum did not spontaneously mix with the buffer, we first incubated the sputa with an equal volume of 6 mM DTT dissolved in sputum buffer. After 1 h incubation at room temperature the sputa were further diluted with sputum buffer to a final 13.5-fold dilution.

After incubating the lipoplexes with a single CF sputum component or with the diluted sputum, we added 5 µl of loading buffer to 45 µl of the lipoplex/sputum (component) mix. Subsequently, 20 µl of the resulting mix was loaded into the slots of a 1.1% [wt:vol] agarose gel prepared in EB. Also free pDNA and lipoplexes incubated with sputum buffer were loaded on the gel. Over the gel, which was submerged in EB, we set a voltage of 100 V for 1 h. After this time span the gel was transferred into a staining bath containing a solution of Sybr Green I in EB at pH 8.0. After 1 h the gel was illuminated with ultraviolet light and photographed.

Southern Blotting Experiments

A biotinylated probe was made from the pDNA as follows. A 704-bp-long sequence was cut out of the pDNA using the Pvu II restriction enzyme and separated by gel electrophoresis. The 704 bp sequence was cut out of the gel and purified using the Concert Gel Extraction System-kit. This DNA sequence was then biotinylated using nick translation (BioNick Labeling System kit). After biotinylation, unincorporated (biotinylated) nucleotides and enzymes were separated from the labeled DNA by the Concert Rapid PCR Purification System kit.

After agarose gel electrophoresis on the lipoplex/sputum or lipoplex/linear DNA mixtures, the nucleic acids in the gel were made single-stranded by incubating the gel in denaturation buffer for a controlled period of time. The nucleic acids on the agarose gel were first transferred to a nitrocellulose membrane overnight, to make a precise replica of the gel. The next day the membrane was removed and the DNA was fixed on the membrane by placing it in a vacuum oven for 2 h at 80° C. In a next step the membrane was prehybridized in prehybridization buffer. After removal of the prehybridization buffer the membrane was incubated with the biotinylated single-stranded probe, dissolved in hybridization buffer, overnight at 42° C. The next day, unbound probes were washed off from the membrane using wash buffers 1, 2, and 3. To visualize the hybridized probes we used the BluGene Nonradioactive Nucleic Acid Detection kit.

Size and Zeta Potential

The size and  of the lipoplexes were measured using respectively dynamic light scattering (Autosizer 4700; Malvern, Worcestershire, UK) and particle electrophoresis (Zetasizer 2000; Malvern). The influence of CF sputum components (DNA, DPPG, DPPC, and albumin) on the size and  of the lipoplexes was studied as follows. A constant volume of lipoplexes (50 µl) was mixed with increasing concentrations of sputum components dissolved in 50 µl sputum buffer. This resulted in sputum component:pDNA concentration ratios similar to the ones used in the gel electrophoresis experiments. After 30 min these mixtures were diluted with sputum buffer to a final volume of 2.0 ml and the size and  were measured.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS AND DISCUSSION REFERENCES

This work aimed to elucidate the physicochemical stability of lipoplexes in CF mucus. For this purpose lipoplexes were brought in contact with CF sputum components or native CF sputum at different concentration ratios of the sputum component to the complexed pDNA. One of these ratios was the clinical concentration ratio which was calculated from the average concentration of a particular component in the sputum of patients with CF (without acute lung infection) and the concentration of pDNA expected in the mucus after nebulization of lipoplexes into the lungs (14). Also the range in which the clinical concentration ratio may vary, due to variations in the concentration of the component in sputum, was calculated (Tables 1 and 2).

