Structural studies of the monolayers and bilayers formed by a novel cholesterol-phospholipid chimera.

Langmuir isotherm, neutron reflectivity, and small angle neutron scattering studies have been conducted to characterize the monolayers and vesicular bilayers formed by a novel chimeric phospholipid, ChemPPC, that incorporates a cholesteryl moeity and a C-16 aliphatic chain, each covalently linked via a glycerol backbone to phosphatidylcholine. The structures of the ChemPPC monolayers and bilayers are compared against those formed from pure dipalmitoylphoshatidylcholine (DPPC) and those formed from a 60:40 mol % mixture of DPPC and cholesterol. In accord with previous findings showing that very similar macroscopic properties were exhibited by ChemPPC and 60:40 mol % DPPC/cholesterol vesicles, it is found here that the chimeric lipid and lipid/sterol mixture have very similar monolayer structures (each having a monolayer thickness of ∼26 Å), and they also form vesicles with similar lamellar structure, each having a bilayer thickness of ∼50 Å and exhibiting a repeat spacing of ∼65 Å. The interfacial area of ChemPPC, however, is around 10 Å(2) greater than that of the combined DPPC/cholesterol unit in the mixed lipid monolayer (viz., 57 ± 1 vs 46 ± 1 Å(2), at 35 mN·m(-1)), and this difference in area is attributed to the succinyl linkage which joins the ChemPPC steroid and glyceryl moieties. The larger area of the ChemPPC is reflected in a slightly thicker monolayer solvent distribution width (9.5 vs 9 Å for the DPPC/cholesterol system) and by a marginal increase in the level of lipid headgroup hydration (16 vs 13 H(2)O per lipid, at 35 mN·m(-1)).

Sedimentation field flow fractionation of liposomes.

Quantitative analyses of the particle size distribution of liposomes were performed in 30 to 60 minutes by exponential-field sedimentation field flow fractionation. This gentle new separation method exhibits great potential for the high-resolution fractionation and the size or molecular weight analysis of a wide variety of biological macromolecules and colloidal suspensions.

Evidence for the prion hypothesis: induction of the yeast [PSI+] factor by in vitro- converted Sup35 protein.

Starting with purified, bacterially produced protein, we have created a [PSI(+)]-inducing agent based on an altered (prion) conformation of the yeast Sup35 protein. After converting Sup35p to its prion conformation in vitro, we introduced it into the cytoplasm of living yeast using a liposome transformation protocol. Introduction of substoichiometric quantities of converted Sup35p greatly increased the rate of appearance of the well-characterized epigenetic factor [PSI+], which results from self-propagating aggregates of cellular Sup35p. Thus, as predicted by the prion hypothesis, proteins can act as infectious agents by causing self-propagating conformational changes.

Liposome-encapsulated doxorubicin targeted to CD44: a strategy to kill CD44-overexpressing tumor cells.