Composition of CF Sputa and Determination of the Clinical Concentration Ratios

Table 1 shows the average concentration of the major electrolytes (Na⁺, Cl, and total calcium) and the main sputum components (DNA, mucin, phospholipids, and proteins) as measured in 24 CF sputa obtained from patients without exacerbations. The average concentrations of the major components in these CF sputa were 2.7 mg/ml DNA, 16 mg/ml mucin, 2.3 mg/ml phosphatidylcholine, 0.3 mg/ml phosphatidylglycerol, and 25 mg/ml protein. Based on the average electrolyte concentrations in these sputa, we also prepared a buffer that simulated the electrolyte composition in CF sputum. This "sputum buffer" contained 85 mM Na⁺, 75 mM Cl, 3 mM Ca²⁺, and 20 mM Hepes at pH 7.4. As the concentration of biopolymers present in CF sputum may depend on the state of infection of the patient, we also determined the DNA and mucin concentration in four CF sputa obtained from patients with acute lung infections (Table 1). It shows that the mean concentration of DNA and mucin is significantly higher when the patients are heavily infected.

Table 2 shows the calculated clinical concentration ratios based on the average concentrations of the sputum components in sputa from patients without exacerbations. It also shows the range in which the clinical concentration ratio may vary owing to variations in the concentration of the sputum components. This range is calculated on the basis of the minimal and maximal concentrations of the components in the CF sputa as mentioned in Table 1. As most of the phospholipids in CF mucus contain two palmitic fatty acid chains, DPPC and DPPG were respectively chosen to represent phosphatidylcholine and phosphatidylglycerol (20).

Size, Zeta Potential, and Dissociation of Lipoplexes by Albumin

Figure 2 shows that upon mixing the lipoplexes with albumin (a negatively charged protein at pH 7.4),  changed from positive to negative values. At the smallest albumin:pDNA concentration ratio tested (6.6),  equaled 21 ± 1 mV. This value decreased steadily to 53 mV upon further increasing the albumin:pDNA concentration ratio to 445. At this concentration ratio  reached a plateau. Except at the albumin:pDNA concentration ratio of 6.6, the size of the lipoplexes was always approximately 280 nm. Albeit albumin changed  of the lipoplexes to negative, gel electrophoresis showed that albumin was never able to release pDNA from the cationic liposomes (data not shown). This indicates that albumin forms ternary complexes with the lipoplexes without liberation of pDNA. Ternary complexes between albumin and cationic polymer-based gene complexes were also reported by Dash and coworkers (21).

View larger version (18K): [in this window] [in a new window]   Figure 2.   The zeta potential and size of lipoplexes mixed with albumin at different albumin:pDNA concentration ratios. The arrow indicates the clinical concentration ratio (n = 3).

In summary, in a clinical situation, where the albumin: pDNA concentration ratio is expected to range between 3 and 33 (Table 2), albumin will not cause dissociation of the lipoplexes. However, the negative  of the lipoplexes after exposure to albumin at these concentration ratios may prevent their binding to the negatively charged surface of the cells, which may lead to a decrease of gene transfer.

Destabilization of Lipoplexes by Mucin

Mucins are negatively charged biopolymers owing the presence of N-acetylneuraminic acid and sulfated sugars (22). Figure 3 shows that mucins changed the  of the lipoplexes to negative values at all the tested concentration ratios (0-220). Moreover, we also found large aggregated lipoplexes at mucin:pDNA concentration between 1.5 and 5, which is just below the clinical concentration ratio (9.5). However, as for albumin, mucin did not cause release of pDNA even when the mucin:pDNA concentration ratio exceeded 23 times the clinical concentration ratio (data not shown).

View larger version (18K): [in this window] [in a new window]   Figure 3.   The zeta potential and size of lipoplexes mixed with mucin at different mucin:pDNA concentration ratios. The arrow indicates the clinical concentration ratio (n = 3).

Size, Zeta Potential, and Dissociation of Lipoplexes by Linear DNA

Lipoplexes were incubated with increasing amounts of linear DNA, dissolved in sputum buffer. The linear DNA:pDNA concentration ratios in these mixtures ranged between 0 and 44. As Figure 4A shows, after gel electrophoresis of an aliquot of these mixtures, a DNA smear was observed at the highest concentration ratios (lanes 5-10 of Figure 4A). Lipoplexes mixed with sputum buffer did not show any release of pDNA (lane 11 of Figure 4A). Depending on the pDNA batch and the amount of loaded pDNA, free pDNA resulted in 1, 2, or 3 bands (lanes 12 and 13 of Figure 4A and lanes 4 and 5 of Figure 8A). Because it was impossible to distinguish between released pDNA and the linear DNA in the DNA smear, Southern blotting was used to visualize pDNA in the gels (Figure 4B). Release of pDNA by linear DNA was observed at a concentration ratio of 1.6 (lane 5 of Figure 4B), which corresponds to the clinical concentration ratio (Table 2). This indicates that lipoplexes nebulized in the CF lung may lose pDNA as soon as the DNA concentration in the mucus reaches 2.7 mg/ml. This is a quite realistic concentration, even in patients without acute lung infection.

View larger version (96K): [in this window] [in a new window]   Figure 4.   Gel electrophoresis on lipoplexes mixed with linear DNA dissolved in sputum buffer (A). The corresponding Southern blot is shown in (B). The linear DNA:pDNA concentration ratios are 0.073 (lane 1), 0.48 (lane 2), 0.79 (lane 3), 1.3 (lane 4), 1.6 (lane 5), 2.9 (lane 6 ), 4.0 (lane 7), 7.9 (lane 8), 16 (lane 9), and 44 (lane 10). Lane 11 contains lipoplexes mixed with sputum buffer. Lanes 12 and 13 contain 0.01 and 0.05 µg of free pDNA, respectively. The arrow indicates the clinical concentration ratio.

View larger version (123K): [in this window] [in a new window]   Figure 8.   Gel (A) and Southern blot (B) after electrophoresis of native sputum/lipoplex mixtures (lanes A1, B1, C1, D1, SS1) and sputum/free pDNA (0.01 µg on gel) mixtures (lanes A2, B2, C2, D2, and SS2). Lane 3 contains lipoplexes exposed to sputum buffer only. Lanes 4 and 5 contain 0.005 µg and 0.01 g of free pDNA, respectively.

Based on the intensity of the pDNA bands on the blot (Figure 4B) and experiments with lipoplexes containing fluorescein-labeled pDNA (data not shown), we found that the percentage of liberated pDNA at concentration ratios of 1.6, 2.9, 4.0, and 7.9 was only 0.6%, 0.7%, 2.0%, and 2.8%, respectively. Consequently, this release of pDNA from the cationic liposomes by linear DNA may not drastically alter transfection efficiency of the gene complexes.

As no smear of linear DNA was detected at low concentration ratios (Figure 4A), it seems that the lipoplexes are able to bind a certain quantity of linear DNA. Again, this binding may lead to changes in the physicochemical properties of the lipoplexes. The influence of this binding of linear DNA on the size and  of the lipoplexes was studied (Figure 5). The ± charge ratio in the linear DNA/lipoplex mixtures, i.e., the ratio of positive charges (arising from the cationic lipids in the liposomes) to the negative charges (on the pDNA and linear DNA) was calculated. Figure 5 shows that this binding of linear DNA is only reflected in a decrease of  and an increase in size of the lipoplexes when the concentration ratio of linear DNA:pDNA becomes higher than 1, which corresponds to a ± charge ratio below 2. This implies that the nebulized lipoplexes in the lungs of patients with CF will neither dissociate nor undergo changes in their size and surface charge when the DNA concentration in the sputum in less than 1.7 mg/ml. At a concentration ratio of 1.6 (corresponding to a ± charge ratio of 1.5),  of the lipoplexes became neutral and large aggregates were revealed from the dynamic light scattering measurements. Gel electrophoresis experiments showed that together with these changes in physicochemical properties of the lipoplexes, pDNA also started at this concentration ratio to get released from the lipoplexes (Figures 1 and 4).