Certain tumors, including many that are found in the lung, overexpress the CD44 cell-surface marker. CD44 is a receptor that binds to hyaluronan (HA), a carbohydrate consisting of beta1,3 N-acetyl glucosaminyl-beta1,4 glucuronide. We hypothesized that the incorporation of phosphatidylethanolamine lipid derivatives-containing HA oligosaccharides (HA-PE) into liposomes could target drug-containing liposomes to tumor cells that express CD44. HA-PE containing palmitoyl oleoyl phosphatidylethanolamine or dipalmitoyl phosphatidylethanolamine (HAn-PE) were incorporated into the lipid bilayer at various mole percentages of the total lipids; and the physicochemical properties (diameter, surface charge, and stability) of the resulting liposome preparations were characterized. HA-targeted liposomes (HALs) avidly bound to the CD44-high-expressing B16F10 murine melanoma cell line but not to the CV-1 African green monkey kidney cells, which express CD44 at low levels. Binding of the HALs to the B16F10 cells was rapid, concentration dependent, and saturated at a lipid concentration of about 250 microM. HAL binding to B16F10 was inhibited by HA with high Mr and by an anti-CD44 monoclonal antibody. Binding to the B16 melanoma cells occurred at a lipid composition that contained a > or =0.1 mol % of the HAn-PE lipid. The bound liposomes were internalized by a temperature-dependent process. The IC50s of doxorubicin (DOX) encapsulated in either HALs or nontargeted liposomes and of nonencapsulated DOX were compared in two protocols: continuous exposure of the cells to treatment for 24 h and transient exposure in which the treatment was applied for a 3-h period, and in which non-cell-associated drug was replaced with drug-free medium for the duration of the experiment. The IC50s of free DOX, DOX-loaded nontargeted liposomes, and DOX-loaded HAL (HAL-DOX) for the transient exposure were 6.4 microM, > 172 microM, and 0.78 microM, respectively. For the continuous exposure protocol, the IC50s were 0.60 microm, 25.0 microl, and 0.14 microm, respectively. Thus, in both protocols, delivered DOX was significantly more potent than the nonencapsulated DOX in cells expressing high levels of CD44, which suggests that HALs may be a useful targeted drug carrier to treat CD44-expressing tumors.

The use of aqueous space markers to determine the mechanism of interaction between phospholipid vesicles and cells.

A method has recently been introduced that quantitates the extent of phospholipid vesicle-cell interactions by following the amount of a vesicle-entrapped water-soluble fluorescent probe, carboxyfluorescein (CF) that becomes cell associated (Weinstein, J.N., Yoshikami, S., Henkart, P., Blumenthal, R. and Hagins, W.A. (1977) Science 195, 489--492). We have characterized some of the properties of this probe in sonicated phospholipid vesicles. The CF undergoes a pH-dependent quenching as previously reported and both a pH- and temperature-dependent efflux from vesicles. Decreasing the pH from 7.4 to 5.0 results in almost a 100-fold increase in CF efflux from the vesicles. The simultaneous measurement of cell-associated tritiated lipid and CF fluorescence reveals a discrepancy between the two markers with the tritiated phospholipid becoming associated to 5--10-fold greater extent than the CF. In the presence of cells the leakage of CF from vesicles increases from 1.5- to 10-fold depending on the vesicle composition. This data suggests that interpretations of cell-vesicle interactions followed by the CF technique or other aqueous space markers should be done with caution. However, in experiments where the leakage of CF from vesicles can be controlled, the technique can provide useful information.

Effect of serum components on the physico-chemical properties of cationic lipid/oligonucleotide complexes and on their interactions with cells.

The interactions among serum components and cationic lipid-nucleic acid complexes are central to the understanding of how serum inhibits cellular delivery of oligonucleotides in vitro and in vivo. In this study, we show that several serum proteins, in particular bovine serum albumin (BSA), lipoproteins (HDL and LDL) and macroglobulin, interact with cationic lipid/oligonucleotide complexes, alter the complex diameter and zeta potential (from positive to negative values), and significantly interfere with the ability of 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP) to deliver phosphorothioate oligonucleotides (ODN) into cells. Serum and BSA do not dissociate the ODN and lipid components, therefore inhibition of delivery cannot be attributed to a displacement of cationic lipid from the ODN. Rather BSA at 2.5 mg/ml, comparable to the amount found in 10% serum, decreases the cell association of ODN by about 5-fold and nuclear uptake of ODN by greater than 20-fold. In contrast, immunoglobulin G, the other major serum component, alters the zeta potential from positive to near neutral, has a modest effect on the diameter of the complex but does not affect cell association or nuclear delivery of the ODN at amounts found in 10% serum. Other molecules found in serum, specifically oleic acid and heparin, displace the ODN from the complex and thus interfere with delivery. This displacement is attenuated by first incubating the complex with BSA. Another manifestation of serum-complex interactions is that ODN significantly and cationic liposomes slightly, activate complement. However, formation of the complex markedly reduces the complement activation of the ODN. Finally, the effect of serum can be partially counteracted by the selection of the helper lipid (DOPE or cholesterol). Inclusion of a helper lipid reduces the effective charge ratio (cationic groups/anionic thioates) required to deliver ODN into cells and permits delivery in the presence of greater percentages of serum in the culture medium. These results support the current view that the binding of serum proteins to the complex is a significant factor in modulating the activity of cationic lipid-ODN complexes in culture and after intravenous administration.