View larger version (23K): [in this window] [in a new window]   Figure 5.   The zeta potential and size of lipoplexes as a function of the linear DNA:pDNA concentration ratio and the ± charge ratio in the linear DNA/lipoplex mixtures. The arrow indicates the clinical concentration ratio (open circle = inaccurate data resulting from the formation of big aggregates [> 3 µm]; n = 3).

As release of pDNA and changes in physicochemical properties of the lipoplexes are determined by charge neutralization, it is expected that the higher the positive charge of the lipoplexes the more resistant they will be toward liberation of pDNA and alterations in physicochemical properties. This may be important to take into account in the design of new gene complexes for CF gene therapy.

We also observed that  of the cationic liposomes did not change upon complexation with pDNA (see METHODS). This indicates that pDNA gets completely entrapped and surrounded by a lipid (bi)layer (Figure 1). This agrees with the observations of others who showed that pDNA becomes coated with lipids (either with a lipid monolayer or double layer) after complexation with DOTAP:DOPE liposomes (23). The more or less constant positive  after adding small amounts of linear DNA (Figure 5) indicates that the linear DNA also is probably surrounded by a lipid double layer. Assuming that the pDNA and linear DNA are sandwiched between a double lipid layer implies that only half of the positive charges from the cationic liposomes can participate in the binding of pDNA and linear DNA (Figure 1). Knowing this, it is expected that  will decrease when the ± charge ratio of the linear DNA/lipoplex mixture falls below 2, because the added linear DNA will then bind to cationic lipids at the surface of the lipoplexes (Figure 1). An abrupt decrease in  at a ± charge ratio below 2 is indeed observed in Figure 5.

Release of pDNA from Lipoplexes by Linear DNA Degraded by rhDNase I

Because the release of pDNA from the lipoplexes by linear DNA may also be influenced by the length of the DNA chains (24), we analyzed the liberation of pDNA by linear DNA having a lower average molecular weight. The linear DNA, dissolved in sputum buffer, was degraded by rhDNase I (3.9 ng of rhDNase I per µg of linear DNA, for 45 min at 37° C). To prevent degradation of liberated pDNA by rhDNase I, we inactivated the rhDNase I by heat treatment (10 min at 100° C) before adding the lipoplexes to the linear DNA/rhDNase I solutions. The lipoplexes were mixed with the degraded DNA at the same concentration ratios as in Figure 4. Figure 6 shows the gel after electrophoresis. In contrast to Figure 4, no pDNA bands were detected in these mixtures using gel electrophoresis (Figure 6) and Southern blot (data not shown). Additionally, compared with Figure 4A, the DNA smear observed on Figure 6 was more concentrated in one spot and located much further from the slots. This was due to fragmentation of linear DNA by rhDNase I. So, it seems that the ability of linear DNA to release pDNA from the lipoplexes depends on the molecular weight. Similar molecular-weight-dependent exchange reactions were observed by Izumrudov and coworkers (25) and Katayose and Kataoka (26) in the competitive binding between DNA and other types of anionic macromolecules to cationic polymers. As it is known that complexed pDNA is not degraded by rhDNase I (27), pretreatment of the patients with a rhDNase I aerosol may completely prevent the dissociation of lipoplexes administered to the CF lung. Although release of pDNA was not observed with low-molecular-weight DNA, the absence of a DNA smear at the lower concentration ratios in the gel electrophoresis experiments (Figure 6) indicates that the small DNA fragments still bind to the cationic lipoplexes, which may alter their gene transfection properties.

View larger version (72K): [in this window] [in a new window]   Figure 6.   Agarose gel after electrophoresis of lipoplexes exposed to rhDNase I degraded linear DNA. The linear DNA:pDNA concentration ratios are 0.073 (lane 1), 0.48 (lane 2), 0.79 (lane 3), 1.3 (lane 4), 1.6 (lane 5), 2.9 (lane 6), 4.0 (lane 7), 7.9 (lane 8), 16 (lane 9), and 44 (lane 10). Lipoplexes mixed with sputum buffer were loaded in slot 11. The amount of free pDNA in lanes 12 and 13 is 0.01 µg and 0.05 µg, respectively. The arrow indicates the clinical concentration ratio.