Photoaffinity labeling of three renal cyclic 3',5'-adenosine monophosphate-binding proteins.

Evidence is presented for the presence of multiple cyclic AMP binding components in the plasma membrane and cytosol fractions of porcine renal cortex and medulla. N6-(Ethyl-2-diazomalonyl)-3',5'-adenosine monophosphate, a photoaffinity label for cyclic AMP binding sites, exhibits non-covalent binding characteristics similar to cyclic AMP in membrane and soluble fractions. Binding data for either compound to the plasma membrane fraction yields biphasic Scatchard plots while triphasic plots are obtained with the dialyzed cytosol. When covalently labeled fractions are separated on SDS-polyacrylamide gel electrophoresis, the cyclic AMP photoaffinity label is found on 49 000 and 130 000 dalton components in each kidney fraction. DEAE-cellulose and gel filtration chromatography of the labeled cortical cytosol fraction establishes that the three components suggested by the binding data correspond to two 49 000 dalton species and a 130 000 component. The 49 000 species have higher affinities for cyclic AMP than the 130 000 component (Ka(1) = 2.0 . 10(9), Ka(2) = 1.7 . 10(8), Ka(3) = 1.0 . 10(7)). The 49 000 components are associated with protein kinase activity while the 130 000 component does not exhibit protein kinase, adenosine deaminase, or cyclic nucleotide phosphodiesterase activity. Immunologic results and effects of phosphorylation and cyclic GMP on cyclic AMP binding further suggest that the 49 000 components are regulatory subunits of cyclic AMP-dependent protein kinases. Cyclic AMP binding to the 130 000 component is markedly inhibited by adenosine and adenine nucleotides, but not cyclic GMP. Thus, this component may reflect an aspect of adenosine control or metabolism which may or may not be a cyclic AMP-related cellular function.

Preparation of liposomes of defined size distribution by extrusion through polycarbonate membranes.

Liposomes of defined size and homogeneity have been prepared by sequential extrusion of the usual multilamellar vesicles through polycarbonate membranes. The process is easy, reproducible, produces no detectable degradation of the phospholipids, and can double the encapsulation efficiency of the liposome preparation. Multilamellar vesicles extruded by this technique are shown by both negative stain and freeze-fracture electron microscopy to have mean diameters approaching the pore diameter of the polycarbonate membrane through which they were extruded. When sequentially extruded down through a 0.2 micron membrane, the resulting vesicles exhibit a very homogeneous size distribution with a mean diameter of 0.27 micron while maintaining an acceptable level of encapsulation of the aqueous phase.

Incorporation of LPS in liposomes diminishes its ability to induce tumoricidal activity and tumor necrosis factor secretion in murine macrophages.

We investigated the effect of lipopolysaccharide (LPS) incorporated into phospholipid vesicles (liposomes) on the induction of macrophage-mediated tumor cytotoxicity and tumor necrosis factor (TNF) secretion. The incorporation of Salmonella minnesota rough (Re)-LPS into multilamellar or small unilamellar vesicles (liposomes) resulted in an 100- to 1,000-fold reduction in its potency to activate both the macrophage cell line RAW 264.7 and murine thioglycolate elicited peritoneal macrophages to become cytotoxic for L929 and P815 tumor cells. Liposomal LPS was also a 100- to 1,000-fold less potent inducer of TNF secretion from RAW 264.7 cells. Cytokines secreted by the activated macrophages contributed to the cytotoxic effect on the L929 cells but not the P815 cell line. Human recombinant TNF was not cytotoxic for either cell line but was cytostatic for the L929 cell line. Morphological examination of the cells after uptake of fluorescent, free, and liposomal LPS revealed that both forms were internalized by the endocytic pathway. This, together with the considerably reduced potency of liposomal LPS to induce tumor cytotoxicity and TNF secretion, suggests that the interaction of the hydrophobic part of the lipid A moiety of LPS with the macrophage plasma membrane is needed to optimally activate these cells. Incorporation of LPS into liposomes effectively abrogates this interaction.