Size, Zeta Potential, and Dissociation of Lipoplexes by DPPC

In vivo the DPPC:pDNA concentration ratio is expected to vary between 0.26 and 2.1 (Table 2). We found (data not shown) that at these concentration ratios DPPC did not cause any release of pDNA or alteration in size and  of the lipoplexes.

Size, Zeta Potential, and Dissociation of Lipoplexes in the Presence of DPPG

DPPG bears one negative charge at pH 7.4 which may interact with the positive charges on the cationic lipoplexes. We found that, in a clinical situation, where the DPPG:pDNA concentration ratio is expected to range between 0.04 and 0.30 (Table 2) DPPG will cause no release of pDNA (Figure 7A); however, a small decrease and increase of respectively  and lipoplex size was seen when the concentration ratio became higher than 0.18, corresponding with DPPG concentrations in the sputum > 0.3 mg/ml (Figure 7B).

View larger version (52K): [in this window] [in a new window]   Figure 7.   Gel after electrophoresis of lipoplexes mixed with DPPG dissolved in sputum buffer (A). The DPPG:pDNA concentration ratios are 0.18 (lane 1), 0.45 (lane 2), 0.90 (lane 3), 1.8 (lane 4), 2.7 (lane 5), 5.5 (lane 6), 9.1 (lane 7), 15 (lane 8), and 18 (lane 9). Lane 10 contains lipoplexes mixed with sputum buffer. Free pDNA was loaded in lane 11 (0.01 µg) and lane 12 (0.05 µg). In (B), the zeta potential and the size of the lipoplexes as a function of the DPPG:pDNA concentration ratio and the ± charge ratio of the DPPG/lipoplex mixtures are shown. The arrow indicates the clinical concentration ratio (n = 3).

Release of pDNA from Lipoplexes by CF Sputum

From the dissociation experiments of lipoplexes by single sputum components we can conclude that only (nondegraded) linear DNA is able to disassemble lipoplexes at concentration ratios that are clinically relevant. The linear DNA in our experiments, as a model for sputum DNA, was a mixture of DNA fragments having a molecular weight distribution analogous to the molecular weight range of DNA in CF sputum (300 to > 50,000 bases) (28). Therefore, we can expect that in CF sputum also the DNA concentration will be a main parameter governing the liberation of pDNA from lipoplexes. To evaluate whether indeed the DNA concentration in CF sputum will determine whether or not lipoplexes dissociate, we mixed lipoplexes in clinically relevant ratios with different CF sputa. In these experiments 13.5-fold diluted sputa were used. We realize that dilution may to some extent affect the dissociation equilibrium of the lipoplexes. However, the strong viscoelasticity of the CF sputa did not allow gel electrophoresis experiments as the pDNA became sterically blocked in the sputum. Therefore, we were forced to make use of diluted sputa.