Lipoplex-mediated gene delivery to the lung occurs within 60 minutes of intravenous administration.

We have defined the critical time period for gene delivery in the lung after intravenous administration of cationic lipoplex. We accomplished this through the displacement of intravenously injected cationic lipoplexes from the lungs by the subsequent administration of anionic liposomes. When reporter gene-bearing lipoplexes were injected intravenously and followed by anionic liposomes 5 min later, reporter gene expression was reduced up to 400-fold compared with animals into which lipoplex alone was administered. Administration of anionic liposomes 60-90 min after lipoplex injection yielded no significant reduction in lung transfection. When lipoplexes were disrupted 5 min after administration, the pulmonary distribution of the cationic lipid and DNA components was reduced by 80%. Lipids subsequently accumulated primarily in the liver, while the plasmid DNA constituent distributed into the blood and liver. As the interval between lipoplex and anionic liposome injection increased, the degree of lipoplex displacement from the lung decreased to such a point that, 60 min after lipoplex injection, the anionic liposome injection did not displace significant quantities of the lipoplex. We conclude that cationic lipid-DNA complexes can be disrupted in vivo via the administration of anionic liposomes; moreover, we have employed this phenomenon to demonstrate that transfectionally active DNA is taken up within 60 min of systemic lipoplex administration.

Effects of complement depletion on the pharmacokinetics and gene delivery mediated by cationic lipid-DNA complexes.

We show that lipoplexes activate complement in human serum in vitro and deplete complement when administered intravenously (i.v.) to mice. This raised the possibility that complement proteins might alter gene expression mediated by lipoplex in animals. To investigate this phenomenon, complement levels were depleted to less than 5% in ICR mice by intraperitoneal (i.p.) injection of cobra venom factor and anti-C3 antibodies. The pharmacokinetics and distribution of radio labeled DOTAP-cholesterol (1.0:0.9 molar ratio)-DNA (5:1 positive charge ratio) complexes containing 131I-labeled p-hydroxybenzamidine phosphatidylethanolamine and 125I-labeled DNA were measured in mice after i.v. administration. Greater than 75% of the injected lipoplex appeared in the lungs 5 min following injection. The lipid and DNA were eliminated from the lungs at a constant ratio. Distribution in various organs was not affected by complement depletion. Expression of luciferase was highest in the lungs and showed a dose-dependent increase as the amount of DNA injected increased from 3 to 60 microg. Reporter gene expression was not affected by complement depletion. In addition, complement depletion had no effect on either the distribution or gene expression in the heart, spleen, or liver. We conclude that cationic lipid-DNA complexes interact with serum complement proteins upon i.v. injection in mice, but this interaction does not influence the lipofection efficiency or systemic distribution of the lipoplex.

Branched cationic peptides for gene delivery: role of type and number of cationic residues in formation and in vitro activity of DNA polyplexes.

To examine the suitability of synthetic peptides as DNA-binding and -compacting agents for receptor-mediated gene delivery, we have synthesized and characterized a series of branched oligocationic peptides that differ in the number and type (lysine, arginine, ornithine) of cationic amino acids in the DNA-binding moiety. The peptides were designed as branched molecules to provide a coupling site via a spacer for the attachment of effectors at a flexible distance from the DNA-binding moiety. This design provides torsional flexibility in the peptide backbone of the DNA-binding moiety to maximize cation-DNA phosphate interactions and also minimizes the potential for interference by the effector with DNA binding. The branched peptides bind DNA with affinities that increase with the number of cationic groups. The peptides compact DNA into microparticulate structures as judged by an ethidium bromide displacement assay, dynamic light scattering, and electron microscopy. In general, differences in DNA binding and compaction owing to variation in the cationic side chain were modest, with the rank order being arginyl > lysyl approximately ornithyl. Incorporation of tryptophans into the DNA-binding moiety had no major effect on apparent binding affinity but clearly reduced the DNA-compacting potency of the peptides. Compared with polylysine, the peptides and their DNA complexes are weak activators of the complement system. Complement activation by an octaarginyl peptide was stronger than that induced by an octalysyl peptide. The microparticulate peptide-DNA complexes are suitable for receptor-mediated gene delivery as evidenced by transferrinfection of K562 cells in the presence of chloroquine. The results obtained in gene delivery in vitro suggest that a minimum chain length of six to eight cationic amino acids is required to compact DNA into structures active in receptor-mediated gene delivery.

Activation of the complement system by synthetic DNA complexes: a potential barrier for intravenous gene delivery.

We have examined the complement-activating properties of synthetic cationic molecules and their complexes with DNA. Commonly used gene delivery vehicles include complexes of DNA with polylysine of various chain lengths, transferrin-polylysine, a fifth-generation poly(amidoamine) (PAMAM) dendrimer, poly(ethyleneimine), and several cationic lipids (DOTAP, DC-Chol/DOPE, DOGS/DOPE, and DOTMA/DOPE). These agents activate the complement system to varying extents. Strong complement activation is seen with long-chain polylysines, the dendrimer, poly(ethyleneimine), and DOGS (half-maximal at about 3 microM amine content in the assay used). Compared to these compounds, the other cationic lipids (in liposome formulations) are weak activators of the complement system (half-maximal approximately 50-100 microM positive charge in assay). Complement activation by polylysine is strongly dependent on the chain length. Short-chain oligolysines are comparable to cationic lipids in their activation of complement. Incubation of these compounds with DNA to form complexes reduces complement activation in virtually all cases. The degree of complement activation by DNA complexes is strongly dependent on the ratio of polycation and DNA (expressed as the charge ratio) for polylysine, dendrimer, poly(ethyleneimine), and DOGS. To a lesser degree, charge ratio also influences complement activation by monovalent cationic lipid-DNA complexes. For polylysine-DNA complexes, complement activation can be considerably reduced by modifying the surface of preformed DNA complexes with polyethyleneglycol (half-maximal approximately 20 microM amine content). The data suggests that, by appropriate formulation of DNA complexes, complement activation can be minimized or even avoided. These findings should facilitate the search for DNA complex formulations appropriate for reproducible intravenous gene delivery.

Gene expression following direct injection of DNA into liver.

The liver is an attractive target tissue for gene therapy. Current approaches for hepatic gene delivery include retroviral and adenoviral vectors, liposome/DNA, and peptide/DNA complexes. This study describes a technique for direct injection of DNA into liver that led to significant gene expression. Gene expression was characterized in both rats and cats following injection of plasmid DNA encoding several different proteins. Luciferase activity was measured after injection of plasmid DNA encoding the luciferase gene (pCMVL), beta-galactosidase (beta-Gal) activity was evaluated in situ using plasmid DNA encoding Lac Z (pCMV beta), and serum concentration of secreted human alpha-1-antitrypsin was measured following injection of plasmid DNA encoding this protein (pRC/CMV-sHAT). Several variables, including injection technique, DNA dose, and DNA diluent, were investigated. Direct injection of pCMVL resulted in maximal luciferase expression at 24-48 hr. beta-Gal staining demonstrated that the majority of transfected hepatocytes were located near the injection site. Significant concentrations of human alpha-1-antitrypsin were detected in the serum of animals injected with pRC/CMV-sHAT. These findings demonstrate the general principle that direct injection of plasmid DNA into liver can lead to significant gene expression.

A procedure for the efficient incorporation of wild-type lipopolysaccharide into liposomes for use in immunological studies.