In a first series of experiments we used four CF sputum samples obtained from patients with mild lung infections (samples A to D in Table 3), and SS. For each sputum sample we calculated, on the basis of the measured DNA concentrations in these sputa, the expected linear DNA:pDNA concentration ratios in vivo (Table 3). Because CF sputum may show deoxyribonuclease (DNase) activity (owing to the presence of endogenous DNase or administered rhDNase I in the sputum), which can lead to degradation of liberated pDNA from the lipoplexes, we also incubated free pDNA with 50 µl of the sputa as a control. Figure 8A shows that after gel electrophoresis of the sputum/lipoplex and sputum/pDNA mixtures, a smear of DNA, which was more intense for the sputum/ pDNA mixtures, was observed. This indicated that DNA from the sputa had bound to the lipoplexes. To find out whether pDNA was released from the cationic liposomes, the gels were blotted. Figure 8B shows that the pDNA in the controls (i.e., the sputum/pDNA samples) results in a smear (lanes A2, B2, C2, and D2). This smear of pDNA had not migrated as far as the pDNA bands of the standards, which was probably attributed to the higher viscosity of the sputum/pDNA mixtures. Lanes A1, B1, C1, and D1 of Figure 8B show that the lipoplexes mixed with the CF sputa did not dissociate. On the other hand, lipoplexes mixed with SS clearly released pDNA (lane SS1 of Figure 8B). Based on the DNA concentration in SS and the critical DNA concentration of 2.7 mg/ml, this observation was as expected (Table 3). However, dissociation of lipoplexes was also expected in sputum sample C, which has a DNA concentration slightly higher than 2.7 mg/ml. As the DNA concentration in sputum fractions taken from the same sample varies, the DNA concentration in the fraction of sputum sample C used in the gel electrophoresis experiments might have been slightly lower than 2.7 mg/ml.

In a second series of experiments we used two CF sputum samples obtained from patients with acute lung infections (samples E and F in Table 3). The infectious state of the patients was confirmed by the high DNA concentration (> 2.7 mg/ml) in their sputum. The bands in lanes E1 and F1 of Figure 9, which are situated at the same level as the free pDNA bands in the control (lanes E2 and F2), indicate, as expected, that a huge amount of pDNA was released from the lipoplexes. The lower intensity of the pDNA band in E2 is probably due to degradation by DNase or binding of a certain amount of the pDNA to cationic proteins in CF sputum.

View larger version (72K): [in this window] [in a new window]   Figure 9.   Southern blot after gel electrophoresis on native sputum/lipoplex mixtures (lanes E1 and F1) and native sputum/free pDNA (0.01 µg on gel) mixtures (lanes E2 and F2). These CF sputa were obtained from patients with acute exacerbations (i.e., sputa E and F in Table 3). Lane 3 contains 0.01 µg of free pDNA.

Conclusion

In this study we investigated the influence of CF sputum on the integrity of a model lipoplex. This was done as close as possible to the in vivo situation. Destabilization of gene complexes by CF mucus can occur in different ways. First, CF mucus constituents may cause a release of pDNA from the gene complexes. Second, binding of negatively charged CF mucus components to the gene complexes may change their surface charge and size, resulting in a decreased transport of the gene complexes through the mucus and a decreased cellular uptake.

Albumin, mucin, DPPC, and DPPG did not liberate pDNA from our lipoplexes at concentration ratios belonging to the clinical concentration range. However, pDNA was released from the lipoplexes by linear DNA at concentration ratios that are clinically relevant. These concentration ratios will occur in vivo when the DNA concentration in the sputum becomes higher than 2.7 mg/ml. The dissociation of lipoplexes by linear DNA was confirmed in experiments on native CF sputa. Sputa with a DNA concentration lower than the critical DNA concentration of 2.7 mg/ml did not cause any release of pDNA, contrary to sputa with DNA concentrations significantly higher than 2.7 mg/ml. However, the release of pDNA was completely abolished when the linear DNA was fragmented by rhDNase I before it was mixed with the lipoplexes.

Although albumin was not able to disassemble lipoplexes, it clearly bound to the lipoplexes at concentration ratios that were clinically relevant. This changed the surface charge of the lipoplexes to negative values. Also linear DNA and mucin changed at concentration ratios around the clinical ratio  of the lipoplexes to negative values, together with the formation of large aggregates. However, at higher concentration ratios small complexes, with a negative  were formed again. As the amount of released pDNA by linear DNA is low, a decrease in transfection efficiency is mainly expected to be due to the change of  to negative value by albumin, mucin, and linear DNA and also due to the formation of big aggregated lipoplexes by linear DNA at clinically relevant concentration ratios. These aggregates may drastically decrease gene transfer as we showed recently that particles larger than 500 nm became completely entrapped in CF sputum.