Previous studies on the mechanism of action of lipopolysaccharides (LPS) on macrophages have used wild-type lipopolysaccharide (wt-LPS) containing liposomes. In these studies the endotoxin was incorporated into liposomes by suspending the wt-LPS in the buffer used to rehydrate the lipid. Using this approach (buffer method), we observed that less than 10% of Salmonella minnesota smooth LPS is incorporated into multilamellar vesicles (MLV). If the non-incorporated material is not effectively separated from the liposomal form, erroneous conclusions on the mechanism of action of LPS can be drawn. Prolonged sonication of the wt-LPS-MLV suspension resulted in almost complete incorporation of the LPS into the resulting small unilamellar vesicles (SUV). In order to prepare MLV, we briefly soniated the buffer preparation, dehydrated the resulting smaller vesicles and then rehydrated the mixture (dry method). This procedure resulted in almost complete incorporation of the wt-LPS into MLV. The ability of wt-LPS in MLV prepared by the dry method to activate macrophages or trigger gelation of Limulus amoebocyte lysate was reduced by 100-1000-fold compared to the non-incorporated wt-LPS. This indicates that at least 99% of the wt-LPS is incorporated in MLV made by the dry method.

Synthesis and characterization of long chain alkyl acyl carnitine esters. Potentially biodegradable cationic lipids for use in gene delivery.

A series of alkyl acyl carnitine esters (alkyl 3-acyloxy-4-trimethylammonium butyrate chloride) were synthesized as potential biocompatible cationic lipids for use in gene transfer. The physicochemical properties of the lipids, liposomes prepared from them, and their complexes with DNA were characterized by differential scanning calorimetry (DSC), particle size, zeta potential, and surface monolayer measurements. The transition temperatures and behavior at an air-water interface for this series are similar to phosphatidylcholines with the same hydrocarbon chain length. The physical properties of the l derivatives were not significantly different from the dl derivatives. At 70 degrees C, the acyl chains were readily hydrolyzed at pH 7. The influence of the aliphatic chain length (n = 12-18) on transfection efficiency in vitro was determined using cationic liposomes prepared from these lipids or their mixtures with the helper lipids, dioleoylphosphatidylethanolamine (DOPE), dioleoylphosphatidylcholine, monooleoylglycerol, and cholesterol (Chol). The mixture of myristyl 3-myristoyloxy-4-trimethylammonium butyrate chloride (MMCE, 4d) with DOPE at a 1:1 molar ratio mediated the highest transfection efficiency in cell culture. The mixture of oleyl 3-oleoyloxy-4-trimethylammonium butyrate chloride (OOCE, 4f) with Chol at a 1:1 molar ratio gave the highest transfection efficiency after intravenous administration in mice. In vivo gene expression using 4f was comparable to values obtained with the best cationic lipids reported to date.

A simple in vitro model to study the release kinetics of liposome encapsulated material.

A simple in vitro model was developed to study the release kinetics of liposome encapsulated material in the presence of biologic components. Liposomes were embedded in an agarose gel (bottom layer) formed in a glass vial and separated from the receptor compartment buffer by a second layer of agarose gel (top layer). To follow the release of liposomal contents, aqueous space markers differing in molecular weight (from 205 Dalton to 17500 Dalton) were encapsulated. The isotonic buffer in the receptor was completely changed at various time points and the amount of marker released from the agarose matrix containing the liposomes into the receptor medium determined. The release of non-encapsulated markers from the gel followed a time0.5 relationship with about 75% of a 17500 Dalton protein being released from the matrix in 48 h. In the same period, about 7% of the intact liposomes added to the agarose gel appeared in the receptor phase. The release of calcein from various liposome compositions including: (A) egg phosphatidylcholine (EPC)/egg phosphatidylglycerol (EPG) 9:1, (B) dioleoylphosphatidylethanolamine (DOPE)/cholesterylhemisuccinate (CHEMS) 2:1, and (C) dioleoylphosphatidylglycerol (DOPC)/dioleoylphosphatidylglycerol (DOPG) 2:1 was measured. Components of the biological milieu such as serum proteins and calcium influenced release of encapsulated material. This in vitro model is a convenient and reproducible system that permits the study of the release of high molecular weight molecules such as proteins from liposomal formulations in the presence of serum. It may find applications with respect to release of proteins from a variety of colloidal drug delivery systems.

Optical measurement of aqueous potassium concentration by a hydrophobic indicator in lipid vesicles.

An assay was developed for K+ in aqueous solution at neutral pH. The method was based on the change in optical absorbance of the hydrophobic indicator 7-(n-decyl)-2-methyl-4-(3',5'-dichlorophen-4'-one)indonaphthl++ +-1-ol (MEDPIN) in phospholipid vesicles. Formation of a ternary complex between a valinomycin-K+ pair and the anionic form of MEDPIN in the bilayer resulted in an absorption band at 584 nm. K+ concentration was determined by monitoring the MEDPIN absorbance at 584 nm and MEDPIN quenching of lissamine rhodamine B sulfonylphosphatidylethanolamine (L-RhB-PE) fluorescence by an energy-transfer mechanism. Both the fluorescence intensity and lifetime of L-RhB-PE decreased by more than 25% upon addition of 50 mM K+. Kinetic studies using stopped-flow photometry showed a single-exponential reaction of MEDPIN and valinomycin in vesicles with aqueous K+ (maximum rate 1.7 s-1) that was dependent upon [valinomycin] and [K+]. The lipid surface charge was shown to influence the ratio of anionic to neutral MEDPIN at constant pH, and to alter the sensitivity of MEDPIN absorbance to aqueous [K+]. A 1:20 neutral/negative lipid mole ratio was optimal for K+ detection at pH 7.4. Spectroscopic and kinetic data suggest that the optical response of MEDPIN to K+ involves the formation of a ternary complex between K+, valinomycin and MEDPIN.

Gene delivery and expression in human retinal pigment epithelial cells: effects of synthetic carriers, serum, extracellular matrix and viral promoters.

Non-viral gene therapy is a potential treatment to many incurable retinal diseases. To fulfill this promise, plasmid DNA must be delivered to the retinal target cells. We evaluated the efficacy of synthetic DNA complexing compounds in transfecting primary human retinal pigment epithelial (RPE) cells in vitro. Fetal human RPE cells were cultured with or without extracellular matrix (ECM), produced using calf corneal endothelial cells. Plasmids encoding nuclear localizing beta galactosidase or luciferase (pRSVLuc, pCLuc4, pSV2Luc) were complexed in water at various +/- charge ratios using cationic lipids (Lipofectin, DOTAP, DOGS), polyethylene imines (25 and 750 kDa), and with degraded 6th generation starburst polyamidoamine dendrimers. Luciferase was quantified using a luminometric assay and beta galactosidase with X-gal staining. Toxicities of transfections were evaluated with the MTT-assay. Using beta galactosidase as the reporter gene naked DNA did not transfect RPE cells at measurable levels whereas 1-5% of the cells expressed histochemically detectable amounts of the gene after transfection with cationic lipid DNA complexes. In RPE cells, Rous sarcoma virus and cytomegalovirus (CMV) were more efficient promoters than SV40 in driving luciferase expression, and CMV was chosen for further experiments. At optimal complex charge ratios, expression levels of luciferase were > 10(9) light units/mg protein after transfection using dendrimers and PEI25, while transfection mediated with the other carriers resulted in luciferase expression levels of 10(7)-10(9) light units/mg protein or less. In general, dendrimers and large molecular weight PEI were less toxic than cationic lipids or PEI25 to RPE cells. Serum and ECM decreased gene expression to the RPE cells with all carriers. Despite low percentage of transfected cells the transgene expression per RPE cell is high, important feature in the retinal tissue with small dimensions, in particular in the case of secreted gene products. Degraded dendrimers and high molecular weight PEI exhibited the best combination of high activity and low toxicity in RPE cell transfection.